Full Report

The Trade Desk sells software that helps advertisers buy ads. To understand the company, you first have to understand the plumbing of the modern digital-advertising market — who sits on which side of the table, where the money leaks out, and why an independent, buy-side-only middleman can be a durable business inside an arena dominated by Google and Amazon. This tab builds that mental model from the primary record, and cites every material claim back to the exact filing page that supports it.

1. The thirty-second orientation

Digital advertising is now the largest and fastest-growing part of the global ad market — reported annual spend of over $700 billion, more than 70% of all ad dollars, inside a total advertising market that has surpassed $1 trillion [1]. The Trade Desk operates a demand-side platform ("DSP") — a self-service, cloud-based cockpit that ad buyers use to plan, execute, optimize and measure campaigns across connected TV, video, display, audio and native inventory [2]. It does not own media, and it does not sell its own inventory. It charges a platform fee, generally a percentage of the advertiser's total spend, plus fees for data and value-added services [3].

That "percentage of spend" model is the single most important fact about the economics: TTD's reported revenue is a thin slice of the far larger river of ad dollars — gross spend — flowing across its platform.

Source: FY2025 Form 10-K — gross spend $13,394.7M, revenue $2,896.3M and Adjusted EBITDA $1,196.4M per MD&A Executive Summary [4]; retention rate that has exceeded 95% for over a decade [5]; 3,843 full-time employees in 21 countries [6].

The implied take rate — revenue divided by gross spend — was 21.6% in FY2025, up from 20.3% in FY2024 [7]. That rising monetization is the crux of the bull case; a falling take rate is management's own stated warning sign, because revenue "may not necessarily grow at the same rate as spend" as pricing competition, volume discounts and channel mix shift against it [8].

2. How the money moves: the programmatic value chain

"Programmatic" simply means buying and selling ad impressions electronically, in real time, one impression at a time, through automated auctions — instead of humans negotiating insertion orders. When you load a web page or a streaming app, an auction for the ad slot in front of you resolves in milliseconds. TTD's AI system alone evaluates roughly 15 million ad opportunities every second, weighing hundreds of variables on each one [9].

The ecosystem divides cleanly into buyers, sellers and the marketplace in between — and management's governing belief is that a participant should represent one side, not referee a game it also plays in [10].

Source: FY2025 Form 10-K, Our Industry (buyers/sellers/marketplace structure) [11]; competition described as highly competitive and fragmented, competing with DSPs and divisions of Google and Amazon [12].

TTD sits squarely in the DSP layer, and its whole brand is that it is a pure buy-side advocate with no conflicting interest in any particular publisher's inventory. Its industry, in its own words, is "highly competitive and fragmented," and it competes with smaller private DSPs as well as the DSP divisions of "large, well-established companies such as Google and Amazon" [13].

3. The growth engine: programmatic maturing, CTV taking over

Three structural currents drive the arena, all documented in the filings:

• The shift to programmatic. Automated, data-driven buying is becoming the predominant way advertisers reach consumers, replacing manual media buying [14].
• The generational shift to Connected TV (CTV). Viewers are abandoning linear cable for ad-supported streaming, and those impressions are biddable, data-rich and premium — "a generational shift from linear television to connected television" [15]. CTV is now TTD's growth engine: video, which includes CTV, has grown to a "high 40s percentage share" of the platform's business [16].
• Expanding global TAM. The total addressable market for advertising has crossed $1 trillion, and management sees international markets — the U.K., Germany, France, China, Japan, India, Australia — as the long runway [17].

The way to see the industry's trajectory is through TTD's own revenue arc — a decade of rapid but visibly decelerating growth as programmatic matures from novelty to mainstream.

Source: Company SEC filings (FY2016–FY2025); FY2024 revenue $2,444.8M and FY2025 revenue $2,896.3M per FY2025 Form 10-K MD&A [18].

Source: Company SEC filings (FY2017–FY2025); FY2025 revenue grew 18% year-over-year per FY2025 Form 10-K MD&A [19].

Growth of 18% in FY2025 is still enviable for a company this size, but it is roughly a third of the ~55% pace of 2018 [20]. The industry is transitioning from land-grab to share-grab — which raises the stakes on execution and competitive positioning, the subjects of the next two sections.

4. Unit economics: why the model is attractive

A DSP is, at heart, software: once built, each incremental dollar of ad spend routed through it costs very little to serve. That shows up as high gross margins and expanding operating leverage. The FY2025 income statement walks revenue down through the four cost buckets to a 20% operating margin [21].

Source: FY2025 Form 10-K, Results of Operations — revenue $2,896M; platform operations $619M; sales & marketing $644M; technology & development $525M; G&A $518M; income from operations $589M [22].

Two features of the model are worth teaching a newcomer:

• Recurring, sticky relationships. TTD contracts through ongoing master services agreements, not one-off orders, and reports a customer retention rate that has exceeded 95% for over a decade [23]. Ad budgets are decided at the agency/branch level, but if aggregated to the holding-company parent, two holding companies each represented more than 10% of gross billings in 2025 — a concentration risk hiding inside a fragmented client base [24].
• Working-capital and credit risk sit in the plumbing. Because most spend flows through agencies, TTD carries "sequential liability" — it often must pay inventory and data suppliers on shorter cycles than it collects from agencies, and if an advertiser doesn't pay the agency, TTD must chase the advertiser directly [25]. This is a structural feature of the whole ad-buying industry, not a TTD quirk.

5. Competitive structure: the independents vs. the walled gardens

The industry's most important dividing line is ownership of media. On one side are the walled gardens — Google and Amazon — which monetize their own enormous owned-and-operated inventory and run buy-side and sell-side tools. On the other side is the independent open internet, where TTD is the largest buy-side operator and a cluster of independent sell-side platforms (Magnite, PubMatic) and rival DSPs (Viant) compete.

The scale asymmetry is staggering and worth internalizing: TTD's entire platform routes ~$13.4 billion of gross spend, while Google's advertising revenue alone is roughly twenty times TTD's spend.

Source: Alphabet FY2025 10-K, Google advertising revenue $294,691M [26]; Amazon FY2025 10-K, advertising services revenue $68,635M [27]; TTD gross spend $13,395M [28].

Among the independents, TTD is not just the largest — it is by far the most profitable, the clearest evidence that scale and a buy-side focus compound into superior economics. Magnite bills itself as "the world's largest independent omni-channel sell-side advertising platform" [29]; PubMatic positions itself as an "independent infrastructure provider prioritizing transparency" that owns no media [30]; and Viant runs a rival cloud-based DSP [31].

Source: FY2025 10-K / income statements — TTD revenue $2,896M, operating income $589M, net income $443M [32]; Magnite revenue $714M, operating income $98M, net income $145M [33]; PubMatic revenue $283M, operating loss $17M [34]; Viant revenue $344M, operating income $12M, net income $8M [35].

Criteo, a French commerce/retail-media specialist, rounds out the independent set but reports on a different revenue basis, so it is excluded from the table above. The takeaway is structural: TTD's ~20% operating margin dwarfs every other independent — Magnite (14%), Viant (4%), PubMatic (loss) — a gap that reflects both its scale and the fact that buy-side share is where budgets and pricing power concentrate [36].

Two competitive dynamics from management's own mouth sharpen the picture:

• Google is de-prioritizing the open internet. Jeff Green argues TTD competes against something like "the 47th highest priority at Google" as the company reallocates attention to Gemini, Cloud, AI, Search and YouTube — an opening for a focused independent [37].
• Amazon is the rising threat but structurally narrow on retail data. TTD counters that the retailers in its data marketplace represent more than 80% of sales from top U.S. retailers, versus Amazon's own share of under 15% of U.S. retail spend — i.e., TTD aggregates the rest of retail against Amazon's first-party walled garden [38].

6. Where the cycle sits: the 2024–2025 inflection

Ad tech is sensitive to the macro cycle — advertising budgets flex with the economy — but the defining recent event was self-inflicted, not macro. In Q4 2024, after 8 years and 33 consecutive quarters of beating its own guidance as a public company, TTD missed for the first time, calling it "our fault" [39]. Even so, platform spend that year exceeded $12 billion, the highest in its history [40].

Management attributed the stumble to a slower-than-planned migration to Kokai, its "most recent and major upgrade" to the platform — an AI-forward rebuild it chose to roll out deliberately rather than fast [41], coupled with an internal reorganization. The following quarter the business reaccelerated, growing revenue 25% year-over-year and surpassing its own expectations in Q1 2025 [42]. For an industry newcomer, the lesson is that a DSP's moat is its software and its client trust — a botched platform transition, not a demand shock, was what briefly cracked the story.

7. Regulation and identity: the industry's existential debate

Two intertwined forces — privacy regulation and the mechanics of ad targeting — shape the arena's long-term structure more than anything else.

Privacy law is tightening and fragmenting. In the U.S., the FTC polices the industry under Section 5 of the FTC Act, and a growing patchwork of state omnibus privacy laws now forces opt-outs for "targeted advertising" and treats device IDs, cookies and hashed emails as regulated personal data [45]. Compliance cost and legal exposure are rising for every buyer, seller and platform in the chain [46].

The identity war is the response. Targeting has historically relied on third-party cookies. TTD's strategic bet is Unified ID 2.0 (UID2) — an open-source framework that turns emails and phone numbers into a pseudonymous advertising identifier, with a European variant (EUID) — as the open internet's alternative to cookies and to the walled gardens' logged-in data [47]. Critically, in 2025 Google reversed course and announced it will not eradicate third-party cookies from Chrome — removing a near-term cliff but leaving the long migration to first-party identity intact [48].

Antitrust may reshape the walled gardens. Also in 2025, U.S. courts declared Google an illegal monopoly in two separate matters [49]; Google's own filing acknowledges the DOJ ad-tech ruling that its publisher tools unfairly excluded rivals [50]. Any forced separation of Google's buy-side and sell-side would be a direct tailwind for independent operators like TTD.

TTD's own answer on the supply side is OpenPath, which lets it connect more directly to publishers and strip out intermediary toll-takers; early adopter The New York Post reported a 97% boost in programmatic display revenue [51].

8. The watchlist: what would change the industry view

Source: take-rate and spend metrics per FY2025 Form 10-K MD&A [52]; CTV mix [53]; Google ad-tech ruling [54]; identity strategy [55].

The one-paragraph synthesis. The Trade Desk sits in the buy-side seat of a maturing, ~$700bn-and-growing digital-ad market, monetizing a small but rising slice of a $13.4bn river of spend at industry-best margins because it is the largest independent operator in an arena otherwise ruled by Google and Amazon [56] [57]. Its growth engine is the generational move of TV to streaming; its risks are decelerating growth, take-rate compression, Amazon's ascent, privacy regulation and its own platform execution [58]. The whole thesis hinges on one question a newcomer should keep front of mind: can a neutral, buy-side-only platform keep taking share of the open internet faster than the walled gardens can wall it off?

---

The Business: A Toll Booth on Open-Internet Ad Spend

The Trade Desk is one of the highest-quality business models in software — a capital-light, buy-side-only operating system that skims a rising percentage fee off a $13.4 billion river of advertising spend it never owns, at margins no independent rival comes close to matching [1]. It is also, as of mid-2026, priced like a broken one: the stock has fallen roughly 78% from its 2025 high, and the market now pays about 3.3 times sales for a business still growing 18% with a 27% free-cash-flow margin. The entire investment question sits in that gap between quality and price — and it turns on one number the bulls and bears both watch: the take rate.

Sources: gross spend $13,394.7M, revenue $2,896.3M, Adjusted EBITDA per FY2025 10-K MD&A Executive Summary [2]; operating income $589.3M per Consolidated Statements of Operations [3]; free cash flow derived from operating cash flow $992.7M less capex $197.0M per Consolidated Statements of Cash Flows [4]; FCF yield computed on a $9.64B market cap at the $19.53 close (10-Jul-2026).

1. The economic engine — a percentage fee on spend it never owns

The single most important fact about The Trade Desk is that reported revenue is a thin slice of a much larger flow. Clients route their advertising budget through the platform to buy inventory, data and value-added services; on top of that spend the company adds a platform fee, generally a percentage of the client's total spend [5]. Because the company acts as an agent rather than a principal, it recognizes revenue net — it keeps the fee, not the media dollars. That is why $13.4 billion of gross spend became $2.9 billion of revenue in FY2025 [6].

The take rate — revenue divided by gross spend — is therefore the master variable. It rose to 21.6% in FY2025 from 20.3% the year before, driven by higher utilization of value-added services and data and higher platform fees [7]. A rising take rate is the crux of the bull case: it means the company is monetizing each dollar of spend more richly. A falling take rate is management's own stated warning sign — the filing explicitly cautions that revenue "may not necessarily grow at the same rate as spend on our platform" as pricing competition, volume discounts and channel-mix shifts bite [8].

Source: gross spend from FY2025 10-K MD&A and prior-year segment KPIs; revenue per Consolidated Statements of Operations [9] [10].

Source: derived from reported gross spend and revenue; the FY2025 take rate of 21.6% (up from 20.3% in FY2024) reflects higher platform fees and greater utilization of value-added services and data, per FY2025 10-K MD&A [11].

The take rate held in a tight 19–20% band for years and then stepped up to 21.6% in FY2025 — the visible signature of the company selling more high-margin data and measurement on top of media, not just routing more dollars. It is the number to track every quarter.

2. Why the revenue is recurring — and where the moat starts

Two structural features make this a durable toll, not a spot fee. First, the company contracts through ongoing master services agreements (MSAs) rather than one-off insertion orders, and reports a customer retention rate that has exceeded 95% for over a decade [12]. Retention above 95% for ten-plus years is rare in any software category and is the strongest single piece of moat evidence in the filing.

Second, the platform gets stickier the more a client uses it: buyers integrate their own first-party data into the platform to optimize campaigns, and the bidding engine lets them tune hundreds of granular bid factors against specific business objectives [13]. A team that has wired its data, its measurement and its optimization logic into The Trade Desk does not casually re-platform. The self-service model reinforces this: clients are trained to run campaigns independently, which deepens habit and lowers the company's own cost to serve [14].

The counterweight — and it is material — is client concentration at the holding-company level. Budgets are decided at the agency/brand level, so the client base looks fragmented; but if aggregated to the parent, two holding companies each represented more than 10% of gross billings in 2025 (up from one in 2024) [15]. The moat is real, but a couple of agency holding companies hold genuine negotiating leverage over the take rate.

3. The balance sheet is a mirage — read it as agency plumbing

A newcomer opening The Trade Desk's balance sheet sees something alarming: $3.77 billion of receivables against $3.01 billion of payables on a company with under $3 billion of revenue. This is not leverage or aggressive accounting — it is the agency float inherent to the take-rate model. The company invoices agencies for the gross media spend (media dollars plus its fee), then pays inventory and data suppliers. Because most spend flows through agencies, it carries "sequential liability": it often must pay suppliers on shorter cycles than it collects from agencies, and if an advertiser fails to pay its agency, the company must chase the advertiser directly [16].

Source: FY2025 10-K Consolidated Statements of Cash Flows and balance-sheet lines; receivables grew $432.7M in the year and payables $291.1M [17].

Two investor takeaways follow. (1) There is no financial-leverage risk — the company carries no funded debt, and cash plus short-term investments of roughly $1.3 billion exceed every liability that is not agency float. (2) Working capital is a permanent, growing use of cash: receivables climb with gross billings faster than payables, so a slice of reported profit is continuously reinvested into the float. That is why operating cash flow, though strong, does not simply track net income up in a straight line — the model funds its own growth in the receivable book [18].

4. Unit economics: real operating leverage, with an asterisk

A DSP is fundamentally software — once built, each incremental dollar of routed spend costs almost nothing to serve. That shows up as expanding operating margins as revenue scales. Margins are also noisy year to year because the FY2021 CEO stock-option grant runs through general and administrative expense, but the direction is unmistakable: from a ~7% operating margin in the 2022 trough to 20.3% in FY2025 [19].

Source: derived from FY2025 10-K Consolidated Statements of Operations (three-year detail) and prior filings; FY2025 operating income $589.3M on revenue $2,896.3M [20].

The asterisk is stock-based compensation. The company adds back $490.6 million of stock-based compensation to reach its headline Adjusted EBITDA of $1,196.4 million [21]. That add-back is roughly 17% of revenue and is a real economic cost — it dilutes shareholders whether or not it touches cash. The honest way to read profitability is to walk from the promotional number down to owner cash earnings.

Source: Adjusted EBITDA $1,196.4M and SBC $490.6M per FY2025 10-K Non-GAAP reconciliation [22]; reported free cash flow $795.7M (operating cash flow $992.7M less capex $197.0M) per Statements of Cash Flows [23]. Bridge steps are illustrative allocations of the same reported items; the endpoints — Adjusted EBITDA, FCF and SBC — are as reported.

The point is not that the numbers are fake — reported free cash flow of $795.7 million (a 27% margin) is genuine cash [24]. The point is that free cash flow adds back the $490.6 million of SBC, so a conservative "owner earnings" figure that treats stock comp as the real cost it is lands closer to $300–350 million. An investor who values TTD on unadjusted FCF and one who values it on SBC-adjusted earnings are underwriting two very different multiples — a distinction the valuation section returns to.

5. Capital allocation: the buyback treadmill

The company pays no dividend and reinvests almost nothing in fixed assets (capex was 6.8% of revenue even after a FY2025 step-up to build capacity), so the entire capital-allocation story is share repurchases [25]. In FY2025 it bought back and retired 26.2 million shares for $1.4 billion — and management is explicit that the program is "designed to help offset the impact of future share dilution from employee stock issuances" [26].

That framing matters. Despite spending $1.4 billion — nearly triple the year's SBC expense — the diluted share count fell only modestly, from 501.9 million to 493.6 million [27]. Much of the buyback is a treadmill: it mops up the dilution that SBC creates rather than compounding per-share ownership. The silver lining in 2026 is opportunism — buying back stock at $19.53 that traded above $100 in 2024 retires far more shares per dollar, so the depressed price makes the treadmill unusually accretive if the business is worth anything like its history suggests.

Source: repurchases and stock-based compensation per FY2025 10-K Consolidated Statements of Cash Flows [28].

One consequence worth flagging: cumulative buybacks now exceed cumulative retained profit, so the balance sheet carries a small accumulated deficit — retained earnings swung to roughly negative $591 million at year-end 2025 [12]. That is a bookkeeping artifact of returning capital, not distress — but it is why equity shrank year over year despite positive earnings.

6. The moat, pinned to mechanism and evidence

Moat claims are only worth as much as the mechanism behind them. Here is the honest scorecard.

Source: retention and philosophy [29]; data marketplace of over 370 vendors [30]; UID2/EUID identity strategy [31].

The data network effect deserves emphasis because it is the least obvious: a two-sided marketplace of over 370 third-party data vendors sits on the platform, and each additional buyer makes the platform a more attractive distribution channel for data, which in turn makes it more useful to the next buyer [32]. Its neutrality — buy-side only, owning no media — is a genuine differentiator against Google and Amazon, but it is the moat most exposed to erosion as Amazon leans into its own DSP.

Where the moat is not proven: Unified ID 2.0. The company is trying to make UID2 (and its European variant EUID) the open internet's replacement for the third-party cookie, alongside OpenPath, which routes buyers more directly to publisher inventory [33]. If UID2 becomes the industry standard it would deepen the moat enormously; today it is optionality, not a demonstrated advantage — and Google's 2025 decision to keep cookies in Chrome removed the near-term catalyst that would have forced adoption.

7. Competitive reality: the largest independent, dwarfed by the walled gardens

The competitive map has two tiers. Against the walled gardens the scale asymmetry is staggering — Google's advertising revenue alone is roughly twenty times TTD's entire spend — and the filing openly concedes the industry is "highly competitive and fragmented," naming Google and Amazon as its most formidable competitors [34]. TTD's counter is that it aggregates the rest of retail against Amazon's first-party garden: on its Q1 2026 call management argued the retailers in its data marketplace represent over 80% of the sales of top U.S. retailers, versus Amazon's own share of under 15% of U.S. retail spend [35].

Against the independents, TTD is not merely the largest — it is the only one that earns a serious margin, which is the clearest evidence that buy-side scale compounds into superior economics.

Source: TTD revenue, operating income and net income per FY2025 10-K [36]; Magnite income statement [37]; Viant income statement [38]; PubMatic income statement [39].

The read-through: TTD's ~20% operating margin dwarfs Magnite (14%), Viant (4%) and PubMatic (a loss) — and it does so at roughly four times the revenue of the next-largest independent. The genuine competitive threat is not on this table; it is Amazon's DSP, the only rival with the data, the scale and the retail-media flywheel to contest buy-side share directly. Share loss to Amazon is the core bear risk, and no amount of margin superiority over the small independents offsets it.

The growth engine underneath all of this is connected TV: video, which includes CTV, has grown to a high-40s percentage share of platform spend and continues to rise as the mix shifts [40]. The migration of TV budgets from linear to biddable streaming is the reason an 18%-growth business can still credibly talk about a long runway — and it is heavily U.S.-weighted today (the United States was 85% of FY2025 revenue), leaving international as the second leg.

8. Cyclicality and the self-inflicted 2024 stumble

Advertising is cyclical — budgets flex with the economy, the fourth quarter is the seasonal high and the first quarter the low, with extra lift in U.S. election years [41]. But the defining recent event was not macro. In Q4 2024, after eight years as a public company, TTD missed its own guidance for the first time, with the CEO calling it plainly "our fault" and attributing it to a slower-than-planned rollout of Kokai — its "most recent and major upgrade" to the platform — alongside an internal reorganization [42] [43].

9. Governance and alignment — an underwriting factor, not a footnote

Two governance facts change how an investor should size position and risk.

First, control is entrenched. The dual-class structure gives Class B shares ten votes each versus one for the Class A that public investors own, so insiders — led by founder-CEO Jeff Green — hold effective voting control regardless of their economic stake [44]. Public shareholders are, in practice, along for the ride on major decisions.

Second, the 2021 CEO Performance Option is one of the largest executive grants in corporate history and its incentives are now badly out of the money. It entitles the CEO to buy up to 16 million shares (a maximum of 19.2 million with a relative-TSR modifier) at a $68.29 exercise price, vesting only if the stock hits price hurdles ranging from $90 to $340 per share — against a $819 million grant-date fair value [45]. With the stock at $19.53, every tranche is deeply underwater and the first $90 hurdle is more than 4x away. The grant has also drawn shareholder litigation [46]. The alignment cuts both ways: the CEO makes nothing on the option unless the stock roughly quintuples, but the company keeps expensing it, and the size of the grant is a real governance red flag.

10. How to value it — the entire story is the de-rating

This is the section that matters. The Trade Desk was, for years, one of the most expensively valued stocks in software — the market paid 38 times sales at the end of 2021 and still 24 times sales at the end of 2024. By mid-2026 that had collapsed to 3.3 times sales. Nothing in the business broke commensurately: revenue still grew 18% and free-cash-flow margins held at 27%. What changed was the multiple.

Source: derived from year-end share prices (price feed) and reported annual revenue; 2026 point uses the $19.53 close on 10-Jul-2026 and FY2025 revenue [47]. Market-cap inputs from the numeric price feed, as reported.

At the $19.53 price the business carries a ~$9.6 billion market cap and, because it holds ~$1.3 billion of net cash and no debt, an ~$8.3 billion enterprise value. The valuation looks cheap on almost every conventional lens — and demanding on the one lens that treats stock comp honestly.

Source: multiples computed on a $9.64B market cap and ~$8.3B enterprise value at the $19.53 close; earnings, FCF, Adjusted EBITDA and SBC per FY2025 10-K [48] [49] [50].

The right lens. This is a capital-light, high-return compounder, so the correct frame is price-to-free-cash-flow and FCF yield, adjusted for stock-based compensation — not EBITDA (which flatters by adding back the SBC) and not revenue multiples (which tell you sentiment, not value). Read that way, the two poles of the debate are clear:

• The bull case underwrites the 8.3% reported FCF yield on a business still growing 18% with a rising take rate and a CTV tailwind — an extraordinary combination that has almost never been available this cheaply in TTD's history.
• The bear case notes that once you charge shareholders for the $490.6 million of annual dilution, the honest yield is closer to 3.2%, growth is decelerating, and Amazon's DSP is the one competitor that could compress the take rate the whole thesis rests on [51].

The decisive variables are the same three to watch every quarter: the take rate (is monetization still rising or is it rolling over) [52], CTV mix (the growth engine) [53], and share versus Amazon [54]. A high-quality business does not automatically make a good stock, but at 3.3 times sales the price has already done most of the work of pricing in the risks — which is precisely why the take rate is now the only number that decides the outcome.

---

Long-Term Thesis — The Trade Desk, Inc. (TTD)

The underwriting question is not "is TTD cheap?" — after an ~86% de-rating it plainly is. The question is whether, five-to-ten years out, this is still the neutral operating system for the open internet's ad spend, charging a fee that holds or rises on a spend base several times larger than today. Everything durable about the thesis reduces to one identity: reported revenue is a take rate on a much larger river of gross spend — $13.4 billion of spend became $2.9 billion of revenue in FY2025, at an implied take rate of 21.6% [1][2]. To win over a decade, two things must compound together: the spend that flows across the platform, and the fee TTD keeps on it. The company owns the first through the two most durable secular currents in advertising — programmatic and connected TV — and defends the second with a retention rate that has exceeded 95% for over a decade [3]. The bear case is that both legs are quietly weakening at once — growth has fallen nine straight quarters and the take rate is undefended against Amazon — while a founder-controlled board sheds its independent oversight. This tab frames the durable bet, states what must be true, and names the multi-year signals that would prove it working or breaking.

1. What has to be true — the five load-bearing propositions

A 5-to-10-year thesis is only as strong as its weakest structural assumption. The table below decomposes the bet into the five propositions that must each hold, the evidence that they are holding today, and the single signal that would first reveal each one breaking. The order is deliberate: proposition 2 (the take rate) is the master variable — the others feed it or defend it.

Sources: take rate and margin per FY2025 10-K MD&A [4]; retention [5]; TAM and CTV shift [6]; CTV share [7]; retail-data marketplace [8]; revenue-versus-spend risk [9].

2. The reinvestment runway — small fish, enormous pond

The most durable thing in the whole file is the size of the pond relative to the fish. TTD routes $13.4 billion of gross spend across a digital-advertising market of over $700 billion — itself more than 70% of a total ad market that has surpassed $1 trillion [10]. Even measured on gross spend — not the thinner revenue line — the platform intermediates under 2% of digital ad dollars. This is the structural reason growth can decelerate for years and still leave a long runway: the constraint is share capture and monetization, not a saturated market.

Source: digital ad spend "over $700 billion" and the $1 trillion total market per FY2025 10-K, Our Industry [11]; gross spend $13,394.7M and revenue $2,896.3M per MD&A Executive Summary [12].

The runway has a specific shape, and it matters for how the next decade plays out. The engine is connected TV — video (which includes CTV) is already a high-40s percentage share of platform spend and rising, as a "generational shift from linear television to connected television" plays out [13][14]. Linear TV budgets moving to biddable streaming is exactly the kind of high-value, premium inventory that lifts both spend and take rate — and it is the position Amazon's owned-media conflict structurally cannot occupy for a Fortune-500 advertiser. The second leg is deeper monetization of existing spend — the FY2025 take-rate step-up came from "increased utilization of our value-added services and data; and higher platform fees," not just routing more dollars [15]. International (the U.K., Germany, France, Japan, India, Australia) and the retail-data marketplace are the third leg. The point for an underwriter: the reinvestment runway is high, but it is a monetization and share runway, not a market-creation one — the market already exists; the question is TTD's slice of it.

3. The compounding math — where the take rate and spend meet

Because revenue = take rate × gross spend, a decade of value creation is a two-variable problem, and small differences in either variable compound into very different outcomes. The scenarios below are illustrative underwriting paths, not forecasts — they hold the structure fixed (a fee on a spend base) and vary the two things that matter. The base case assumes gross spend compounds at low-teens as CTV migration continues and the take rate holds near 21–22%; the bull case adds re-acceleration and continued take-rate lift; the bear case is the structural thesis — spend still grows but the take rate erodes as Amazon contests price.

Source: illustrative scenario paths constructed by this analyst; anchored to FY2025 revenue of $2,896.3M and the take-rate / gross-spend structure per FY2025 10-K MD&A [16][17]. Not a forecast.

The spread between the paths is almost entirely a take-rate story, which is why proposition 2 dominates the frame. Management's own risk factor is the tell that this is the live variable, not a settled one: revenue "may not necessarily grow at the same rate as spend on our platform" as pricing competition, volume discounts and channel mix bite [18]. The bull path is not a fantasy — it is roughly what the last decade delivered — but it requires the take rate to keep rising into a market where the best-capitalized rival on earth is contesting price. The bear path still grows revenue; it simply earns a value multiple rather than a growth one. An underwriter should size positions to the reality that the decade's outcome is set by a single line item the company itself flags as uncertain.

4. The moat that must survive a decade — durable, but narrow by construction

The moat is what defends the take rate, so its width is the thesis. The strongest evidence it is real is not an adjective but a multi-year fact: retention has held above 95% through two opposite shocks — the 2022 ad recession and the self-inflicted 2024 Kokai-migration miss the CEO called "our fault" [19]. A retention floor that survives both a demand collapse and a product failure is telling you the advantage is structural, not cyclical.

Source: client-retention disclosures — FY2021 "over 95% in 2021, 2020 and 2019" [20] and FY2025 "exceeded 95% for over a decade" [21]; the Kokai miss per the Q4 FY2024 call [22].

But underwrite the moat's width honestly, because it is where the thesis is most fragile. The lock-in is behavioral, not contractual: master service agreements are terminable on 60 days' notice with no spend minimums, and management concedes there is "limited cost and difficulty" for a client to move its spend to a competitor [23]. The reconciliation of "cheap to leave" with "95% stay" is the moat — buyers wire first-party data, measurement and optimization logic into the platform and the switching cost is the disruption of re-plumbing it — but it is a moat a sufficiently better or cheaper rival could erode. Two structural pressure points compound the fragility over a decade: client concentration (two agency holding companies each represented more than 10% of gross billings in 2025, each with take-rate leverage) [24], and the neutrality edge narrowing as Amazon leans into its own DSP. The one genuinely compounding, countable advantage is the two-sided data marketplace, which has grown from roughly 200 integrated data vendors in FY2021 to more than 370 in FY2025 — a network a subscale rival cannot replicate quickly because the buyers and the vendors must show up first [25][26].

Source: "more than N third-party data vendors" disclosed in each Form 10-K — 200+ in FY2021 [27], 370+ in FY2025 [28]. Intermediate years are disclosed floors, so the chart understates the true count.

The two management responses worth tracking as moat-widening options over the decade — both still unproven — are Kokai, the AI-driven platform upgrade whose success would deepen optimization lock-in [29], and OpenPath, which routes buyers to a "simplified, direct connection to publishers," strengthening supply-side control [30]. Neither yet counts as moat; both are options on a wider one.

5. The economic engine funds itself — but read the cash honestly

A long-term thesis needs a balance sheet that lets the company self-fund through downturns without dilutive raises or debt maturities, and TTD has one. It carries zero debt and roughly $1.3 billion of net cash (cash plus short-term investments), and it converts earnings to cash at a high rate — FY2025 operating cash flow was $992.7 million against $443 million of net income [31][32]. That fortress is what lets management repurchase stock through the drawdown — 26.2 million Class A shares retired in FY2025 — at prices where each dollar buys more than twice the shares it did at the FY2025 average [33].

The honest deduction an underwriter must carry for a decade: the headline 27% free-cash-flow margin is before charging shareholders for stock-based compensation. SBC of $490.6 million is added back inside the $795.7 million of FY2025 free cash flow — more than half of it [34]. The buyback has, to date, largely offset that dilution rather than shrinking the share count. So the decade-long test on capital allocation is narrow and specific: does the buyback, funded by genuine owner cash, actually reduce shares outstanding over time, or does it merely run on a dilution treadmill? On real, SBC-charged cash, the business is far less cheap than the tape's ~12x reported FCF advertises — but it is still self-funding, which is what the balance-sheet leg of the thesis requires.

Source: derived from reported FY2022–FY2025 cash-flow statements; FY2025 no-debt position and 26.2 million shares repurchased per FY2025 10-K Liquidity [35]; SBC add-back per Consolidated Statements of Cash Flows [36].

6. The structural discount — governance is a permanent feature of this bet

Unlike the take-rate question, which the next few years will resolve, the governance overhang is a standing feature an owner must accept for the full holding period — and it is deteriorating, not improving. This is a founder-controlled company: insiders held approximately 49.9% of voting power at year-end 2025, via a dual-class structure whose Class B super-vote does not convert to Class A until December 22, 2035 [37]. The control lasts, by design, roughly as long as this very underwriting horizon. That concentration is not automatically bad — a founder who bought ~$150M of stock personally at the lows is aligned with minority holders on price — but it means minority investors are passengers, and the two live negatives compound the risk.

The first is oversight erosion at the worst possible moment: independent board control visibly thinned around the 2026 annual meeting, leaving audit-committee independence in question exactly as the growth story came under scrutiny [38]. The same people who define the flattering non-GAAP metrics control the votes. The second is a live legal tail: TTD's own FY2025 10-K discloses a securities complaint asserting a Section 20A claim that the CEO, then-CFO and Chief Strategy Officer "engaged in insider trading during the proposed class period" [39] — a claim that survived a motion to dismiss and is now in discovery. And the incentive backdrop is aggressive: the FY2021 CEO Performance Option granted the founder up to 16 million shares vesting on stock-price milestones running to $340, a grant valued at roughly $819 million [40]. For a decade-long holder, the governance dial is the reason this is a narrow, condition-gated long rather than a clean compounder: it lowers the trust an investor can place in the very metrics the bull case leans on.

7. The four dials, and what to watch for a decade

Netting the frame: the driver that makes this a superior investment is deep monetization of a still-tiny share of a $700B-plus market as CTV migrates to biddable streaming; the failure mode that breaks it is structural take-rate erosion as Amazon contests price on inventory TTD cannot access. The evidence sits in the middle — genuinely unresolved — which is why the honest four-dial read pairs a high runway and real durability with only moderate evidence confidence.

Source: analyst synthesis of the cited primary record — TAM and penetration [41][42]; retention [43].

The signals that separate long-term thesis evidence from short-term noise, ordered by how much they would move the decade view:

Bottom line for the underwriter. The durable frame is a conditional long. TTD owns two of advertising's most durable secular currents, defends a real (if narrow) moat with a decade of 95%+ retention, self-funds from a net-cash balance sheet, and trades at a value multiple on less than 2% penetration of its market — the ingredients of a genuine multi-year compounder. What keeps it off an unconditional long is that the single line that governs the decade — the take rate — is the very one management flags as uncertain [44], the rival contesting it is the best-capitalized company on earth, and a founder-controlled board is shedding oversight while a Section 20A claim proceeds. Underwrite the franchise; gate the position on revenue re-accelerating with the take rate intact; and treat the governance dial as the permanent discount it is.

---

Competition — The Trade Desk, Inc. (TTD)

The Trade Desk runs the largest independent, buy-side-only demand-side platform (DSP) — it plans, buys and measures programmatic advertising for agencies and brands across connected TV (CTV), video, display, audio and native, and, unlike its two biggest rivals, it does not own the media it buys against. That single design choice — objectivity — is the spine of the whole competitive case, and management leans on it as its core differentiator against "divisions of large, well-established companies such as Google and Amazon" [1].

This tab takes a side on whether that moat is real, names the one rival that matters, benchmarks the peer set with receipts, and ends with the signals that would change the call.

Bottom line

The arena and why these five peers

Digital advertising is the largest and fastest-growing slice of global ad spend — over $700 billion, more than 70% of the total market — split into buyers, sellers and the marketplaces between them [2]. The Trade Desk sits permanently on the buy side. That fixes who its true comparators are, and it is not a generic "big tech" list. The peer set falls into three tiers, each confirmed from the peer's own filing:

• Walled-garden buy-side rivals — Alphabet/Google and Amazon. Both are named by The Trade Desk as competitors [3]. Google's customers "purchase advertising inventory through … Google Display and Video 360" — its DSP [4]. Amazon sells advertising "through programs such as sponsored ads, display, and video advertising" — Amazon DSP [5]. Both also own media, so they are simultaneously suppliers and rivals.
• The pure-play independent DSP — Viant Technology (DSP), whose "cloud-based demand side platform … enables the programmatic purchase of advertising" — the identical business model and revenue driver to The Trade Desk [6]. It is the single cleanest public comparable.
• Independent sell-side counterparts — Magnite (MGNI), "the world's largest independent omni-channel sell-side advertising platform" and "largest independent programmatic CTV marketplace" [7], and PubMatic (PUBM), an "independent sell-side platform" [8]. These are the other side of the same programmatic transaction — ecosystem peers routinely traded alongside The Trade Desk, not same-side substitutes.

That the peers name each other is the best evidence the set is right: Viant lists its competitors as "public companies exclusively serving our industry, such as The Trade Desk, and … divisions of large, well-established public companies such as Google and Amazon" [9], and PubMatic tells investors that "The Trade Desk and Google DV360 in particular … account for a significant portion of the ad impressions purchased on our platform" [10].

One genuine peer sits out of the benchmark: Criteo (CRTO), an independent commerce-media / performance demand platform that competes for the same advertiser budgets. The run's staged data for Criteo resolved to the wrong entity (a French regional bank), and no Criteo annual report is indexed, so its market cap and enterprise value cannot be sourced reliably. It is carried below as N/A with that reason rather than benchmarked on bad data or silently dropped.

Peer comparison — with receipts

Sources: market caps from the run's staged pricing snapshots (as of 2026-07-11); EV derived from reported cash and debt; ad-related revenue and margins from FY2025 filings — Alphabet advertising (Search + YouTube + Network) [11], Amazon advertising services revenue of $68.6 billion [12], Magnite [13], Viant [14], PubMatic [15]. EV/Rev and Op Margin are left blank for Google and Amazon because their enterprise value spans whole conglomerates and is not comparable to an advertising-unit multiple.

For Google and Amazon the "Ad Revenue" column is the advertising business, not the whole company: Google generated over 70% of its revenue from online advertising [16], and Amazon's ad business reached $68.6 billion in FY2025, up 22% [17]. Against those, The Trade Desk's $2.9 billion is a rounding error on total advertiser spend — the runway is enormous (management pegs its own share at "less than 2% of the global advertising TAM" [18]) — but it also means the two giants can fund the fight indefinitely.

The four independents map cleanly on growth-versus-margin, with bubble size showing enterprise value. The Trade Desk is the only name in the top-right quadrant: fast growth and a high-teens/twenties operating margin at multiples of peers' scale.

Source: FY2025 filings and derived enterprise values — Viant [19], Magnite [20], PubMatic [21]; Google and Amazon excluded as off-scale conglomerates.

Where The Trade Desk wins

1. Objectivity it can prove, not just assert. The Trade Desk is "free from the conflicts of interest inherent in our competitors that also own and operate media" [22], and it turns that into measurable performance. In a head-to-head test run by "one of the world's leading appliance manufacturers" against Amazon DSP on CTV, The Trade Desk "reach[ed] 70% more unique households … at 30% lower total cost" and "performed six times better in terms of delivering their campaign goals" [23]. Management's framing of the rivals is blunt: Amazon "is mostly playing in selling their owned and operated inventory … if they say they're objectively trying to buy the open Internet and then spend most of the money on their owned and operated … it imposes some hypocrisy or creates some channel conflict" [24], and "DV360 is primarily technology to buy YouTube [while] Amazon's DSP is primarily a product built to buy Amazon's Prime Video" [25].

2. Client stickiness few software businesses match. Revenue comes from ongoing master service agreements, not episodic orders, producing "a customer retention rate that has exceeded 95% for over a decade" [26]. That lock-in is deepening as spend consolidates into Joint Business Plans, which by end-2025 "accounted for well over half of our business" [27]. Neither sell-side peer, nor Viant, can point to comparable buy-side switching costs.

3. A performance engine and identity standard rivals lack. The platform's bid-factor architecture lets buyers express campaigns across "quadrillions of permutations," which management calls expressiveness [28]. Its latest platform, Kokai, has delivered "on average, 26% better cost per acquisition, 58% better cost per unique reach, and a 94% better click-through rate compared to Solimar" [29]. And its Unified ID 2.0 — an open-source identity framework that turns emails and phone numbers into a privacy-safe identifier [30] — has, on management's account, become "the primary identity currency for ads in the open Internet around the world" [31]. This is the closest thing the open web has to a Google/Amazon-independent targeting rail.

4. The premium-CTV franchise. The most-watched streaming inventory transacts through The Trade Desk: "Partners like Disney, NBCU, Walmart, Roku, LG, Netflix, and many others are deepening their relationships with us around the growing … programmatic opportunity in CTV" [32]. Magnite is the sell-side leader in the same channel [33], but as the seller's agent — a partner to The Trade Desk on the buy side, not a substitute for it.

5. It is gaining share on the open web while Google's open-web engine shrinks. Google's third-party "Network" ad business — the piece that competes most directly with The Trade Desk's turf — has declined every year since 2022, even as The Trade Desk's revenue has grown over 80% cumulatively. The objectivity advantage is compounding: in 2025 supply of ad impressions "meaningfully outpaced demand," which The Trade Desk calls "a positive development … [that] shifts the balance of power to the buy-side" [34].

Source: The Trade Desk revenue from FY2025 10-K MD and A [35]; Google Network revenue as reported in Alphabet FY2025 segment data [36]. Indexed by this analyst from reported figures.

Where competitors are better

1. Amazon's data and inventory advantage is structural, not a technology gap. Amazon couples first-party purchase data with owned Prime Video and retail-media inventory. The Trade Desk itself warns that "walled garden inventory providers … may exclusively sell their own inventory directly to advertisers, which prevents us from competing with them entirely for such inventory," and that such rivals have "significantly more financial, technical, marketing, and other resources … [and] rich first party data sets" [37]. Amazon's 22% ad growth off a $68.6 billion base [38] is being funded by a business The Trade Desk cannot touch.

2. Google's distribution and scale dwarf everything independent. Even a shrinking Google Network is a roughly $30 billion open-web ad business [39], and DV360's tie to YouTube gives it default access to the single largest premium video destination [40]. The Trade Desk's own risk language concedes rivals may be "better positioned to execute … over certain channels, such as social media, mobile, and video" [41].

3. Price is a real vector against the mid-market. Competitors "differentiate … primarily on the basis of artificially low prices" and "some customers may be price sensitive"; The Trade Desk admits there is "no guarantee that new and existing customers will value our offerings as we intend" [42]. Viant (DSP) is growing 19% — matching The Trade Desk's pace — precisely by serving that price-conscious segment with a comparable cloud DSP [43].

4. Execution risk is self-inflicted and recent. For the first time in its public life, The Trade Desk missed its own forecast in Q4 2024. The CFO owned it — "this miss was not due to lack of opportunity or increased competition, it was on us" [44] — after "the largest reorganization in company history" [45] and a forced migration to Kokai, "the most significant company upgrade in our 16 years" [46]. A stumble competitors did not cause is still a stumble competitors can exploit.

Threat map

The threats that matter are not evenly weighted. Amazon is the one to watch; Google is the incumbent whose open-web pressure is actually easing under antitrust and its own cookie reversal; the SSPs and Viant are lower-order.

Sources: Amazon ad scale [47]; walled-garden and pricing risk from TTD FY2025 10-K [48]; Viant competitive positioning [49]; execution miss [50]; Google cookie reversal [51].

Why Amazon over Google. Two structural shifts have reduced Google's open-web threat. On April 22, 2025 Google abandoned its multi-year effort to deprecate third-party cookies in Chrome [52] — removing the biggest overhang on the open-web ad economy The Trade Desk serves — and a U.S. federal court ruled Google's "publisher tools unfairly excluded rivals" [53], an antitrust outcome that, if it forces separation of Google's buy- and sell-side, hands independent DSPs a tailwind. Amazon faces no such constraint and is growing three times faster on the open web's fringes. Management agrees on the ranking — Amazon's "objectivity problem is way worse than Google's because Amazon competes with nearly every company in the Fortune 500" [54] — but "objectivity problem" is a reason advertisers should prefer The Trade Desk, not proof they will; Amazon's data and inventory can win budgets regardless.

The valuation reset — the misunderstood part

The most striking competitive fact is not in any filing's competition section — it is in the tape. After peaking near $139, The Trade Desk shares changed hands at $19.53 on 2026-07-10, an ~86% drawdown that has compressed the category leader to a ~$9.6 billion market cap and roughly $8.3 billion enterprise value — about 2.9x its ad revenue. That is a lower EV/revenue multiple than Magnite (3.8x), a slower-growing (7%) sell-side peer, and barely above loss-making PubMatic and single-digit-margin Viant.

Source: EV/revenue derived from staged pricing snapshots (2026-07-11) and FY2025 reported revenue — Magnite [55], Viant [56], PubMatic [57].

The bear reads that discount as the market pricing in the walled gardens winning and take-rate compression. Management's counter is that "for as long as we've been public … there's been a narrative that our … take rate must compress … [but] those business models deliver less value overall" [58], and that rising complexity is "more of a moat than a threat … we're often competing with companies whose DSP is their 37th highest priority, not their #1" [59]. The competitive read: the moat evidence (95%+ retention, UID2 ubiquity, head-to-head CTV wins, gaining open-web share) has not deteriorated to match the multiple — which is what makes the position misunderstood rather than impaired, provided execution holds.

How The Trade Desk manages the sell side

The Trade Desk's relationship with Magnite and PubMatic is deliberately kept as partner, not rival: "we will always only represent the buy side … Everything we do that interacts with the supply side, such as OpenPath and OpenAds, is intended to drive better signal and a more transparent marketplace for our advertiser and agency clients" [60]. But OpenPath — a direct pipe to publishers — is explicitly a check on the SSPs: management calls it "a canary in the coal mine … keeping exchanges and SSPs in check" [61]. PubMatic already routes a "significant portion" of its impressions to The Trade Desk [62] and is integrating with The Trade Desk's Deal Desk [63]: the buy-side platform holds the power in that relationship, which is why the SSPs rank only a Low threat.

Moat watchpoints

Five measurable signals that would tell an investor whether the position is strengthening or cracking:

1. Amazon DSP win-rate on open-web / CTV. Watch for third-party or client-disclosed head-to-heads that contradict the appliance-maker result (70% more reach, 30% lower cost, 6x performance) [64]. If Amazon starts winning objective CTV bake-offs, the core thesis breaks.
2. Client retention holding above 95%. The decade-plus streak [65] is the single cleanest moat gauge; any dip signals the switching costs are eroding.
3. Take rate / gross margin. Gross margin has already drifted from ~82% (FY2022) to ~79% (FY2025); watch whether the "take rate must compress" narrative [66] starts showing up in the numbers.
4. UID2 adoption breadth. Its status as "the primary identity currency for … the open Internet" [67] is a network-effect moat only while adoption widens; a stall — or a walled garden pushing a rival ID — is a warning.
5. Google antitrust remedy and open-web share. A structural break-up of Google's ad stack [68] plus continued decline in Google's Network revenue [69] would confirm share shifting to independents; a reversal would not.

---

Current Setup & Catalysts — The Trade Desk, Inc. (TTD)

The one-line read. After an ~86% drawdown, TTD is a genuinely cheap, net-cash, GAAP-profitable franchise trading on ~3.3x sales and ~11x forward adjusted EPS — but the entire debate now hangs on a single unresolved question the tape has not settled: is the fall from high-teens to ~12% revenue growth a self-inflicted Kokai-migration stumble that reverses, or Amazon-driven structural take-rate erosion that does not? Everything on this page is the bridge between that durable 5-to-10-year question and the near-term evidence path that will start to answer it. The single most decision-relevant event is the Q2 FY2026 print and the Q3 guide it carries, on August 6, 2026 — not because one quarter decides a decade, but because it is the next clean read on the one variable (revenue versus gross spend — the take-rate direction) that governs the whole thesis.

This is not a binary or distressed name. There is no maturity wall, no covenant, no going-concern question — zero debt and ~$1.3B net cash [7]. So no single print "decides the case." What the prints do is move the probability the market assigns to cyclical-versus-structural, and that probability is what the ~3x-sales multiple is arguing about.

The variant view, sized. On direction we are consensus-aligned — this is a watchlist name, cheap but growth-gated, and the Street's "Hold" (mean target $24.42 vs $19.53, ~25% upside) roughly captures that. Where we differ is on the near-term numbers and the skew:

• Modestly below Street on FY2026 revenue. Consensus FY2026 revenue of ~$3.18B (+9.7%) embeds an H2 re-acceleration: H1 lands near $1.44B (Q1 actual $689M plus a Q2 guide of "at least $750 million" [1]), which requires H2 to grow ~9% after Q2's guided ~8%. Given nine straight quarters of deceleration and management's own CPG/auto/tariff caveats, we model FY2026 revenue nearer +7–8% (~$3.10B), roughly 1–3% below consensus, and treat the Aug-6 Q3 guide as the number that resolves it.
• The skew into Aug 6 has flipped asymmetric to the upside. The market's estimate revisions are overwhelmingly negative (18–20 downward EPS cuts versus 0–2 up in the last 30 days; FY2026 adj. EPS trimmed from ~$2.07 to ~$1.85 over 90 days), and — critically — the earnings-day gap-downs are shrinking (−33% and −39% in early/mid-2025 to just −4.8% and −1.7% on the last two prints) even as growth kept sliding. Bad news is increasingly in the price. A stabilization or a take-rate-intact Q3 guide could re-rate hard (the historical up-gap base rate is +12% to +19%); a fresh sub-8% signal likely gaps down, but by less than history (est. −8% to −15%), not another −35%.

Sources: Q2 guidance and Q1 results per Q1 FY2026 earnings call [1]; consensus, target, and revision figures from the analyst-estimates and earnings-calendar data feeds (as reported); price-reaction base rate derived from the staged daily price series.

Where we are right now

Source: share price and consensus mean target from the market-data / analyst-estimates feed (as of 2026-07-11, ~36 analysts); next-earnings date from the earnings-calendar feed (2026-08-06). Company filings, as reported.

The stock crashed in three discrete guidance shocks, not on a reported-earnings collapse: −41% in February 2025 (the first revenue-guidance miss in 33 quarters [13]), −39% in August 2025, and a long grind lower through FY2025 results into mid-2026. Through all of it the business kept compounding: FY2025 revenue +18% to $2.90B, with a rising take rate of 21.6% (from 20.3%) as $13.4B of gross spend it never owns was recognized net into revenue [4][3]. This is a multiple de-rating, not an earnings impairment — which is exactly why the forward evidence path, not the backward tape, is what matters now.

What actually changed in the last 3–6 months

Sources: Q1 results and CSO departure per Q1 FY2026 call [1][2]; Section 20A claim and client concentration per FY2025 10-K [8][11]; one-member audit committee per 2026 proxy [10]; Publicis, CFO turnover, Nasdaq cure clock, CEO purchase, and Koa Agents from recent press (Ad Age, Adweek, StockTitan), as reported.

The historical earnings price-reaction base rate

Every "high impact" claim below is anchored here — how TTD actually moves on a print. Two facts dominate. First, the EPS surprise barely predicts the reaction — the moves are driven by revenue and the forward guide, not the headline beat/miss (Feb 2025 beat EPS and fell 33%; Aug 2025 was in-line on EPS and fell 39%). Second, the down-moves are shrinking — the last two prints moved just −4.8% and −1.7% despite continued deceleration, the signature of exhausted positioning. Average absolute earnings-day move over the last eight quarters: ~15%.

Source: EPS surprises from the earnings-calendar feed; next-day price moves derived from the staged daily price series (data/prices/) — not filing figures.

The live debate — what the market is watching now

Sources: take rate and value-added-service driver per FY2025 10-K MD&A [3]; revenue-may-not-track-spend risk [5]; CTV low-50s mix [1]; SBC add-back [6]; voting control [9]; one-member audit committee [10].

The near-term growth path is the spine of the whole debate. Consensus already models Q2 down to ~+8% and needs H2 to firm back to ~+9% to hit the full-year number — the Aug-6 Q3 guide is the first test of that bet.

Sources: quarterly YoY growth from company results (as reported); Q2 FY2026 point is the company's "at least $750 million" guide (~+8% vs the year-ago $694M) per the Q1 FY2026 call [1].

Ranked catalyst timeline

Ranked by decision value to an institutional investor — not by date. The most thesis-resolving event sits at the top whether it lands in weeks or quarters. Positioning context for the high-impact rows: no official short interest is reported for TTD, but the tape trades crowded-negative-momentum (deep ~19M-share ADV, estimates cut hard into the print), so a surprise in either direction is amplified — a short here carries real squeeze-into-a-beat risk given the +12% to +19% historical up-gaps.

Sources: Q2 guidance ("at least $750 million"), $164M Q1 buyback, and the CSO/OpenAI departure per the Q1 FY2026 call [1][2]; Section 20A claim per FY2025 10-K Note 13 [8]; one-member audit committee per 2026 proxy [10]; FY2025 buyback / zero debt per 10-K MD&A [7]; Q2/FY2026 consensus and next-earnings date from the analyst-estimates and earnings-calendar feeds; Nasdaq cure date, Google-remedy status, and Publicis settlement from recent press (StockTitan, AdExchanger, Ad Age), as reported — timing only, not a filing fact.

Impact view — what resolves the debate vs what only adds information

Only two of these actually close underwriting questions on a durable variable; the rest add information or move a quarter. Note the honest split: the events that resolve the near-term evidence path (the prints) are not the same as the events that resolve the durable thesis-breaker (governance/litigation), which no single quarter settles.

Sources: linked-thesis reads synthesized from the Bull, Bear, Long-Term Thesis and Forensic tabs; underlying facts cited elsewhere on this page — take rate [3], Section 20A [8], walled-garden risk [12].

The next 90 days

The window through mid-October is dominated by one hard high-impact date and one governance deadline; everything else is a soft window or continuous watchpoint. The first real catalyst is Aug 6 — there is no meaningful pre-earnings datable event.

Sources: earnings date from the earnings-calendar feed (2026-08-06); Nasdaq cure date, Google-remedy status, and litigation status from recent press, as reported; gross-margin/share-count context from the Q1 FY2026 call [1].

What would change the view

Over the next ~6 months, three observable signals would most move the underwriting debate — tied to the durable thesis, not to Stan's final verdict:

1. The take-rate direction (revenue vs gross spend), read across the Aug and Nov prints. Two quarters of revenue outgrowing gross spend with a take rate holding above ~22% would tilt the case Long — it is the direct evidence that the deceleration is cyclical, not structural [3]. Two quarters of revenue growing slower than gross spend — the failure mode management itself flags [5] — would confirm the bear and justify the de-rating.

2. The Q3 guide trajectory (given Aug 6). Flat-to-up sequential guidance is trough confirmation; a guide below ~8% extends the nine-quarter slope and reprices the name as a high-single-digit grower. This is the near-term evidence marker, and it is cleaner than any single reported number.

3. The governance/litigation path — the durable thesis-breaker. A clean Nasdaq cure with credible independents and a permanent CFO, plus a benign turn in the Section 20A case [8], would remove the standing discount on the ~49.9%-founder-controlled structure [9] that lowers trust in the very non-GAAP metrics the bull case leans on. Further oversight erosion or an adverse 20A step does the opposite — and unlike the growth question, no single print resolves it.

The upside wildcard sitting outside all three: a structural Google ad-tech remedy that widens the open-web pool — near-zero priced, and the one catalyst that could re-rate the multiple for reasons entirely outside TTD's own execution.

---

Bull and Bear

Verdict: Watchlist — the valuation is genuinely cheap, but the one variable the whole debate turns on (is the deceleration cyclical or structural?) is not resolvable from the current record, and governance is deteriorating in real time. After an ~86% drawdown, The Trade Desk trades near a value multiple — ~3x sales, ~12x reported free cash flow — for the only durably GAAP-profitable independent DSP, with net cash, zero debt, and 95%+ client retention. That is the bull's floor. Against it stands a nine-quarter, monotonic slide in revenue growth (28% to 12%), a "free cash flow" number that is more than half added-back stock compensation, and a board whose independent oversight visibly collapsed at the exact moment the metrics most need scrutiny. The single tension that decides everything is whether the drop from high-teens growth to 12% is a self-inflicted Kokai-migration stumble that reverses, or Amazon-driven structural take-rate erosion that does not — and the honest answer is that no print in the record yet settles it. The evidence that would change this call is concrete: two or three quarters of revenue re-accelerating toward the high teens with the take rate intact would tilt it Long; a print below ~10% growth, or revenue growing slower than gross spend, would tilt it toward Avoid.

Bull Case

The three sharpest points are the valuation-versus-franchise gap, the rising take rate, and the CTV-plus-retail-data growth engine. At $19.53 the market pays ~3.3x sales and ~12x free cash flow for a business that grew revenue 18% to $2.90 billion and generated a record $796 million of free cash flow — the economic engine being a percentage platform fee skimmed off a $13.4 billion river of gross spend the company never owns, recognized net [1]. The take rate — revenue divided by gross spend, the master variable — stepped up to 21.6% from 20.3%, driven by "increased utilization of our value-added services and data; and higher platform fees" [2], sitting on a base with retention above 95% for over a decade [3]. And the durable growth engine — TV budgets migrating to biddable streaming plus a retail-data marketplace covering over 80% of top-U.S.-retailer sales against Amazon's under-15% retail share [4], with connected TV already a high-40s percentage of the business [5] — is precisely the neutral, no-media-conflict position Amazon structurally cannot offer.

Sources: bull points sourced as cited above — FY2025 Annual Report (Form 10-K) MD&A [6][7], Item 1 Business [8]; Q1 FY2026 call [9] and Q2 FY2024 call [10].

Bull's price target is $40 over 12–18 months, method: a re-rate to ~22x FY2026E adjusted EPS of ~$1.85 (less than half TTD's five-year-average multiple), cross-checked at ~20x FY2026E free cash flow of ~$850M plus $1.3B net cash on ~490M shares. The disconfirming signal the bull itself names: the take rate rolling over — revenue growing slower than gross spend — for two consecutive quarters, or growth sliding into high-single digits, either of which confirms Amazon-driven structural share loss and breaks the case.

Bear Case

The three sharpest points are the deceleration, the quality of the cash flow, and the board. Revenue growth has fallen in a near-straight line for nine consecutive quarters, with Q1 FY2026 delivering $689 million at just 12% [11], and TTD's own filing concedes revenue "may not necessarily grow at the same rate as spend on our platform" [12]. The 27% FCF margin is largely non-cash: stock-based compensation of $490.6 million is added back inside the $796 million FCF [13], and the FY2025 operating-cash jump was flattered by a non-recurring ~$168M deferred-tax lift under the July-2025 OBBBA law that will not repeat [14]. And oversight has collapsed: all three audit-committee members resigned within two weeks before the 2026 annual meeting, leaving a one-member committee [15], a live class action asserts a Section 20A insider-trading claim against the CEO and CFO [16], and insiders control ~49.9% of the vote on ~11% of the economics [17].

Sources: bear points sourced as cited above — Q1 FY2026 call [18]; FY2025 10-K Risk Factors [19][20], Cash Flows [21], Tax MD&A [22] and Note 13 [23]; 2026 Proxy [24].

Bear's downside target is $14 over 12–18 months, method: an EV/sales de-rating to ~2.0x on FY2025 revenue of $2.9B (a decelerating 10–12% grower earns no premium to faster-growing Viant at 2.0x), cross-checked at ~20x SBC-adjusted "cash FCF" of ~$305M. The cover signal the bear itself names: revenue re-accelerating to the high teens with the take rate intact and a genuine reduction in share count from buybacks — proof the 12% was a trough and that FCF compounds per share rather than funding dilution.

The Real Debate

The two advocates do not argue past each other — they read the same three facts in opposite directions. The 12% Q1 FY2026 growth print [25], the $796M FCF built on $490.6M of added-back SBC [26], and the same 10-K that reports a rising take rate [27] while conceding clients face "limited cost and difficulty" in leaving [28] — each is a coin both sides flip their own way.

Sources: shared facts traced to the Q1 FY2026 call [29] and the FY2025 Annual Report (Form 10-K) — Cash Flows [30], MD&A Results of Operations [31] and Item 1A Risk Factors [32].

Verdict

Watchlist. On the current record the bear carries slightly more weight, for a simple reason of burden-of-proof: at ~3x sales the multiple still embeds a re-acceleration, so the bull needs something to happen, while the bear only needs the observed nine-quarter downtrend to continue — and nothing in the tape has yet broken that slope. The single most important tension is the first one: whether the fall from high-teens growth to 12% is a self-inflicted Kokai stumble that reverses or Amazon-driven structural take-rate erosion that does not, and the current evidence genuinely cannot settle it. The bull can still be right — this is the only durably GAAP-profitable independent DSP, net cash of $1.3B, zero debt, 95%+ retention, and a rising take rate are real, and a de-rating this violent has repeatedly overshot on quality franchises; if growth stabilizes, $19.53 will look like a gift. But two things keep this off a Long: the durable thesis-breaker is structural take-rate erosion plus a governance apparatus (one-member audit committee, three CFOs in a year, a live Section 20A claim) that is a present-tense negative independent of any single print and lowers trust in the very non-GAAP metrics the bull leans on; the near-term evidence marker is narrower and cleaner — the next two-to-three revenue prints. The verdict flips to Lean Long on revenue re-accelerating toward the high teens with the take rate intact (revenue outgrowing gross spend) and a genuine share-count reduction; it flips toward Avoid on a sub-10% print or two quarters of revenue growing slower than gross spend. Until one of those arrives, the right posture is to watch, not to own.

---

The Moat: Real, Proven — and Narrower Than the Bulls Think

The Trade Desk clears the bar most companies fail: it has a genuine, company-specific, multi-year-durable economic moat, not just an attractive industry. The evidence is not an adjective — it is a client-retention rate that has exceeded 95% for over a decade [1], an operating margin that dwarfs every independent rival, and a two-sided data marketplace that has compounded for years. But the same primary record caps how wide the moat is: management itself concedes there is "limited cost and difficulty" for a client to move its spend to a competitor [2], the neutrality advantage is eroding as Amazon leans into its own demand-side platform, and the one variable the whole thesis rests on — the take rate — is contestable. The honest verdict is a narrow moat: durable enough to protect returns through two separate shocks, not wide enough to be assumed permanent.

Source: retention "exceeded 95% for over a decade" and more than 370 integrated data vendors per FY2025 Form 10-K, Item 1 Business [3] [4]; evidence-strength and durability are this analyst's scores.

1. The one test that separates a moat from a good year: did it survive stress?

A single filing can show a business is profitable. Only the multi-year record can show a business is protected — that the advantage held when the environment turned against it. The Trade Desk has now been stress-tested twice, from opposite directions, and the retention number did not break either time.

The first shock was demand: the 2022 advertising slowdown, when revenue growth decelerated and GAAP operating margin collapsed to roughly 7%. The second was self-inflicted and structural: in Q4 2024, after 33 straight quarters of meeting its own guidance, the company missed for the first time — its CEO called it plainly "our fault," tied to a botched rollout of the Kokai platform upgrade and a large internal reorganization [5]. A platform migration going wrong is precisely the moment clients would defect if the moat were thin. They didn't.

Source: client-retention statements in each year's Form 10-K, Item 1 Business — FY2021 [6], FY2022 [7], FY2023 [8], FY2024 [9], FY2025 [10]; the Q4 2024 miss and its cause per the Q4 FY2024 earnings call [11].

Retention above 95% held through every one of those years. That is the strongest single piece of moat evidence in the entire record, and it is what elevates this from "good software business" to "moated software business." A retention floor that survives both a demand recession and a self-inflicted product failure is telling you the advantage is structural, not cyclical.

But read the mechanism honestly, because it is not what most investors assume. The stickiness is not contractual lock-in. The master services agreements have one-year terms that auto-renew, are terminable on 60 days' notice by either party, and carry no minimum-spend commitment [12]. Management goes further and states the quiet part in its own risk factors: clients "choose the amount they spend," relationships are non-exclusive, and "there is limited cost and difficulty to moving their media spend to our competitors" [13].

2. The moat, pinned to mechanism, evidence and durability

Every claimed advantage below is tied to a specific economic mechanism and a checkable number, and graded on whether it is company-specific, copyable, and battle-tested. Adjectives alone score nothing.

Source: retention and self-service model [14]; MSA terms and 370+ vendors [15]; 20% operating margin [16]; "limited cost and difficulty to move" [17]; competition vs Google and Amazon [18].

The rest of this tab defends the three "Proven" rows with the multi-year record, then explains why "neutrality" is downgraded to narrowing and why UID2 is optionality rather than a moat.

3. The data network effect — the least obvious advantage, and the most measurable

The moat investors underrate is the two-sided data marketplace. The Trade Desk was built "data-management platform first, before building our ad-buying technology" [19], and it describes itself as "an important distribution channel" for third-party data vendors [20]. The flywheel is classic: every additional buyer makes the platform a more valuable place for a data vendor to distribute, and every additional data vendor makes the platform more useful to the next buyer. Unlike retention (a floor that can only be re-affirmed), this one leaves a rising, countable trail across five annual reports.

Source: "more than N third-party data vendors" disclosed in each Form 10-K, Item 1 Business — 200+ in FY2021 [21], 250+ in FY2023 [22], 350+ in FY2024 [23], 370+ in FY2025 [24].

The vendor count nearly doubled from 2021 to 2025 while the platform kept the same take-rate discipline — the mark of a network that is deepening, not just growing. The caveat: these are disclosed floors ("more than"), so the chart understates the true count and the 2021–2022 flat reading is a disclosure artifact, not a stall. Still, the direction is unambiguous and it is the kind of advantage a subscale rival cannot replicate quickly, because it requires the buyers and the vendors to show up first.

4. Scale economics — the moat that shows up straight in the margin

If buy-side scale genuinely compounds into a cost-and-data edge, it should appear as a profitability gap that rivals cannot close. It does, starkly. Among independent ad-tech platforms, The Trade Desk is not merely the largest — it is the only one earning a serious margin, at roughly four to six times the revenue of the next independent.

Source: TTD operating income $589.3M on $2,896M revenue (20% margin) per FY2025 10-K MD&A [25]; Magnite income statement [26]; Viant income statement [27]; PubMatic income statement [28].

A ~20% operating margin against Magnite's 14%, Viant's 4% and PubMatic's loss is not a rounding difference — it is the signature of a fixed software cost spread over a far larger spend base, plus the pricing power that comes from being the default independent buying seat. This is company-specific (the peers run the same programmatic model and cannot match it) and it has been durable: the margin gap has persisted for years and widened as TTD scaled. The one honest deduction, carried over from the Financials tab, is that this margin is struck before the ~$491M of annual stock-based compensation that TTD adds back to its headline numbers — real, but it does not change the relative verdict, since it dwarfs the independents on cash economics too.

5. The neutrality moat — real, but this is where the wall is cracking

The Trade Desk's most-cited differentiator is buy-side neutrality: it owns no media and sells no inventory of its own, so it never bids against its own clients — unlike Google and Amazon, which are buyer, seller and referee at once. The company frames its whole competitive stance around this, and the industry is, in its own words, "highly competitive and fragmented," with its most formidable rivals being "divisions of large, well-established companies such as Google and Amazon" [29].

Neutrality is a genuine advantage, but I grade it narrowing, for three reasons pulled from the record:

• Structural dependence on a competitor. Google is simultaneously "one of our largest advertising inventory suppliers in addition to being one of our competitors" [30]. A moat that depends on buying supply from the rival it is trying to displace is, by construction, not a fortress.
• Amazon is the one rival that can contest the take rate. Amazon's DSP pairs first-party retail data with effectively unlimited capital. TTD's counter — that the retailers in its data marketplace represent "more than 80% of sales from top U.S. retailers," versus Amazon's own share of "less than 15% of U.S. retail spend" [31] — is a strong argument that TTD aggregates the rest of retail against Amazon's garden. But it is an argument that has to be won, quarter after quarter, and that is a different thing from a moat that makes the fight unnecessary.
• The pitch is trust, and trust is a softer barrier than switching cost. Management leans on the conflict narrative — that "Amazon, for instance, is soliciting ad budgets while competing against numerous Fortune 500 companies" [32]. That resonates with brands wary of handing data to a competitor, but "advertisers prefer a neutral party" is a preference, not a lock — it can be outbid by better performance.

The offsetting positive is the growth engine neutrality plugs into: connected TV. Video, which includes CTV, has grown to a "high 40s percentage share" of platform spend [33], and the streaming platforms that supply that inventory are themselves wary of routing budgets through Amazon or Google. In CTV, neutrality is a live selling point — which is why the moat is narrowing, not gone.

6. What would make the moat fade — and the signal that warns first

A narrow-moat verdict is only useful if it names the disconfirming evidence. Here is what erosion looks like, in order of how much it would matter, and where it shows up first.

Source: "revenue may not necessarily grow at the same rate as spend" [34]; FY2025 gross spend $13,394.7M vs revenue $2,896.3M [35]; two holding companies each above 10% of gross billings [36]; Kokai as the major platform upgrade [37].

The single decisive variable is the take rate. Because reported revenue is a percentage fee on a gross-spend flow roughly five times larger — $13.4 billion of spend became $2.9 billion of revenue in FY2025 [38] — the moat is ultimately a claim that TTD can keep charging its fee. Management's own risk factor is the tell: revenue "may not necessarily grow at the same rate as spend on our platform" as pricing competition and volume discounts bite [39]. A moat you can see in retention and margins is real; but the day the take rate rolls over is the day the market learns whether the switching costs were deep enough to defend price, or only deep enough to defend the relationship. Watch revenue-versus-gross-spend before you watch anything else.

Two management responses are worth tracking as moat-widening attempts, both still unproven: OpenPath, which routes buyers to a "simplified, direct connection to publishers," strengthening supply-side control [40], and the continued rollout of Kokai, the AI-driven platform upgrade whose fumbled launch caused the 2024 miss but whose success would deepen the optimization lock-in [41]. Neither yet counts as moat; both are options on a wider one.

The bottom line

The Trade Desk has a narrow moat, and that verdict is deliberate on both words. Moat, because the advantage is company-specific, mechanistically grounded, and — the rarest proof of all — demonstrably durable: retention above 95% survived a demand recession and a self-inflicted product failure, the margin gap over independent peers has widened with scale, and the data network has compounded from 200 to 370 vendors. Narrow, because the lock-in is behavioral rather than contractual (management admits it is cheap to leave), the neutrality edge is being contested head-on by Amazon, the identity-standard bet is still optionality, and the entire structure rests on a take rate the filings themselves flag as not guaranteed. This is a business that can protect its returns — but an investor should underwrite it as a franchise that must keep earning its moat every quarter, not one that has already locked it in. The number that will settle the debate is the take rate; everything else in this tab is context for that one line.

---

Financial Shenanigans — The Trade Desk, Inc. (TTD)

Forensic verdict: Watch (score 38/100), sitting at the top of the band. The Trade Desk's books are among the cleaner ones in ad-tech: revenue is recognized conservatively on a net (agent) basis, cash flow is genuine cash rather than a payables trick, there is no acquisition roll-up to disguise, and Ernst & Young — auditor since 2015 — has issued clean opinions with a single, well-telegraphed Critical Audit Matter [1]. The risk here is not fabricated earnings — it is optics and governance: a business whose headline cash generation is roughly two-thirds funded by adding back ~$490 million of annual stock-based compensation, a balance sheet that dwarfs the income statement because TTD passes ~$13 billion of client spend through it, a one-time deferred-tax tailwind flattering FY2025 cash flow, and a founder-controlled governance structure carrying an $819 million disputed CEO option, a Nevada re-domicile, a revolving cast of CFOs, and live securities litigation.

The two things to watch are (1) the stock-based-compensation-funded cash-flow narrative and (2) the breeding ground — founder super-voting control, the mega-option, and finance-leadership turnover. The cleanest offsetting fact: operating cash flow is not manufactured from working capital — receivables actually grow slower than revenue and TTD funds a net receivable float to its own clients, so cash conversion is real. The single data point that would most move the grade: whether, under a new finance team, the revenue-recognition Critical Audit Matter widens the "gross-basis Supplier Components" scope or the non-GAAP/DSO definitions drift.

Sources: score and flag counts are this analysis's assessment; CFO/NI, FCF/NI, accrual ratio and the Adjusted-EBITDA-to-net-income gap are derived from reported financials, FY2023–FY2025 Statements of Cash Flows and Operations [2]; Adjusted EBITDA reconciliation [3].

The intermediary lens — why the balance sheet looks nothing like the P&L

Read this first: everything else follows from it. TTD is a demand-side platform that sits between advertising agencies and media suppliers. In FY2025 roughly $13.4 billion of client spend ("gross spend") flowed across the platform, but only $2.90 billion was recognised as revenue — TTD keeps a platform fee of roughly one dollar in five and passes the rest through [4]. Because TTD invoices agencies for the gross amount and owes suppliers the gross amount, its receivables and payables are sized to the ~$13 billion flow, not the ~$2.9 billion of revenue.

Source: FY2025 10-K, MD&A gross-spend table [5] and Consolidated Balance Sheets [6].

The forensic consequence: standard balance-sheet screens misfire on TTD. Accounts receivable of $3.77 billion is 130% of net revenue, which would flag as a five-hundred-day DSO on any naive screen — but that is an artifact of gross billing, not slow collection or channel-stuffing [7]. Any DSO, accrual, or receivables-to-revenue metric on TTD must be read against gross spend, not revenue. This same structure is the reason the SEC challenged TTD's gross-vs-net presentation before the IPO (see Breeding ground), and it is why revenue recognition remains the auditor's sole Critical Audit Matter today.

Cash-flow quality — real cash, but SBC-funded and partly one-time

Takeaway: the cash is genuine, but the "quality" flattering it is stock-based compensation add-backs and a one-off deferred-tax swing, not operating leverage — and it is not a payables trick. Over FY2023–FY2025 operating cash flow ran 2.3x reported net income and free cash flow 2.0x net income — the kind of spread that should trigger a "why is CFO so far above earnings?" investigation.

Source: FY2025 10-K Consolidated Statements of Cash Flows (FY2023–FY2025) [8]; earlier years company filings, as reported.

The mechanism is unambiguous once you decompose the FY2025 bridge from $443 million of net income to $993 million of operating cash flow. The single biggest reconciling item is $490.6 million of stock-based compensation — larger than net income itself — followed by a $167.7 million deferred-income-tax add-back and $115.8 million of depreciation. Working capital was a net drag of roughly $275 million, not a source [9].

Source: FY2025 10-K Consolidated Statements of Cash Flows [10].

Three findings fall out of this bridge:

1. The deferred-tax add-back is a non-recurring tailwind (yellow). Deferred income taxes swung from reducing cash flow by $76.9 million in FY2024 and $61.6 million in FY2023 to adding $167.7 million in FY2025 — a ~$245 million year-over-year swing that alone explains most of FY2025's 34% CFO growth. It is driven by the July 2025 OBBBA tax law, which let TTD recognise $175 million of domestic R&D costs as an income-tax receivable with a matching reduction in deferred tax assets (the balance-sheet deferred-tax asset fell from $230 million to $56 million) [11]. This is legitimate, but it is a one-time cash-flow lift that will not repeat, and it makes FY2025 CFO growth look better than the underlying run-rate.

2. Working capital is a drag, not a lifeline (green — clean). The classic "CFO propped up by stretching payables" shenanigan is absent here. In FY2025 the AR build (−$433 million) exceeded the AP build (+$291 million), a net use of cash — because, by TTD's own risk disclosure, it "generally pay[s] our accounts payable on shorter cycles than we collect on our accounts receivables, requiring us to remit payments from our own funds" [12]. TTD funds float to its clients rather than living off supplier float. Receivables have grown in line with or slower than revenue every year, further evidence there is no collection-timing game.

Source: FY2025 10-K Consolidated Balance Sheets and prior-year filings [13]; earlier years company filings, as reported.

3. There is no financing-into-operating or acquisition-into-operating distortion (green — clean). TTD carried no drawn debt at year-end, borrowings under its revolving facility are shown in financing, its lone FY2025 acquisition (Sincera) cost $4.4 million of cash and sits in investing, and a prior-year cash-flow reclassification was limited to aggregating the credit-loss provision line with "no impact to cash flows from operating, investing or financing activities" [14]. CF1, CF2, and CF3 all test clean.

Earnings quality — conservative recognition, rate-sensitive income, thin reserves

Takeaway: recognition is conservative and there is no capitalization game, but two yellow items deserve tracking — the rate-sensitivity of "other income" and a very thin credit-loss allowance.

Revenue recognition is conservative (green, with a caveat). TTD recognises revenue at a point in time, when a bid is won and the client's purchase clears the platform, and books the platform fee net (as agent) — it only grosses up the narrow category of "Supplier Components" (value-added services and data where it controls the service) [15]. Net presentation is the conservative choice; the risk is one-directional — that the gross-basis Supplier-Components scope quietly widens — which is exactly why E&Y flags revenue recognition as its Critical Audit Matter, citing "the high degree of audit effort in performing procedures related to client purchases" [16]. No pull-forward is visible: receivables grew 13% against 18% revenue growth in FY2025.

"Other income" is rate-sensitive, not operating (yellow). With a securities-and-cash portfolio around $1.3 billion, interest income has been a meaningful and rate-cycle-dependent contributor to pre-tax income — peaking near 34% of operating income in FY2023 before fading as rates and the cash balance fell. Adjusted EBITDA correctly deducts the $68.7 million of net interest income, but GAAP pre-tax income and EPS carry it, so a chunk of "profit growth" in 2022–2023 was the Fed, not the platform [17].

Source: derived from reported financials, FY2021–FY2025 Statements of Operations [18].

No expense-shifting via capitalization (green). TTD capitalises almost nothing: capitalised software development costs were just $12.8 million in FY2025 against ~$490 million of expensed technology spend, so there is no "route R&D into an asset" game. Capital expenditure did surge — $47M → $98M → $197M across FY2023–FY2025 — but that is a genuine data-center build-out, disclosed as such, and depreciation is rising behind it (D&A up from $80M to $116M). Capex running ahead of depreciation is expected during a build phase; it is worth tracking, not a red flag, but note that $104.5 million of FY2025 capital assets were financed through accounts payable (a non-cash item), so the cash capex line understates the commitment [19].

Source: FY2025 10-K Consolidated Statements of Cash Flows [20]; earlier years company filings, as reported.

Under-reserving is a low-grade yellow. The allowance for credit losses is $12.2 million against $3.77 billion of gross receivables — 0.32% [21]. That looks thin, and it is partly defensible: "sequential liability" terms mean many agency receivables are not TTD's credit risk until the advertiser pays. But the allowance has barely moved ($11.2M → $12.2M) while receivables have more than doubled since FY2021, and the CTV/agency concentration is rising (below). A reserve that never grows into a fast-growing, concentrated book is a place a future problem would hide.

No big bath (green). The December-2024 reorganization that accompanied the Q4-2024 revenue miss was framed as an operational reset "to accelerate opportunities across CTV, retail media, identity, supply chain optimization, and audio" [22]. Critically, it produced no discrete restructuring charge and no asset write-off — severance was expensed within operating expense, and TTD states "no material impairments have been recorded" on its long-lived and intangible assets [23]. There is no evidence of kitchen-sinking a weak quarter to lower a future base.

Metric hygiene — stock-based comp is the whole game

Takeaway: the single largest forensic issue on the page is metric optics — adjusted profitability and "free cash flow" are both heavily subsidised by adding back real, recurring, dilutive stock compensation. This is the strongest yellow flag.

In FY2025, GAAP net income was $443 million but Adjusted EBITDA was $1,196 million — a $753 million gap, of which stock-based compensation ($491 million) is roughly 65% [24]. SBC is not an unusual or "one-time" cost for TTD — it has run remarkably flat at ~$490 million for three straight years and is the primary tool the company uses to pay its people. Adding it back to reach an "operating" metric, then also adding it back inside operating cash flow, means the same ~$490 million is excluded twice from the numbers investors anchor on.

Source: FY2025 10-K Statements of Cash Flows (SBC and FCF) [25].

Stated as a ratio, SBC equalled 89% of free cash flow in FY2023 and still 62% in FY2025. The trend is genuinely improving as revenue scales (SBC fell from 25.3% to 16.9% of revenue), which is why this is a yellow and not a red. But a PM should treat TTD's ~$796 million of FY2025 free cash flow as roughly $305 million once the economic cost of the equity used to generate it is respected — and note that TTD spent $1.38 billion on buybacks in FY2025, more than its entire free cash flow, largely to mop up that dilution and shrink the share count; that buyback drained cash from $1.37 billion to $658 million and pushed the company to a small accumulated deficit [26].

Source: FY2025 10-K Adjusted EBITDA reconciliation [27].

"Gross spend" is a scale-flattering metric, but honestly labelled (yellow). Management leads with $13.4 billion of gross spend — a number 4.6x revenue — "to assess our market share and scale" [28]. It is disclosed with a clear reconciliation and caveats, so it is not deceptive — but it is the same gross-flow framing the SEC pushed back on pre-IPO, and it lets TTD present a far larger scale figure than its actual take. Balance-sheet metric distortion (KM2) is structural, not managed: the concentration disclosure shows two agency holding companies accounted for 30% of gross billings in FY2025 (each above 10%), up from one holding company above 10% in FY2024 [29]. Rising customer concentration against a static, thin allowance is the intersection to monitor.

Breeding ground — founder control, an $819M option, a re-domicile, and a CFO carousel

Takeaway: the governance breeding ground is the elevated part of this file and it amplifies, rather than dampens, the metric-optics risk — because the people who define the adjusted metrics also control the votes.

Founder super-voting control. Class B shares carry ten votes to Class A's one; as of December 31, 2025, Class B holders — including executives, directors and their affiliates, led by founder-CEO Jeff Green — held ~49.9% of total voting power, enough to influence or control most shareholder matters, with the dual-class structure not sunsetting until December 2035 [30]. Concentrated founder control with limited independent challenge is a classic breeding-ground condition.

The $819 million CEO Performance Option. In October 2021 the board granted Green a market-based option over up to 16 million shares (up to 19.2 million with a TSR modifier), exercise price $68.29, price hurdles from $90 to $340, grant-date fair value ~$819 million, expensed graded-vesting over ~five years [31]. This one grant drove $198M, $128M and $67M of SBC in FY2023–FY2025 respectively — i.e. it is a large slice of the very SBC add-back that inflates Adjusted EBITDA and CFO, tying the governance flag directly to the metric-hygiene flag. The option has been the subject of shareholder litigation, and TTD's November 2024 reincorporation from Delaware to Nevada — an anti-takeover-tightening move — is widely read as related to that dispute.

Live litigation. Following the Q4-2024 miss and stock drop, securities class actions were filed in February and March 2025, alongside consolidated data-privacy/wiretapping suits — legal exposure that management notes will make G&A "fluctuate period to period" [32]. These are disclosed as reasonably-possible with no material accrual, consistent with early-stage cases, but they are a governance-and-contingency watch item.

Finance-leadership turnover (yellow). The CFO seat has changed hands repeatedly across the corpus — Blake Grayson, then Laura Schenkein, then Alex Kayyal, then Tahnil Davis — and FY2025 G&A includes "an acceleration of stock-based compensation in connection with an executive transition" [33]. Repeated turnover in the office that owns the non-GAAP definitions and the DSO/DPO narrative raises the odds of a definitional drift and is the reason "watch the metrics under the new team" is the grade-moving variable.

The one already-settled negative — historical, and remediated. Before the IPO, the SEC's 2016 comment letters directly challenged TTD's gross-vs-net revenue presentation, its gross-spend metric, agency-customer concentration, and a restatement of 2014 financials tied to identified material weaknesses in internal control [34]. This matters as context — the same pressure points (gross vs net, concentration) persist structurally today — but it was resolved a decade ago, the company IPO'd, and there has been no restatement, auditor change, or material-weakness disclosure in the public-company era since. That clean subsequent record is a major reason the score stays in "Watch" rather than "Elevated."

The 13-category scorecard

Every shenanigan category, graded. Only three carry a live yellow; the rest test clean or benign. The concentration is in metric hygiene and one-time items, not in the core statements.

Sources: severity/evidence derived from FY2021–FY2025 10-Ks — Statements of Cash Flows [35], Balance Sheets [36], Adjusted EBITDA reconciliation [37], revenue-recognition note [38] and concentration note [39].

What to underwrite next

Five specific, named diligence items — not generic caveats:

1. Discount free cash flow for SBC. Treat reported FCF (~$796M FY2025) net of stock-based compensation (~$491M) — a "cash FCF" nearer $305M — when setting a valuation multiple or FCF yield. Track SBC/revenue: continued decline (from 16.9%) confirms the improving trend; a reversal is a downgrade trigger.
2. Normalize FY2025 CFO growth for the OBBBA deferred-tax lift. The $168M deferred-tax add-back and $175M R&D tax receivable are one-time; recompute the FY2026 CFO/NI ratio without them before extrapolating 34% CFO growth.
3. Watch the revenue-recognition CAM and the "Supplier Components" gross-basis scope. A material widening of what is recognised gross (vs the current net platform fee) would inflate reported revenue growth without more economic activity — read Note 2 in each 10-K/10-Q for scope language.
4. Watch the allowance-to-receivables ratio against rising concentration. $12.2M (0.32%) against a book where two holding companies are now 30% of gross billings. A provision that stays flat while concentration climbs, or a single large advertiser default, is where a loss would surface.
5. Watch non-GAAP and DSO/DPO definitions under the new finance team. After repeated CFO turnover, any change in Adjusted EBITDA add-backs, the gross-spend definition, or the DSO/DPO framing in the MD&A/press releases is a red-flag escalation.

Signal that would DOWNGRADE the grade (toward Elevated/High): the gross-basis revenue scope widens materially; a non-GAAP/gross-spend/DSO definition changes without clear reconciliation; the credit-loss allowance proves inadequate on a concentrated default; or securities litigation moves to a material accrual.

Signal that would UPGRADE the grade (toward Clean): SBC/revenue keeps falling toward peer norms; the CEO Performance Option finishes vesting (fully recognised by Q1 2026) and SBC steps down durably; finance leadership stabilises; and interest-rate-independent operating income keeps carrying the profit growth.

Decisive read: the accounting risk at The Trade Desk is a valuation haircut and a metric-discipline caveat, not a thesis breaker. The statements are conservative, the cash is real, and there is no restatement, auditor, or capitalization red flag. What a PM must not do is take Adjusted EBITDA or reported free cash flow at face value — both are meaningfully subsidised by ~$490M of annual, recurring, dilutive stock compensation that management then spends more than all of FCF buying back — and must weigh a founder-controlled governance structure that concentrates power over exactly the metrics and disclosures where the discretion lives. Size the position for optics and governance risk; do not underwrite it for fraud risk.

---

People & Governance — Founder Control Meets a Hollowed-Out Board

Verdict up front: this is a founder-controlled company where the check-and-balance apparatus is thin and, right now, visibly broken. Co-founder Jeff Green holds 49.7% of the total voting power through super-voting Class B stock while owning only about 11% of the economics [1]. He is both Chairman and CEO, the board seats just five people, and in a two-week stretch before the 2026 annual meeting three of the independent directors resigned, leaving the audit committee with a single member [2]. Against that sits a real dose of alignment: Green's paper wealth collapsed with the stock, he is barred from hedging or pledging, and in March 2026 he put roughly $148 million of his own cash into the open market at the lows. The tension between genuine skin-in-the-game and near-absolute, weakly-supervised control is the whole story of this tab.

Sources: voting power and economic stake from 2026 Proxy, Ownership table [3]; CEO pay from 2025 Summary Compensation Table [4]; 5-year total-shareholder-return value from Pay Versus Performance table [5].

The Control Structure: ~50% of the Vote on ~11% of the Economics

The defining governance fact at The Trade Desk is a dual-class structure. Each Class A share carries one vote and each Class B share carries 10 votes; at the record date there were roughly 427 million Class A shares and 43 million Class B shares outstanding [6]. Green owns 97.6% of the Class B and, through it, 49.7% of the entire company's voting power — while his Class A economic position is just 2.6% of that class [7]. Company-wide, holders of Class B stock (executives, employees and directors) together control about 49.9% of the voting power, a concentration the company itself flags as able to "deter, delay or prevent a change of control" [8].

Source: 2026 Proxy, Ownership of Common Stock — % of Total Voting Power column [9].

Two mitigants keep this from being absolute autocracy. First, Class A holders vote as a separate class to elect one dedicated "Class A Director" — currently Andrea Cunningham — a structural minority-shareholder seat [10]. Second, the super-voting Class B converts automatically to one-vote Class A on December 22, 2035, a hard sunset that eventually dissolves the control premium [11]. Notably, in 2025 the board formed a special committee to consider extending the dual-class structure, chaired by then-director Lise Buyer — a live governance flashpoint given that Buyer has since resigned [12].

The Board Just Lost Half Its Independence

For a company of The Trade Desk's scale (roughly $2.9 billion of revenue), the board is strikingly small — five directors, divided into three staggered classes so that only part of the board faces election in any year [13]. Of those five, only three are independent: Green (CEO) and Samantha Jacobson (Chief Strategy Officer) sit on the board as management [14]. Green serves as his own Chairman; the board's stated rationale is that his knowledge "renders him best positioned" to set the agenda, and independent directors merely "intend to appoint a lead independent director" — meaning at the time of the filing, no lead independent director was in place [15].

The acute problem is turnover. In the weeks before the 2026 annual meeting, three directors resigned in rapid succession — Kathryn Falberg (March 23, 2026), and Lise Buyer and Gokul Rajaram (both April 3, 2026) — all of whom had served on the audit committee [16]. The result is a governance apparatus running on fumes: as of the proxy, Andrew Vollero is the sole member of the audit committee, with the board conceding it must "fill the vacancies as expeditiously as possible" [17]. The nominating & governance committee likewise has one member (Cunningham), and the compensation committee has two.

Sources: director bios and independence from 2026 Proxy, Election of Directors [18] and Independence of the Board [19]; committee membership from Board Committees [20].

On the positive side, the two newest independents are substantive appointments: Omar Tawakol (founder of BlueKai and Voicea, both acquired; ad-tech and data-platform veteran) and Andrew Vollero (CFO of Reddit; former first CFO of Snap) bring exactly the operating and financial-oversight expertise the board needs [21]. Whether the board rebuilds genuine independent depth around them is the key thing to watch.

The People: Deep Founder, Thin and Churning Bench

Jeff Green is a credible, industry-shaping founder — he built AdECN, "the world's first online advertising exchange," sold it to Microsoft in 2007, and co-founded The Trade Desk in 2009 [22]. The concern is not capability at the top; it is concentration and succession. The executive bench beneath him has been unstable, most visibly in the finance seat.

Rapid CFO turnover during a period of stock-price collapse and an active securities-fraud suit is a meaningful negative signal about stability in the office that matters most for financial credibility. The rest of the named team is more settled — Chief Legal Officer Jay Grant (since 2020), CSO Samantha Jacobson (since 2021), and COO Vivek Kundra, the former first U.S. federal CIO, who joined in March 2025 [24].

Compensation: Modest Cash, Enormous Equity, and a Pay Curve That Actually Bent

For 2025, Green's total compensation was $27.4 million — a $1.35 million salary, $2.8 million of cash incentive, and roughly $23 million of equity split evenly between restricted stock and options [25]. About 85% of his pay is equity, and the CEO-to-median-employee pay ratio is 125:1 — high in absolute terms but modest for a mega-cap software CEO, helped by an unusually well-paid median employee ($218,847) [26].

Source: 2025 Summary Compensation Table; cash/equity split derived from reported salary, non-equity incentive, stock and option awards [27].

The 2021 CEO "Moonshot" Performance Option

The centerpiece — and the most controversial element — of Green's pay is a market-based performance option granted in October 2021 covering up to 16,000,000 Class A shares at a $68.29 exercise price, vesting in eight tranches only if the stock reaches escalating targets from $90 to $340 per share [28]. This is a genuinely demanding, stockholder-aligned design — the CEO earns nothing unless shareholders first capture large gains. Only the first two tranches (4,800,000 shares, at the $90 and $115 targets) have vested, in 2021 and 2024; no further target has been hit since [29].

Source: 2026 Proxy, 2021 CEO Performance Option — vesting-tranche table [30]; vested tranches per the same section [31].

The design's credibility was tested in court and held: derivative suits (Huizenga and Pfeiffer v. Green) alleged the board breached its fiduciary duties in approving the grant, but the Delaware Court of Chancery dismissed them with prejudice, and the Delaware Supreme Court affirmed the dismissal on November 5, 2025 [32]. The context that matters today: with the stock at $37.96 at year-end 2025, both the exercise price ($68.29) and the next target ($145) are far out of the money, so the option — and most of the 2025 equity grant struck near $49 — is currently deeply underwater [33].

Pay-for-Performance: The Rare Case Where It Actually Worked

The most reassuring number in the entire proxy is that Green's realizable pay moved violently with the stock. The SEC's "compensation actually paid" measure — which marks equity to market — swung from positive $1.2 billion in 2021 to negative $857 million in 2025 as the option value evaporated [34]. Meanwhile a $100 investment at the start of the measurement window was worth just $47.39 by the end of 2025 — versus $137.99 for the peer group. The CEO's fortune fell with shareholders'; the problem is that shareholders fell hard.

Source: 2026 Proxy, Pay Versus Performance — Value of Initial Fixed $100 Investment [35].

Directors are paid conventionally — a $50,000 base cash retainer plus committee fees and a $290,000 annual equity award; 2025 director totals ran from roughly $382,000 to $622,000 [36].

Skin in the Game: Sold High, Then Bought the Crash

Insider behavior here cuts both ways, and the sequence matters. Across 2024 and early 2025 — with the stock in the $90–$120 range — Green sold roughly $443 million of Class A stock [37]. Then, after the collapse, he did something founders rarely do: in early March 2026 he bought about $148 million of stock in the open market — including 2,314,304 shares at $25.08 on a single day — the first material insider purchases on record [38].

Source: derived from SEC Form 4 filings, Insider Activity — Green sales 2024–2025 and open-market purchases March 2026 [39].

The buy-the-dip purchase is a real conviction signal and a point in management's favor. But it is double-edged: the same window in which Green sold ($443 million, late-2023 through mid-2025) is precisely the period a securities class action alleges the CEO, then-CFO and CSO traded on inside information (see below). Two structural protections partly offset the concern — the company prohibits all hedging and pledging of its shares by executives and directors, so there is no margin-loan overhang on Green's stake [40], and all Section 16 ownership reports were filed on time during the last fiscal year [41].

Governance Risk: Reincorporation, Litigation, and Related Parties

The Nevada move. In November 2024 the company reincorporated from Delaware to Nevada. The practical effect is to narrow the liability exposure of directors and officers — the charter limits monetary liability "to the fullest extent permitted by Nevada law" and adopts an exclusive-forum provision routing state-law claims to Nevada courts, which the company acknowledges "may discourage stockholders from bringing a lawsuit" against directors [42]. The switch drew a stockholder class action (Gunderson v. The Trade Desk) alleging the reincorporation was "substantively and procedurally unfair"; the court has rejected the supermajority-vote claim but the fairness claims survive, and a related books-and-records fight (Scarantino) is on appeal to the Delaware Supreme Court as of January 2026 [43].

Securities litigation naming the top of the house. Following the February 2025 stock drop, consolidated federal securities class actions allege false and misleading statements — and, critically, a Section 20A claim that the CEO, then-CFO and CSO engaged in insider trading over a class period running November 2023 to August 2025 [44]. This is unproven, but it directly implicates the same executives and the same trading window discussed above and is the governance overhang most worth tracking.

Related-party dealings are limited and cleanly handled. The only material related-party items are benign or reimbursed: the company reimbursed Green roughly $1 million for legal fees defending litigation brought against him in his CEO capacity (a standard indemnification) [45], and his roughly $628,000 of personal aircraft use was fully reimbursed to the company, so it added nothing to reported pay [46]. There is no evidence of self-dealing, promoter-style asset recycling, or undisclosed insider contracts.

A founder-led philosophy that dismisses outside checks. Management is explicit that it rejects "the 'one size fits all' approach often promoted by major proxy advisory firms," arguing such frameworks "fail to account for the specific context and long-term vision of founder-led companies" [47]. Read charitably, that is conviction; read skeptically, it is a controlling founder pre-emptively discounting the very external scrutiny that a thin, self-selected board fails to supply internally.

Verdict

The single thing most likely to move the grade: whether the board rebuilds genuine independent depth — filling the audit committee, seating a lead independent director, and adding independents who can actually check a controlling founder — and how the securities/insider-trading litigation resolves. Success there could lift this toward a B; another wave of independent departures, or an adverse litigation finding, would push it toward a D.

---

History — The Trade Desk: The Machine That Beat 33 Quarters, Then Broke One

For eight years The Trade Desk was the rarest thing on the public market: a hypergrowth company that hit its own numbers every single quarter — 33 in a row — under the same founder-CEO who built it from nothing. Then, in the results reported February 2025, the machine finally slipped: a first-ever revenue miss that erased roughly a third of the company's value in a day. What makes this history worth reading is not the miss itself but how management handled it — CEO Jeff Green opened the call by saying "it was our fault" [1] — and whether the recovery that followed rebuilt the trust or merely papered over a structural slowdown. The short answer: candor and delivery are largely intact, but the flawless-forecaster premium is gone, growth is decelerating every quarter, the C-suite is churning, and the founder just entrenched his control for another decade.

Sources: credibility score is this analysis's judgment (see verdict below); the 33-quarter streak and the miss per the Q4 FY2024 call [2]; revenue and buybacks derived from reported financials.

The Founder, the Anchor, and the One Stat That Never Moved

There is no "previous regime" to compare against here. Jeff T. Green co-founded The Trade Desk in 2009 and has been its CEO the entire time — the FY2021 10-K names it "a Delaware corporation established in 2009" run by "our two founders, Jeff T. Green, our Chief Executive Officer, and David R. Pickles, our Chief Technology Officer" [3], and he is still introduced as "CEO and Co-Founder" on the Q1 FY2026 call [4]. This matters for every other tab: the quality of this business was built by the current team, not inherited. When we judge capital allocation or execution, Green owns both the credit and the blame.

The single most durable line in the entire record is client retention. It has been quoted at "over 95%" for as long as the company has been public — Green cited it on the Q3 2021 call ("our retention rate remains over 95%") [5], and by the FY2024 10-K it had hardened into "over 95% in each of the last eleven years" [6]. That number never wobbled through the ad recession, the platform migration, or the miss. It is the one claim the history fully vindicates.

Chapter One — The Unbroken Streak (2016–2024)

The FY2021–FY2024 story is consistency. Revenue compounded at 30%+ off a growing base, guidance was beaten and raised every quarter, and management repeated the same handful of themes with almost liturgical discipline: connected TV as a generational shift, a "$1 trillion" advertising TAM, objectivity (TTD represents only the buy side and owns no media), and — above all — gaining share regardless of the macro.

Source: derived from reported financials, FY2021–FY2025 10-Ks [7].

The themes arrived on a schedule you can almost set a watch to. The Solimar platform launched in mid-2021 alongside a 101% revenue jump ("our revenue was up 101% from a year ago … significantly surpassing our own expectations") [8]. Green framed the runway as "$1 trillion industry TAM" [9] and made a bold long-arc promise — "we're going to get from one to 5 billion ahead of anybody else playing the same game" [10]. Then came the supply-path product OpenPath in early 2022 ("we announced another very big initiative, OpenPath") [11] and the identity play UID2 ("we developed and launched UID2 … email and phone number-based instead of cookie-based") [12].

The counter-cyclical thesis was tested in the 2022 ad downturn and, remarkably, held. Green reframed the recession as opportunity — "we look at this as land grab time" [13] — and the numbers backed him: TTD grew 24% in a year when "most of our large competitors were posting between negative 9% and negative 2% growth," which Green called the best relative outperformance "in our six years or so as a public company" [14]. By mid-2023 the flagship Kokai AI platform arrived, and Green declared TTD had "gained more market share than in any other period in our company's history" [15].

Two 2021 events set the governance baseline the reader should remember: a ten-for-one stock split in June 2021 [16], and a CEO Performance Option granting Green up to 16 million shares at a $68.29 strike, vesting on stock-price milestones from $90 to $340 — a grant valued at roughly $819 million [17]. The streak was real; so was the founder's ambition to be paid like it continued forever.

The Tell — The Q4 2024 Miss

Right up to the edge, the tone was peak confidence. In Q1 2024 Green had "never been more optimistic about the future of the open Internet" [18]; in Q2 he called Kokai "the product that I am most proud of our team for shipping in my whole career" [19]; and on the Q3 call he set the bar that would break him — "we estimate Q4 revenue to be at least $756 million" [20].

Q4 came in at $741 million — a $15 million shortfall, the first guidance miss in the company's public life. The response is the single most credibility-defining moment in the record, and it is a case study in owning it rather than spinning it:

Source: Q4 FY2024 Earnings Call, CEO remarks [21].

Green attributed the miss to "a series of small execution missteps while also preparing for the future" [22], and — crucially — his CFO removed the two easiest excuses: "this miss was not due to lack of opportunity or increased competition, it was on us" [23]. That is the honest-miss signature: no macro scapegoat, no competitor to blame. He then committed to concrete structural fixes rather than reassurance — "the largest reorganization in company history in December," a new agile engineering structure, a forced completion of the Solimar-to-Kokai migration, and the launch of the Ventura connected-TV operating system [24]. He even conceded a long-standing gap: "we have managed to operate without a COO for a while, and there is no reason we shouldn't bring in a top-tier COO" [25]. Guidance for the next quarter was cut hard, to "at least $575 million reflecting 17% year-over-year growth," and the buyback authorization was lifted to $1 billion [26].

The guidance record around the break tells the story cleanly: an unbroken streak, one sharp miss, then resumed beats — but off a permanently lower growth bar.

Sources: Q3 FY2024 call (Q4 guide) [27]; Q4 FY2024 call (Q4 actual, Q1 FY2025 guide) [28]; Q4 FY2025 call (Q1 FY2026 guide) [29]; Q1 FY2026 call (Q1 actual) [30].

Chapter Two — Rebuilding the Machine (2025 → now)

The first quarter after the miss was framed as redemption. Q1 2025 revenue grew 25% to $616 million, "far surpassing our own expectations… we did it again" [31]. The promised COO materialized — "we've hired a new COO, Vivek Kundra" [32] — and a genuine external tailwind appeared: "Google has been declared an illegal monopoly in two separate instances in 2025 by the U.S. courts" [33]. By Q3 2025, Kokai adoption reached "nearly 85% using Kokai as their default experience," and TTD named a new Chief Revenue Officer, Anders Mortensen [34]. Capital returns became the defining act of the chapter: cumulative buybacks reached "nearly $2 billion" [35].

Source: derived from reported financials, FY2022–FY2025 cash-flow statements; cumulative repurchases confirmed on the Q3 FY2025 call [36].

But this is not a clean redemption arc, and the honest read has to say so. Growth decelerated in a straight line even as the beats resumed — from 25% in Q1 2025 to 12% by Q1 2026 — and guidance stepped down with it. The recovery restored reliability, not momentum.

Sources: derived from reported quarterly revenue; Q4 FY2024 (22%) [37] and Q1 FY2026 (12%) [38] per the earnings calls.

The second warning sign is churn at the top. Having operated for years with no COO and both founders in the C-suite, the leadership table has turned over repeatedly since the miss:

Sources: Pickles exit, FY2023 10-K [39]; Kundra, Q1 FY2025 call [40]; Kayyal, Q2 FY2025 call [41]; Mortensen, Q3 FY2025 call [42]; Interim CFO Davis, Q1 FY2026 call [43]; Jacob, Q1 FY2026 call [44].

Three different people have held or led the CFO seat inside roughly a year — Schenkein, then Kayyal, now Interim CFO Tahnil Davis [45] — and the Chief Strategy Officer confirmed on the Q1 2026 call that "Samantha is taking a role at OpenAI" [46]. A reorg that promised stability has, so far, produced a revolving door.

What Management Stopped Saying — Narrative Drift

The most revealing signal in a decade of calls is not what recurs but what vanishes. The proud beat-and-raise boast — the drumbeat of "we outperform every quarter" — effectively disappeared after the miss. Solimar gave way to Kokai. And an entirely new, defensive theme appeared in late 2025: sustained weakness in CPG and auto, "verticals [that] represent over a quarter of our business" [47] — the kind of demand-side caveat the pre-miss TTD never needed to lead with.

Source: derived from a full read of the FY2021–Q1 FY2026 earnings calls and FY2021–FY2025 10-Ks; representative anchors — OpenPath launch [48], Kokai adoption [49], CPG/auto weakness [50].

One theme should have shifted but was overtaken by events: the identity bet. For years TTD's UID2 was pitched as the escape hatch from Google's promised death of the third-party cookie. Then Google reversed course — "on April 22, 2025, Google announced that it would maintain its current approach to offering users third-party cookie choice in Chrome" [51]. The premise of a whole strategic chapter dissolved, and management absorbed it without drama, repositioning UID2 as a quality/identity layer rather than a cookie-replacement lifeboat.

Founder Control — The Governance Drift That Cuts the Score

The one place management's actions sit uneasily against its shareholder-friendly words is control. In November 2024 TTD reincorporated from Delaware to Nevada — the FY2024 10-K now opens "The Trade Desk was originally incorporated in 2009 and is a Nevada corporation" [52] — a move a stockholder immediately challenged in a Delaware class action alleging the switch was unfair to investors [53]. Then, in 2025, management proposed extending the dual-class structure Green had once implied would sunset; the FY2025 10-K confirms Class B super-voting shares now "convert automatically into Class A common stock on December 22, 2035" [54] — a full decade later than the previously disclosed 2025 date. Green addressed it directly on the Q2 2025 call, defending "the proposed extension of our dual-class share structure" [55]. Reincorporating for stronger anti-takeover protection and entrenching founder voting control for another ten years — in the same window as the first miss and the C-suite churn — is the clearest evidence that this remains, structurally, a founder's company first.

The Credibility Verdict

Score: 7 / 10. The Trade Desk earns a well-above-average grade, and it is defensible from the record — but it is no longer the 9 the pre-2025 track record commanded.

Leadership and chapter anchors (used by every other tab): the current CEO is co-founder Jeff Green, in the seat since 2009 [56]; the business quality was built, not inherited. The current strategic chapter began in 2024 — the Kokai/AI platform pivot, the December reorg, the Nevada reincorporation, and the first miss together mark the break from the "unbroken streak" era into a "rebuild-and-scale-to-$10B" phase Green now describes explicitly [57].

What to believe versus discount. Believe the retention, the CTV leadership, the objectivity model, and — importantly — that when this team misses, it will tell you the truth. Discount any implied return to 25%+ growth; the deceleration looks structural (scale plus real CPG/auto demand softness), not a one-quarter stumble. The story today is simpler and more honest, but less exciting than the streak era: a maturing, cash-generative, buyback-funded compounder run by a candid but increasingly entrenched founder — with credibility that has stabilized after a real scare, not fully re-earned the premium it once held.

---

The calls in one screen

This tab reads every earnings call The Trade Desk has held from Q2 2021 through Q1 2026 — 20 company calls — plus 17 recent calls across six peers (Amazon, Alphabet, Magnite, PubMatic, Viant and Criteo), so you do not have to open the transcripts. The single most important thing five years of calls tell you is a story of deceleration and a broken streak: revenue growth has slid from triple digits in 2021 to 12% in the most recent quarter, and in February 2025 management missed its own guidance for the first time in the company's public life — an event that reframes how you should read everything since.

Source: Q1 FY2026 earnings call — CFO results and Q2 guide [1]; Q2 2026 guide of "at least $750 million" [2].

In Q1 2026 the interim CFO reported "revenue of $689 million, representing 12% year-over-year growth" and "$206 million of adjusted EBITDA… representing a 30% margin" [3] — a 30% margin that is below the 33-34% the company posted in the first quarters of 2024 and 2025, and a long way from the "revenue was up 101% from a year ago" the founder opened with in mid-2021 [4]. The through-line of this tab: a category-defining compounder that ran a flawless promise-and-deliver streak for eight years, tripped once, and has been decelerating and reinvesting through a defensive stretch ever since.

The guidance track record — the credibility ledger

The highest-value thread in five years of calls is promise versus delivery, and you can only build it by lining the calls up in sequence. The Trade Desk guides to a revenue floor ("we expect revenue to be at least $X") every quarter. For 32 straight quarters it cleared that floor. Then, on the Q4 2024 call, it did not: management had guided to "at least $756 million" and "cautiously optimistic for Q4" [5], then delivered "$741 million, a 22%" increase [6] — a roughly $15 million shortfall that erased more than a third of the market cap in a day.

Source: guided figures are each quarter's forward revenue floor as stated on the prior call (e.g. Q4 2024 guide [7]); actuals are reported quarterly revenue (e.g. Q4 2024 result [8]). Two guides (Q2 2022, Q1 2023) are omitted — the transcript text was corrupted by the OCR privacy-banner and no dollar figure survives.

Source: derived from guided floors and reported revenue per each earnings call, Q4 2021–Q1 2026 [9][10].

Two things this ledger tells you. First, the beats are real but engineered small: the guided floor is deliberately conservative, and the company clears it by single-digit millions most quarters — so the streak was always a low-bar streak, not a signal of accelerating strength. Second, when the model finally broke it broke down, and the language changed permanently. On the miss call the CEO reached for a sports metaphor — "If this were a sporting event, we would still have a championship caliber team, but in this particular instance, we made too many turnovers" [11] — and admitted the cause was internal, including a self-inflicted product delay: "the deliberate slower rollout of Kokai was intentional" [12]. Notably, management could not cleanly reconcile why no other ad-funded company reported the same weakness that quarter, which is why the market treated a small miss as a large tell.

Post-miss, the "beats" resumed — Q1 2025 cleared its floor by a wide $41 million [13] — but only because the floor itself was reset far lower, against a business now growing 12-19% rather than 25%+. Guidance credibility is best read as "mixed": management reliably beats a bar it sets low, but the bar has fallen off a cliff, and the one time expectations mattered most the company missed.

The growth arc — a five-year deceleration

Reading the calls in order, the deceleration is unmistakable and monotonic in the last two years. The 2021 triple-digit prints were pandemic-recovery and pre-split optics; the durable signal is the step-down from a steady low-to-mid-20s cadence through 2022-2024 into 25% → 19% → 18% → 14% → 12% across the five most recent quarters.

Source: derived from reported quarterly revenue across all 20 company calls; representative prints — Q2 2021 +101% [14], Q4 2025 +14% [15], Q1 2026 +12% [16].

Management's defense of the deceleration is consistent and quantified: it blames a discrete set of laggard verticals rather than the core. On the Q4 2025 call the CEO named "a sustained weakness among some large consumer packaged goods companies… as well as some global auto companies" [17], and argued the company "would have been at least 5% higher in growth rate if you don't include those categories" [18]. CPG and auto together are "over a quarter of our business." By Q1 2026 the scapegoat list had grown to add "geopolitical tensions" and tariffs, and the CEO leaned on a racing metaphor to reframe the pain as opportunity: "you cannot overtake 15 cars in sunny weather, but you can when it is raining" [19]. The recurring rhetorical move across every recent call: quantify the drag, then insist the underlying engine is healthier than the headline.

Margins and the Q1 2026 reinvestment cliff

The Trade Desk's calls have always paired high-40s Q4 adjusted-EBITDA margins with a seasonal Q1 trough — but Q1 2026 broke the seasonal pattern. Margin fell to 30%, below both prior first quarters, as management pivoted openly into what it frames as a year of "disciplined reinvestment" in infrastructure.

Source: adjusted-EBITDA margins as stated on each call; Q1 2026 "30% margin" [20]; Q4 2025 "47% margin" [21].

Capital allocation has been a steady tell of confidence. The company runs a debt-free, cash-rich balance sheet; it authorized its first-ever buyback — "a $700 million share repurchase program" — on the Q4 2022 call [22], and has since expanded the authorization to $1 billion, leaning on buybacks to offset the heavy stock-based compensation that funds a founder pay award and aggressive hiring. Free cash flow was a headline point of pride early — "$320 million in free cash flow" for FY2021 [23] — and management now frames the multi-year ambition as scaling "from about $3 billion in revenue to $10 billion in revenue and beyond" [24]. The Q1 2026 margin cliff is the price of chasing that number.

What management started — and quietly stopped — talking about

The small tells matter as much as the headline. Reading the calls in sequence surfaces metrics and products that appeared, got heavily promoted, then vanished — and disclosures the company has never given.

Source: synthesized across all 20 company calls; anchor examples — retention "over 95% across the previous seven years" [25] vs the later qualitative "retention rate has been so high" [26]; Solimar at 100% adoption [27]; OpenPath "no material impact to take rate" [28]; Kokai launch [29]; JBPs "over 40% of spending" [30]; Ventura [31].

The KPI that never appears is the most telling: in five years management has refused to disclose CTV as a share of revenue, even as it calls CTV its fastest-growing channel and lets analysts infer 50%-plus. On the flip side, the KPIs it does volunteer have shifted from durable trust signals (95%+ retention, Rule-of-40) toward product-efficacy proof points designed to answer the AI-competition worry: Kokai now claims "a 42% reduction in cost per unique reach" [32] and "a 94% better click-through rate compared to Solimar" [33]. When a company starts leading with efficacy statistics, it is usually answering a question the market is asking about competitive threat.

Tone and language — how confidence turned

The adjectives moved before the numbers did. In 2021 the register was euphoric ("wind at our backs," "off the charts"); the first real hedges — "we have never claimed to be a bellwether of the economy" [34] — arrived in 2022; 2024 swung back to peak swagger ("the default DSP of the open internet" [35], "an important moment of reckoning… the fall of Rome" for walled gardens [36]); and 2025-26 turned contrite then embattled.

The heatmap below scores the tone on each major theme by year (a simple −2 defensive / negative to +2 confident / positive scale, read from the language management actually used).

Source: author's read of management language across all 20 calls; scale anchors — 2024 peak confidence "very best position of any player" [37]; 2024 guidance-cushion break (the miss) [38]; 2025-26 defensiveness on Amazon [39].

The one row that never wavers is the CTV secular story — management has framed connected TV as an unstoppable channel shift on every single call, and the peer read below confirms that is not just spin. The rows that decayed are demand/macro, management confidence and — most sharply — the competitive tone toward Amazon, which is the subject the Q-and-A keeps circling.

The Q and A — where the truth leaks

Analysts have pressed hardest on three things across the calls, and how management handles each is itself a signal.

1. Amazon. Every call for two years has featured an Amazon-DSP question, and the answer has escalated from dismissal to quantified defensiveness. Q1 2025: Amazon "serves mainly as a purchasing tool for Prime Video" [40]. Q2 2025, more pointed: "if you insist on calling Amazon a competitor, we compete with a tiny division of Amazon" [41] — while conceding in the same breath that Amazon "nearly doubled the supply of Prime Video inventory" [42]. By Q3 2025 the CEO was arguing the numbers, noting Amazon is "projected to generate around $70 billion in advertising revenue" but that "97% to 99%" of its ad effort is its own owned-and-operated inventory [43], and "I hope they do eventually price it at zero because it will be easier to point out then the problem" of selling your own inventory [44]. When a management team has to keep re-litigating that a competitor is not a threat, the market hears an overhang.

2. Where is the growth coming from — and is it decelerating? In Q2 2025 an analyst asked bluntly "who are you taking it from?" given Meta, Amazon and Google were all growing faster [45] — Green conceded walled gardens optimize AI more easily near-term and pitched the open internet as "a longer game." In Q3 2025 an analyst flagged the Q4 guide as "a 3-point deceleration… anything specifically that's giving you concern?" [46] — the CFO deflected to "no change in guidance philosophy." These are the questions with the least crisp answers.

3. Governance and people risk — the newest pressure. The two most recent calls carried hostile-timed disclosures. In Q1 2026, the first analyst question paired the "deceleration in your Q2 outlook" with a public dispute with agency giant Publicis; the CEO confirmed "since 2018, we have done billions of dollars of business with Publicis… negotiations are ongoing" but declined details [47]. On the same call an analyst raised that the Chief Strategy Officer "is leaving the company to join OpenAI" [48]. Underneath it all: three different people have held the CFO seat across the last four calls, the two most recent as interim, and management extended its dual-class structure in 2025 (defended as "not about entrenchment" [49]). The candor is real, but the accumulation of people-and-governance items is a distinct new risk the calls did not carry before 2025.

On the OpenPath conflict-of-interest press reports, the CEO was dismissive but did — for the first time — disclose the economics: "We charge them 4.5%… nearly breakeven" [50], and jabbed that rivals' DSP is "their 37th highest priority, not their #1 priority" [51].

Money quotes — management in its own words

Peer and industry cross-read — is the industry saying the same thing?

The Trade Desk does not report in a vacuum. Its peers split into three camps: SSPs (Magnite, PubMatic) sell the CTV supply TTD buys; independent DSPs (Viant) and commerce-media (Criteo) compete for the same buy-side budgets; and walled gardens (Amazon, Google) are the "big tech encroachment" the whole thesis turns on. Reading their calls against TTD's answers the investor's core question theme by theme.

Sources: TTD Amazon framing [60]; Amazon DSP "growing rapidly" [61]; Magnite OpenPath default-path [62]; PubMatic on Kokai [63]; Viant "margin trap" [64]; Criteo OpenAI partner [65].

CTV demand — the one clean CONSENSUS, and it validates TTD's core story

Every peer that touches connected TV corroborates the secular tailwind TTD leans on hardest. Viant: "CTV contribution ex-TAC increased by more than 40%, over two-and-a-half times the broader industry growth rate" [66]. PubMatic: "For the full year, CTV grew over 50% year-over-year, excluding political" [67]. Magnite: CTV ex-political "increased by 32%" [68]. And the walled gardens are pouring in: Amazon's Prime Video ad audience reached "315 million viewers globally" [69]; YouTube hit "a record high of 12.8% of total TV viewing" [70]. Verdict: consensus. When TTD says CTV is the channel of the decade, the whole ecosystem agrees — this is the part of the bull case the peer set most strongly supports.

Amazon DSP — TTD is the OUTLIER, and this is the crux risk

Here The Trade Desk stands alone in its dismissiveness. TTD tells you Amazon is "a tiny division" [71]. Amazon tells you the opposite — its CEO calls the DSP "growing rapidly" with gaps closed and "far from reaching our growth potential" [72], riding a ~$21 billion-a-quarter ad business growing a metronomic "22% in the fourth quarter" [73][74], and it is actively extending the DSP beyond its walled garden via a "momentous partnership with Roku" for the largest authenticated CTV footprint in the US [75]. The SSPs treat Amazon as a friend, not a threat — Magnite is "one of only three platforms approved for Amazon DSP spend" [76] and calls Amazon's year a "banner year" [77]; PubMatic says Amazon is now "a top five buyer" on its platform [78]. Even fellow independent DSP Viant, which downplays Amazon at the "finish line," now attacks it as "a margin trap" [79] — hardly the language of a non-threat. Verdict: TTD is the outlier. Whether its dismissiveness is founder-confidence or denial is the single biggest judgment call on these calls.

OpenPath — the peers confirm TTD's supply move is real and disruptive

When TTD soft-pedals OpenPath as ecosystem hygiene, its supply partners tell you it hit their numbers. Magnite: TTD "prioritized OpenPath as a default path for supply" and it flowed into a softer DV+ guide, though Magnite gamely says it "support[s] Trade Desk's goal of cleaning up the ecosystem" [80][81], and by Q1 2026 shrugs that "the OpenPath extinction event came and went, and we are still here" [82]. PubMatic first hid it as "a top DSP buyer, which recently shifted a significant number of clients to a new platform" before naming "Kokai" as evaluating media differently [83][84] — yet it also became "the first SSP to integrate the Trade Desk's price discovery" API [85]. Viant is the loudest bear, lumping TTD in with the walled gardens: "The Trade Desk direct spend through OpenPath, all to extract more margin for themselves" [86], and reframes the market as "Viant and The Trade Desk as the two remaining independent" enterprise buying platforms [87] — while an analyst notes Viant grew 18% to TTD's 12% [88]. Verdict: consensus that OpenPath is real and material — a point TTD's own framing understates.

Macro and AI — where the industry SPLITS

On the ad cycle the peers do not agree, which should temper how much you read into TTD's caution. Viant sees "a healthy ad environment" accelerating into 2026 [89]; Criteo cut its guide on "reduced spend from certain large U.S. Performance Media clients" [90]; Magnite lands in the middle — "It is not booming, but we are not prognosticating doom and gloom either" [91]; and the walled gardens see no weakness at all. On AI, the industry consensus is that agentic buying is the next battleground — but here TTD is behind on messaging: Criteo is already "OpenAI's first ad tech partner" [92], Google touts the "fastest-growing AI-powered search ads product" and a new "Universal Commerce Protocol" [93][94], and PubMatic predicts "By 2028… 25% of all digital advertising to be executed autonomously" [95]. TTD's agentic story (Stagwell, LLM-search TAM) is real but earlier-stage in the telling.

What to watch on the next calls

The calls leave four questions unresolved, and each is answerable on the tape ahead:

• Does the deceleration bottom, or break 10%? The Q2 2026 floor is "at least $750 million" [96]; a re-acceleration would validate the reinvestment, another step-down would confirm structural pressure.
• Does the 30% margin recover, or is high-40s gone? Management guided FY2026 margin "approximately in line with 2025" — the quarterly path will show whether Q1 2026 was timing or a new base.
• Do the CPG and auto verticals — "over a quarter of the business" — come back? They are the stated swing factor between 12% and mid-teens growth.
• Does the Amazon-DSP outlier view hold? TTD stands alone against its entire peer set in calling Amazon a non-threat; the next few quarters of relative growth will settle who is right.

The honest read across five years of calls: a genuinely dominant open-internet DSP with an unbroken CTV tailwind that the whole industry corroborates — now working through its first real stumble, decelerating faster than its independent-DSP peer, reinvesting margin, absorbing governance turnover, and defending against a big-tech competitor that every other company in the ecosystem takes more seriously than management does.

---

The Financials: A Cash Machine the Market Priced for Failure

The Trade Desk is one of the rare companies where the operating numbers and the stock price are telling opposite stories. In FY2025 revenue grew 18% to $2.90 billion, operating margin widened to 20%, and free cash flow hit a record $796 million [1]. Over the same window the shares fell from an all-time high of about $140 (December 2024) to roughly $20 (July 2026) — an ~86% drawdown. The entire investment debate for this business now lives in that gap: the market has re-rated a hyper-growth compounder down to a value multiple, and your job is to decide whether the de-rating is justified by a permanent slowdown or is an overshoot on a still-elite franchise.

This page is built to answer one question: does The Trade Desk have the financial quality, balance-sheet strength, and cash generation to justify — or to overturn — how the market now prices it? The short answer is that the quality is not in question; the growth trajectory is. Every material number below is clickable to the exact page of the filing or transcript that proves it.

A note on terms used throughout: DSP (demand-side platform) — the software advertisers use to buy digital ad space; gross spend — the total dollars flowing across the platform (TTD keeps a take-rate percentage of this as revenue); SBC — stock-based compensation, a real but non-cash expense; FCF — free cash flow, operating cash flow minus capital expenditure. Each is defined again where it first matters.

The 30-Second Read

Sources: FY2025 10-K, MD&A Executive Summary and Statements of Operations [2][3]; EV/Sales derived from reported financials and market price.

Why Valuation Comes First: The De-Rating Is the Story

For most companies you build to valuation last. For The Trade Desk in 2026 you must start there, because the collapse in the multiple — not any deterioration in the financials — is what has to be explained.

The Trade Desk earns money as an independent, objective buying platform: advertisers and agencies run campaigns across connected TV (CTV), video, display, and audio through its software, and TTD keeps a percentage of the spend. It takes no inventory of its own, which management argues is the source of its trust advantage over media-owning rivals like Google and Amazon [4]. That model historically commanded a premium multiple. It no longer does.

Source: derived from reported revenue and free cash flow [5] and year-end market prices; company filings, as reported.

At the 2024 year-end the market paid ~24x sales and ~92x free cash flow for The Trade Desk. Today it pays roughly 3.3x sales and 12x free cash flow — value-stock multiples for a business still compounding revenue at double digits with expanding margins. On earnings, the stock trades at about 22x trailing GAAP EPS of $0.90 and ~11x consensus FY2026 adjusted EPS of ~$1.85. Enterprise value (market cap of ~$9.6 billion less $1.3 billion of net cash) sits near $8.3 billion, or ~7x FY2025 adjusted EBITDA of $1.196 billion [6].

The point is not that the stock is "cheap." It is that the valuation now embeds a pessimistic view of growth. The compression from 47x sales (2020) to 3x (2026) is one of the sharpest de-ratings in large-cap software. Whether that is deserved turns entirely on the growth question — which is the next section.

Growth: Still Strong, But Visibly Decelerating

The Trade Desk is not a broken grower — it is a slowing one, and the deceleration is the crux the market is reacting to.

Source: FY2025 10-K, Statements of Operations and prior 10-Ks [7]; as reported.

Revenue has more than quadrupled since FY2019, a 5-year compound growth rate of about 28%. But the annual pace has stepped down — 26% in 2024 to 18% in 2025 — and the quarterly trend tells the sharper story.

Source: quarterly results per the FY2024 and FY2026 earnings calls [8][9]; as reported.

The inflection point was Q4 FY2024, when The Trade Desk delivered $741 million (+22%) but for the first time in 33 quarters as a public company fell short of its own guidance — a miss management attributed not to competition or a smaller market, but to a "series of small execution missteps" during its largest-ever reorganization and the migration of clients from its Solimar platform to the newer Kokai platform [10]. Growth then stepped down every quarter through Q1 FY2026, when revenue of $689 million grew just 12% [11]. Management frames the recent softness as cyclical — pressure in consumer-packaged-goods and auto verticals, tariffs, and geopolitical uncertainty weighing on the Fortune 500 advertisers that dominate its revenue base — rather than structural [12]. For Q2 FY2026 it guided to revenue of at least $750 million and adjusted EBITDA of about $260 million, with full-year adjusted EBITDA margin held at "at least 40%" [13].

Growth quality is genuinely high on the metrics that matter for durability: a client-retention rate that has exceeded 95% for over a decade, revenue earned under ongoing master service agreements rather than one-off insertion orders [14]. Gross spend on the platform reached $13.4 billion in FY2025, up 11%, and the take-rate TTD keeps as revenue has stayed within a consistent historical band [15][16]. Revenue that grows faster than gross spend, as it did in 2025 (18% vs 11%), signals rising monetization from value-added data and services — a positive mix shift. The U.S. still accounts for ~82% of revenue, leaving international (~18%) as the larger structural runway [17].

Earnings Quality: Cash Is the Truth; GAAP EPS Is Noisy

This is the single most important thing to understand about The Trade Desk's income statement: reported net income is a poor guide to the business, and free cash flow is the honest one. Two distortions — taxes and stock-based compensation — make GAAP earnings lurch around while cash generation stays remarkably steady.

Source: FY2025 10-K, Statements of Cash Flows, and prior filings [18]; as reported.

Look at FY2022: GAAP net income collapsed to $53 million (a 3% net margin), yet the company generated $465 million of free cash flow. The culprit was a 58% effective tax rate, driven by non-deductible stock compensation and other book-vs-tax differences, not any operating problem. In FY2020 the opposite happened — a large tax benefit inflated net income to $242 million against $331 million of FCF. The FY2025 effective tax rate of 33% again sat well above the statutory rate, holding GAAP EPS to $0.90 even as pre-tax income reached $659 million [19]. The takeaway: judge this company on cash and operating income, not headline EPS.

On the cash side the story is unambiguously strong. Operating cash flow grew 34% to $992.7 million and free cash flow reached a record $795.7 million in FY2025 — a 27% FCF margin, sustained above 26% for four straight years [20][21]. Free cash flow ran at ~1.8x net income in FY2025, a healthy conversion ratio.

The one honest asterisk is stock-based compensation, the largest gap between GAAP and cash earnings. SBC was $490.6 million in FY2025 — roughly $109 million a quarter — and has been flat near that level for three years [22].

Source: FY2025 10-K, Statements of Cash Flows and Note 10 (Stock-Based Compensation) [23][24].

Because revenue keeps rising while SBC dollars hold flat, SBC has fallen from 25% of revenue in FY2023 to 17% in FY2025 — the definition of operating leverage on the most dilutive line item. That is exactly the trajectory a maturing software business should show, and it is a meaningful quality improvement, not a red flag. It also explains the buyback strategy in the next section: the company repurchases stock largely to neutralize the dilution SBC creates.

The Balance Sheet: A Fortress Wrapped Around a Working-Capital Engine

The Trade Desk carries no debt whatsoever. As of December 31, 2025 it held $658 million of cash and $645 million of short-term investments — about $1.3 billion of net cash — and its $450 million revolving credit facility was completely undrawn, with $445 million of availability and full covenant compliance [25]. The one covenant that would ever bind — a maximum funded-debt-to-EBITDA ratio of 3.50 to 1.00 — is irrelevant while the company borrows nothing [26].

But the balance sheet looks far larger than the equity, and understanding why is essential to reading this business correctly. Total assets of $6.15 billion are dominated by $3.77 billion of accounts receivable, matched against $3.01 billion of accounts payable [27]. This is the DSP "gross-up": TTD is billed by advertisers for the full media spend and in turn owes publishers for that inventory. The receivables and payables largely offset, and the reported $3.67 billion of total liabilities is overwhelmingly trade payables to media sellers — not borrowings. So conventional leverage framing (net debt / EBITDA) is meaningless here; what matters is that the company self-funds and the payables cycle is a source of float, not a claim on shareholders.

Source: FY2025 10-K, Balance Sheets and Liquidity discussion [28][29].

The current ratio of 1.6x and the absence of any maturity wall mean liquidity risk is effectively nil. The balance sheet is a weapon — it lets management buy back stock aggressively through a downturn without touching the business. The one concentration to note: if client relationships are aggregated at the holding-company level, two advertising holding companies each represented more than 10% of gross billings in 2025, so a change in either agency relationship would matter [30].

Capital Allocation: Buying Back Stock — Belatedly at the Right Price

With no debt to service and no dividend, essentially all of The Trade Desk's cash goes to one place: repurchasing its own shares to offset SBC dilution. The scale stepped up dramatically in FY2025.

Source: FY2025 10-K, Statements of Cash Flows and Note 9 (Capitalization) [31][32].

The board first authorized a $700 million repurchase program in February 2023 and has topped it up repeatedly — most recently a further $350 million in February 2026, bringing available authorization to $500 million [33]. In FY2025 the company retired 26.2 million shares for $1.4 billion [34], and continued with $164 million in Q1 FY2026 [35].

There are two ways to read this. The critical read: much of the $1.4 billion spent in 2025 was at an average price near $53 — well above today's ~$20 — so shareholders effectively overpaid to offset dilution, and share count has stayed roughly flat (~493-500 million) rather than shrinking. The constructive read: the program is explicitly designed to neutralize employee-stock dilution, and with the stock now trading at ~12x FCF, every dollar of repurchase is far more accretive than it was a year ago. Either way, capital allocation is disciplined and entirely self-funded — there is no empire-building, no leverage, and only three small acquisitions in the company's history. The swing factor is not whether to return cash but whether management shifts from dilution-offset to genuine per-share shrinkage now that the price is low.

The Year-Wise Statements, in One Place

The full multi-year record. Margins and returns are computed from the same filings; note the tax-driven noise in FY2020-2022 net income discussed above.

Source: FY2025 10-K and prior 10-Ks — Statements of Operations, Cash Flows, and Balance Sheets [36][37][38]; gross profit, margins and ROE derived from reported financials. TTD reports cost as "platform operations"; gross profit is revenue less that line.

A note for the beginner: TTD does not label a "gross profit" line. It expenses "platform operations" (data-center, hosting, and delivery costs — $619 million in FY2025), so gross profit above is revenue minus that cost, implying a ~79% gross margin. Return on equity (net income / shareholders' equity) recovered to 18% in FY2025 as margins expanded; it looks modest for a software business mainly because the equity base is inflated by the receivables/payables gross-up, and because tax noise suppresses the numerator.

The operating-margin trajectory is the clean signal of the business maturing: after dipping to 7% in the FY2022 slowdown, it has climbed to 20% as revenue outgrew the cost base [39].

Source: derived from reported financials, FY2019-FY2025 10-Ks [40].

Peer Positioning: The Biggest Independent — and the Walled-Garden Threat

The Trade Desk's competitive set splits in two, and the financials look completely different across the divide. On one side are the independent ad-tech peers — pure-play buy-side and sell-side platforms. On the other are the walled gardens, Google and Amazon, which dwarf TTD in scale and, critically, own the media inventory TTD does not.

Sources: TTD revenue per FY2025 10-K [41]; peer market caps per latest market data, as reported. (The screen's "CRTO" entry resolved to a French bank, not Criteo, and is excluded.)

The takeaway from this table is that among independent ad-tech companies, The Trade Desk is in a class of its own — larger than Magnite, Viant, and PubMatic combined several times over, and the only one that is durably, substantially GAAP-profitable with a 27% free-cash-flow margin. That scale-and-profitability gap is the real moat in the numbers.

The genuine financial threat is not the small independents but Amazon, whose DSP leverages first-party retail data and whose balance sheet is effectively unlimited. Management's own defense is structural rather than financial: as an objective platform that owns no media, TTD argues it avoids the conflict of interest that comes with Amazon "soliciting ad budgets while competing against numerous Fortune 500 companies" [42]. For a financials reader, the point is that the de-rating partly reflects the market pricing in Amazon's competitive pressure — a risk to future growth, not to current cash generation.

What the Financials Say — and the One Metric That Decides It

What the numbers confirm: This is a high-quality, self-funding franchise. Gross margins near 79%, operating margins at 20% and rising, a record $796 million of free cash flow at a 27% margin, zero debt, $1.3 billion of net cash, and SBC falling as a share of revenue. On quality and balance-sheet strength alone, The Trade Desk is among the best-positioned companies in its sector.

What the numbers contradict: The market's implied verdict. At ~3x sales and ~12x free cash flow, the stock is priced as though growth is structurally impaired — yet the business is still compounding revenue at 12-18%, converting earnings to cash at 1.8x net income, and holding a 40%+ adjusted-EBITDA margin. The valuation and the fundamentals cannot both be right; the gap is the opportunity and the risk.

The swing factor is not margins, leverage, or cash — all of which are excellent and unlikely to break. It is the durability and re-acceleration of revenue growth. The entire de-rating is a growth story, so the financials that decide the stock are the ones that reveal whether 12% is a cyclical trough (as management insists) or the front edge of permanent deceleration under Amazon's pressure. Everything else is already strong enough.

The first financial metric to watch is year-over-year revenue growth (and its driver, gross spend on the platform). A re-acceleration back toward the high-teens — confirming the Q1 FY2026 slowdown to 12% was cyclical — would validate the quality already visible in the cash flows and make today's ~3x-sales multiple look like a mistake. A further slide toward high-single-digits would confirm the market's structural-slowdown thesis and justify the de-rating, no matter how strong the margins and balance sheet remain.

---

Web Research — What the Internet Knows, With Receipts

Bottom line. The filings show a company that still grew revenue 18% in FY2025 to $2.90B with $992.7M of operating cash flow [1] — a picture of health. The web tells a completely different story: TTD has lost roughly 86% of its value since December 2024, and the reason is not the reported income statement but a collapse of narrative control. Growth is decelerating hard (Q1 2026 revenue grew just 12%, Q2 guided to ~10%), a federal judge refused to dismiss an insider-trading class action against CEO Jeff Green and two other executives, the CFO was fired after five months amid serial C-suite and board turnover, the largest agency holding company (Publicis) briefly told clients to stop spending on the platform, and Amazon is visibly taking connected-TV share. The single most important thing the web reveals that the filings do not: the market has concluded TTD's growth algorithm and moat are breaking, and the litigation/governance overhang is now a live, unpriced tail risk — not a footnote.

Sources: price and consensus target per market data as of 2026-07-10 (analyst mean target, 37 analysts); FY2025 revenue per FY2025 Annual Report [2].

1. The 86% de-rating is the whole story — and growth deceleration is the engine

Headline: TTD fell from $139.51 (2024-12-04) to $19.53 (2026-07-10) — an 86% drawdown — as revenue growth collapsed from ~25% to ~10% in six quarters.

The stock has crashed in three discrete steps, each on a guidance problem rather than a reported miss: -41% in February 2025 after its first-ever guidance miss (Nasdaq); -38% in August 2025 when Q2 revenue of $694M (up 19%) came with decelerating guidance (Digiday, Pacific Coast Business Times); and a further slide after Q3 2025 (revenue up 18% to $739M) again dipped on cautious Q4 guidance (Investing.com). Q1 2026 then grew only 12% to $689M with Adjusted EBITDA margin compressing to 30% from 34%, and Q2 2026 was guided to ~$750M — below the ~$771M Street expectation (TIKR, StockTitan). The company still generated $2.90B of FY2025 revenue and $443M of net income [3] — this is a multiple de-rating, not an earnings collapse.

Source: daily price series, as reported (Fiscal v2 feed); waypoints annotated from dated news events cited in this section.

Sources: quarterly YoY growth per company results releases and Q2-2026 guidance (Digiday; TIKR); FY2025 revenue of $2,896.3M was up 18%, per the FY2025 Annual Report [4].

So-what for the stock: This resets the entire thesis. TTD's premium multiple was built on 20%+ durable growth; at ~10% growth it is a different, lower-multiple company. What's priced in: most of the direction is in the price — the stock is down 86% and now trades at ~$19.5 against a ~$24.4 consensus mean target (37 analysts, "Hold"), i.e. the Street already models a re-rate to only modestly higher. The open, unresolved question the market has not settled is whether ~10% is the trough or a way-station lower — that single variable, not any of the discrete controversies below, is the swing factor.

2. Red flag: the insider-trading class action survived — motion to dismiss DENIED in full

Headline: On 2026-03-15 a federal judge denied TTD's motion to dismiss the securities class action in its entirety, letting a Section 20A insider-trading claim against CEO Jeff Green and two other executives proceed to discovery.

Judge Christina A. Snyder (C.D. Cal.) denied the motion to dismiss the amended complaint in full (Cohen Milstein). Plaintiffs allege TTD overstated how quickly customers adopted Kokai — its major platform upgrade — while named executives sold more than $465 million of stock at allegedly inflated prices. Critically, the primary record confirms this is not fringe: TTD's own FY2025 10-K discloses that the First Amended Complaint "asserts a separate claim under Section 20A alleging that the Company's Chief Executive Officer, then-Chief Financial Officer and Chief Strategy Officer engaged in insider trading during the proposed class period" [5]. Parallel shareholder-derivative suits (Silva, Jong, Gardner, Ibrahim v. Green et al.) are stayed pending the securities case [6].

So-what for the stock: A denied motion to dismiss materially raises the odds of a costly settlement and — more damaging — hands plaintiffs discovery into internal Kokai-adoption data at the exact moment the growth story is in question. It caps the multiple and injects headline risk into every quarter until resolved. What's priced in: the drop from the underlying business is well known, but the market treats the litigation as background noise rather than a discrete, quantifiable liability — this is where the edge sits. The insider-selling allegation (executives took $465M off the table on the way up) is also the uncomfortable counterweight to Green's April buy in Finding 6. Red flag.

3. Red flag: C-suite and board turmoil — CFO fired after five months, directors out, Nasdaq non-compliance

Headline: TTD terminated CFO Alex Kayyal on 2026-01-24 after barely five months, capping serial CFO churn, and two directors resigned in March 2026, tipping the company into Nasdaq committee-independence non-compliance.

Kayyal (hired August 2025) was terminated effective 2026-01-24; Chief Accounting Officer Tahnil Davis became interim CFO (Adweek; Yahoo Finance). This follows the earlier exit of long-time CFO Laura Schenkein — three CFOs inside roughly a year. Kayyal then resigned from the board (2026-03-19) and director Kathryn E. Falberg resigned (2026-03-23), leaving TTD non-compliant with Nasdaq audit- and compensation-committee independence rules; Nasdaq granted a cure period to 2026-09-21, and the board appointed Drew Vollero as a director and audit-committee member (StockTitan).

So-what for the stock: Finance-function instability is the worst possible signal in the middle of a growth scare and an active accounting-adjacent lawsuit — it feeds the bear narrative that something is wrong behind the numbers and weakens the "trust management" leg of any bull case. The Nasdaq cure clock (Sept 2026) is a concrete, datable overhang. What's priced in: partly — the CFO firing coincided with a 52-week low (Chronicle Journal) — but the market has not fully connected the governance instability to the litigation risk in Finding 2. Red flag.

4. Amazon is taking CTV share — the moat challenge is now visible in the numbers

Headline: Amazon's ad business grew 23% in Q4 2025 versus TTD's 14%, and premium streaming inventory is increasingly routed through Amazon DSP and Google DV360, bypassing the neutral DSP.

Amazon credits deeper CTV partnerships (Roku, Disney) and first-party data across Prime Video, Fire TV and live sports, and its sales teams have reportedly offered cut-rate CTV deals to pull spend off TTD (AdExchanger). Some advertisers now find ~30% of their target streaming inventory is accessible only via Amazon DSP or DV360 (Digiday). The counter-view: TTD remains the dominant independent DSP and Amazon's threat, while real, is "still distant" (Digiday).

So-what for the stock: This is the structural core of the bear case — if premium CTV supply fragments toward walled gardens, TTD's "neutral, objective" positioning erodes and its take rate faces pressure. CTV is TTD's most important growth channel, so share loss here hits the highest-value part of the mix. What's priced in: this is now consensus among ad-tech reporters and is a major reason for the de-rate, so it is partly priced — but the magnitude of share shift is unresolved, and the market lacks hard TTD-specific CTV-share data to size it. Red flag (structural).

5. The Publicis boycott and truce — a live demonstration of client-concentration fragility

Headline: In March 2026 Publicis pulled TTD from its recommended-DSP list and told clients to stop spending after an audit alleged improper fee-stacking; the two settled in June 2026 on undisclosed terms.

Publicis alleged TTD stacked its ad-tech fee on top of other charges in a way its contract did not support, removed TTD from its recommended list, and advised clients to halt spend (Ad Age). The dispute was resolved around 2026-06-12–15 with Publicis resuming its recommendation; the joint statement disclosed no terms (MediaPost; Digiday). CEO Green publicly moved to reset agency relationships afterward (Adweek).

So-what for the stock: The episode is a real-time proof of the 10-K's own risk that "the loss of advertising agencies… as clients could significantly harm our business" — TTD's revenue runs through a concentrated set of holding companies, and one of them unilaterally cut spend for months. The settlement removes the acute overhang (a modest relief rally followed), but the fee-transparency question it raised — how TTD stacks platform fees — dovetails uncomfortably with the take-rate and litigation threads. What's priced in: the resolution is priced (stock bounced on the June news); the lingering, less-priced question is whether other holdcos press the same fee-transparency point. Neutral-to-negative.

6. Positive: Jeff Green bought ~$150M of stock — but read it against the $465M sold on the way up

Headline: In April 2026, CEO Jeff Green made an open-market purchase of roughly 6.4 million shares (~$150M) after the stock had already fallen more than 80%.

Green's buy is the clearest insider confidence signal in years and frames his bet that TTD becomes "AI-era infrastructure" (Motley Fool; Adweek). It also lands weeks after TTD launched Koa Agents — agentic AI for media planning/buying — with Stagwell as the first pilot partner (Ad Age).

So-what for the stock: A $150M personal buy at ~$23 is a genuine floor-signal and aligns the CEO with minority holders — positive for sentiment and for the "capitulation bottom" case. But it cannot be read in isolation: the same executive team is alleged in the surviving securities suit (Finding 2) to have sold over $465M near the highs. The honest read is a controlling founder averaging back in cheap, not a clean vote of confidence. What's priced in: the buy produced a short-lived pop but did not arrest the slide (stock is lower since), so the market has largely discounted it. Positive, with an asterisk.

7. Optionality: the Google ad-tech antitrust remedy could reshape the pool in TTD's favor

Headline: Judge Brinkema has ruled Google unlawfully monopolized ad exchanges and publisher ad servers; the remedy ruling — potentially forcing a sale of AdX — is pending after November 2025 closing arguments.

The DOJ is pushing for structural divestiture of Google's AdX exchange and open-sourcing of its auction logic; Google is offering behavioral fixes. Judge Brinkema has signaled skepticism of a forced breakup, partly because no buyer for AdX has emerged (AdExchanger; National Law Review).

So-what for the stock: As the largest independent DSP, TTD is a structural beneficiary of anything that weakens Google's grip on the buy-through-to-exchange stack — a genuine, under-discussed source of asymmetric upside. What's priced in: almost nothing — with the ruling pending and the judge leaning behavioral, the market is not underwriting a favorable outcome. This is free optionality on the long side, and a catalyst to watch. Positive (optionality).

8. Background: the dual-class super-vote and Nevada reincorporation

Headline: Green and insiders control ~49.9% of voting power via 10-vote Class B shares that do not sunset until December 2035, and TTD reincorporated to Nevada in late 2024 over a stockholder challenge.

As of 2025-12-31, Class B holders (executives, employees, directors and affiliates) held ~49.9% of voting power, with Class B converting to Class A only on 2035-12-22 [7]. The 2024 Delaware-to-Nevada reincorporation was challenged in Gunderson v. Trade Desk; the court granted summary judgment for the company on the supermajority question in November 2024, though residual fiduciary-duty claims continue (Bloomberg Law; Harvard Corp Gov).

So-what for the stock: Concentrated founder control plus a move to a less shareholder-friendly jurisdiction limits minority recourse and is a structural discount factor — but it is long-standing, already known, and not a new catalyst. Neutral/background.

Recent-news reference layer

The table below is the reference material behind the findings above — the most thesis-relevant items, most recent first. Still-live threads (the securities suit, Amazon competition, the Nasdaq cure clock) are retained regardless of age.

Sources: indexed news corpus (news.pdf, 30-item feed dated 2026-07-11) and first-hand web searches; individual URLs are cited inline in the findings above.

Coverage of specialist questions — importance-first reference grid

The thesis-changing specialist questions are already promoted into the findings above (litigation → #2; CFO/board → #3; Amazon → #4; Publicis → #5; take-rate/fees → #4–5; Google antitrust → #7). The remainder are answered concisely here.

What remains genuinely unresolved

Even after the searches, four questions sit at the center of the PM's uncertainty and no source settles them: (1) is ~10% revenue growth the trough or a way-station lower; (2) how large is the settlement/discovery exposure from the Section 20A case; (3) how much CTV share is Amazon actually taking from TTD specifically; and (4) does the Google remedy land structurally (a real tailwind) or behaviorally (a non-event). These are carried into the research-queries file.

---

Short Interest & Thesis — The Trade Desk (TTD)

Bottom line. Official reported short interest for TTD is not available in the staged data — the FINRA equity short-interest feed returned zero position rows, and no short-sale-volume, borrow-cost, public net-short, or peer-crowding rows were staged either. So this page is not a squeeze/crowding read; it is a thesis-risk assessment built from the primary record. And that record is unusually rich: TTD has lost roughly 86% of its value in eighteen months on repeated 30–40% earnings gap-downs, and the disclosed short case is not a forensic-fraud case (the company is profitable, net-cash, and free-cash-flow generative) but a governance + competition + de-rating case — founder super-voting control just under a majority [1], a pending securities class action that now adds a Section 20A insider-trading claim against the CEO, CFO and CSO [2], the resignation of all three audit-committee members in early 2026 [3], and an Amazon-led walled-garden competitive threat [4].

What is (and isn't) in the data

The short-interest data step staged the file scaffold but every position table came back empty.

Source: reported short-interest / short-sale-volume / borrow / peer figures as staged (all empty); narrative from TTD SEC filings and transcripts.

Because there is no official short position, no legitimate "days-to-cover" or "% of float short" can be computed. Anyone quoting a TTD short-interest number here would be inventing it. The honest institutional answer on positioning is: not decision-useful in this dataset. The decision-useful content is the thesis-risk ledger and the tape.

Tape and liquidity context (inferred — not short interest)

What the price series can tell us is that TTD trades like a crowded, negative-momentum name: a violent, multi-quarter de-rating on rising volume.

Source: staged daily price series (data/prices/), month-open close, split-adjusted; not a filing figure.

The stock peaked near $135 (Dec 2024) and trades near $19 (Jul 2026) — roughly an 86% drawdown. This is the tape a short thesis would point to as vindication. Note the mechanics: TTD reports and then gaps, hard, in both directions. Two of the largest single-day moves in the company's history were earnings-driven collapses.

Source: staged daily price series (data/prices/); percentage moves derived from reported closes — not a filing figure.

Read for a PM: the same catalyst sensitivity that produced −33% (Feb 2025) and −39% (Aug 2025) also produced +33% and +36% up-gaps in prior beats. A short here carries real positive-catalyst / squeeze-into-a-beat risk even without a confirmed crowded position — the name is a coiled spring around every quarterly print. Liquidity, however, is deep: recent volume has averaged roughly 19 million shares/day against ~494 million shares outstanding, so an exit is not the constraint a short would worry about.

Source: shares outstanding from FY2025 10-K financial data (data/financials/); ADV derived from staged daily price series (data/prices/).

The short-thesis ledger (grounded in TTD's own disclosures)

There is no staged short-seller report, so this ledger is assembled from the company's own filings — the disclosures a short would build a case on, each paired with management's framing where it exists. This is the substance of the page.

Sources: FY2025 10-K risk factors and Note 13 [5] [6]; proxy [7]; Q4 FY2024 transcript [8].

Governance: the strongest, most durable leg

Insiders control roughly 49.9% of total voting power through the 10-votes-per-share Class B structure — "this concentrated control limits or precludes your ability to influence corporate matters" in the company's own words [9]. The proxy shows CEO Jeff Green holding 42,071,879 Class B shares across his trusts, the source of that voting block [10]. Layered on top: the November 2024 reincorporation to Nevada, whose interested-stockholder statute restricts 10%+ holders and whose forum provisions channel disputes away from Delaware Chancery [11]. A short frames this as governance that tilts hard toward the founder and away from minority holders.

The sharpest recent signal is the audit committee: all three members — Kathryn Falberg, Lise Buyer and Gokul Rajaram — resigned from the board within two weeks of each other in late March / early April 2026 [12]. Three simultaneous audit-committee departures is exactly the sort of event a short thesis elevates, whatever the benign explanations may be.

The CEO performance option — a 2021 ten-year grant vesting on share-price targets up to $340 — was itself litigated in the Huizenga v. Green derivative suit alleging the board breached its duties in approving it; the dismissal was ultimately affirmed on appeal in November 2025 [13] [14]. That one resolved in the company's favor; the open item is the securities case.

The litigation: an unresolved Section 20A overhang

Following the February 2025 stock collapse, securities class actions were filed [15], and the consolidated amended complaint now asserts a separate Section 20A claim that the CEO, then-CFO and Chief Strategy Officer "engaged in insider trading during the proposed class period," on top of the Section 10(b) fraud-on-the-market claims [16]. This is unresolved (motion to dismiss pending) and is the single most citable "hard" short-thesis hook — it links insider selling to the very disclosures that preceded an 80%+ decline. It is an allegation, not a finding; the company contests it.

The business case: competition and a self-inflicted miss

TTD concedes that walled-garden inventory providers "may exclusively sell their own inventory directly to advertisers, which prevents us from competing with them entirely for such inventory" — the structural Amazon/Google/Meta risk [17]. Client stickiness is thin on paper: master service agreements carry "no material commitments" and are "terminable at any time upon 60 days' notice" [18], and concentration is real — two agency holding companies each represented more than 10% of gross billings in 2025 [19].

Crucially, management handed the short case its narrative. On the Q4 FY2024 call the CEO acknowledged that "for the first time in 33 quarters as a public company we fell short of our expectations," attributing it to "a series of small execution missteps" during the Kokai platform migration — self-inflicted, not competitive [20]. Growth then kept slowing: Q3 FY2025 was guided to "at least $717 million, reflecting 14% year-over-year growth," with explicit tariff/macro caveats around large global brands [21].

Source: FY2025 10-K financial data, revenue by year (data/financials/); growth rates derived. Guidance quote at Q2 FY2025 transcript [22].

The counter to a forensic short: the balance sheet doesn't cooperate

The reason this is a governance-and-multiple short rather than a fraud-and-solvency short: TTD's financials are strong. It is profitable, throws off cash, and holds no meaningful debt — none of the going-concern, covenant, or burn stress a classic activist short leans on.

Source: FY2025 10-K financial data — income, cash flow, balance sheet (data/financials/).

That cash funds a large buyback, which the company runs partly to offset equity-comp dilution — but it warns in its own risk factors that repurchases "cannot guarantee… that it will successfully mitigate the dilutive effect of employee equity awards" and that they "diminish our cash reserves" [23]. Share count has stayed roughly flat despite heavy issuance because the buyback ran hot — $1.4B in FY2025.

Source: FY2025 10-K financial data, cash-flow financing detail (data/financials/); dilution-offset caveat at FY2025 10-K [24].

Peer context

Unavailable. No peer short-interest rows were staged, so no cross-name crowding comparison is possible. For reference, TTD's own filings frame its competitive set as the independent DSPs and, more pressingly, the walled gardens (Amazon, Google, Meta) [25]; the indexed peer set includes Viant (DSP), Criteo (CRTO) and Amazon (AMZN). A relative-crowding read would require reported short interest for both TTD and those peers, which this run does not have.

Evidence quality

Source: staged data/short_interest/ manifest (all position tables empty); narrative items cite the TTD filings referenced above.