Programmatic advertising is the automated buying and selling of digital ad impressions through real-time auctions. Instead of a human sales team negotiating insertion orders, software platforms bid on individual ad slots in milliseconds. Three distinct layers of infrastructure make this work: demand-side platforms (DSPs) that buy impressions on behalf of advertisers, supply-side platforms (SSPs) that sell publisher inventory to the highest bidder, and verification/measurement companies that confirm ads were actually seen by real humans in brand-safe environments. In the US, programmatic channels handled roughly $180 billion of display ad spend in 2025 and are forecast to exceed $200 billion in 2026 — approximately 92% of all digital display advertising. est.
Advertisers send dollars through DSPs. DSPs bid in real-time auctions run by SSPs. SSPs fill publisher ad slots and keep a percentage. Verification firms sit alongside, confirming impressions were viewable, fraud-free, and brand-safe. Each layer takes a cut of the same underlying ad dollar — the unit of product is one served, measured impression.
A DSP lets advertisers (or their agencies) set campaign goals — target audience, budget, desired outcomes — and automatically bids on individual ad impressions across thousands of publishers simultaneously. The DSP evaluates each impression opportunity in under 100 milliseconds, decides whether to bid and at what price, and executes the purchase. The Trade Desk (TTD) is the largest independent DSP. Google's DV360 and Amazon's DSP are major players but operate inside walled gardens (they preferentially serve inventory on their own properties).
Revenue model: TTD charges a percentage of gross ad spend that flows through its platform. In FY2024, $12 billion of gross ad spend passed through TTD and generated $2.4 billion of revenue — an implied take rate of roughly 20%. The advertiser pays $1.00 for an impression; approximately $0.20 goes to TTD and $0.80 flows to the sell side.
Source: TTD FY2024 earnings release ("a record $12 billion of spend on our platform"), TTD 10-K.
An SSP helps publishers (websites, apps, streaming services) sell their ad inventory by connecting to multiple DSPs and running auctions to maximize the price per impression. Magnite (MGNI) and PubMatic (PUBM) are the two largest independent SSPs. Magnite is the dominant SSP for connected TV (CTV) — Netflix, Disney+, and other major streamers route their ad-supported inventory through Magnite's SpringServe technology.
Revenue model: SSPs report revenue in two ways. MGNI reports gross revenue (including traffic acquisition costs paid to publishers) and separately discloses "contribution ex-TAC" — the portion it keeps after paying publishers. In FY2025, MGNI reported $714M in GAAP revenue and $670M contribution ex-TAC, implying it passed roughly $44M to publishers on top of the revenue it kept. PUBM reports on a net revenue basis (already excludes TAC), so its $283M FY2025 revenue is closer to contribution ex-TAC.
Source: MGNI FY2025 10-K, PUBM FY2025 10-K.
Verification companies tag ad impressions and independently confirm three things: (1) viewability — the ad was actually visible on screen, (2) brand safety — the ad did not appear next to harmful content, and (3) fraud detection — the impression was served to a real human, not a bot. DoubleVerify (DV) is the largest public player. Integral Ad Science (IAS) was the #2 player until Novacap acquired it for $1.9 billion ($10.30/share) in December 2025; IAS is now private.
Revenue model: DV charges per measured media transaction. In FY2025 it measured 9.5 trillion media transactions and generated $748M in revenue, roughly $0.08 per thousand transactions measured. Revenue splits into three lines: activation ($427M, 57%), measurement ($250M, 33%), and supply-side ($71M, 10%).
Source: DV FY2025 earnings release; IAS acquisition closing per IAS press release Dec 23, 2025.
Source: TTD Q1 2026 earnings; MGNI Q1 2026 8-K; PUBM Q1 2026 preliminary; DV FY2025 earnings.
Source: eMarketer programmatic forecasts (H1 2025, H1 2026); Adtelligent retail media outlook; MGNI Q1 2026 8-K.
Unlike physical infrastructure businesses, ad tech has no hard capacity ceiling. The "supply" is ad inventory — publisher web pages, app screens, and streaming video slots. This supply is growing: more websites, more apps, more streaming minutes create more impressions every year. est.
The real constraints are attention and quality, not raw inventory:
The binding constraint is market structure. Google operates both the dominant DSP (DV360) and a major SSP (Google Ad Manager), plus owns the largest pool of first-party data (Search, YouTube, Gmail). The DOJ antitrust case against Google's ad tech monopoly (filed Jan 2023, trial completed Nov 2024) could reshape competitive dynamics. If Google is forced to divest parts of its ad tech stack, independent DSPs and SSPs would gain addressable market share.
Demand for programmatic ad buying is growing at 10–15% annually. est. Supply of raw impressions is growing faster (AI-generated content proliferating). CPMs on open-web display inventory have been flat to declining for years as supply outpaces demand. The average programmatic display CPM in the US sits around $2–5. est.
The mix is shifting. CTV, retail media, and video — all higher-CPM channels — are growing at 20–30% annually est. and absorbing a rising share of total programmatic budgets. MGNI's revenue mix illustrates this: CTV crossed 51% of contribution ex-TAC in Q1 2026, up from 43% a year earlier, while desktop shrank from 17% to 15%.
The net effect on ad tech intermediaries: total dollar volume flowing through the pipes is rising, even as per-unit pricing on commodity inventory softens.
Verification demand is growing independently: as inventory volume increases and AI-generated fraud gets more sophisticated, the ratio of "impressions that need checking" to "total impressions" is rising. DV measured 9.5 trillion transactions in FY2025, up 15% YoY.
| Ticker | Company | Role | FY2025 Rev ($M) | Gross Margin | Adj. EBITDA Margin | FCF ($M) | Mkt Cap ($B) | Net Cash / (Debt) ($M) |
|---|---|---|---|---|---|---|---|---|
| TTD | The Trade Desk | DSP (buy side) | 2,896 | 79% | 35% | ~569 | 9.7 | +982 |
| MGNI | Magnite | SSP (sell side, CTV leader) | 714 | 63% | 35% | ~175 | 2.1 | (245) |
| PUBM | PubMatic | SSP (sell side, net rev basis) | 283 | 64% | 22% | 46 | 0.5 | +103 |
| DV | DoubleVerify | Verification | 748 | ~82% | 33% | 173 | 4.4 | +259 |
IAS (Integral Ad Science): Novacap completed its $1.9 billion ($10.30/share, all-cash) acquisition on December 23, 2025. IAS was delisted from Nasdaq. Last public quarterly filing was Q3 2025 (revenue $154M, +16% YoY; TTM revenue ~$591M; gross margin 77%; adj. EBITDA margin ~36%). Acquisition multiple: ~3.2x revenue, ~8.9x adj. EBITDA. No longer investable as a public equity.
Source: TTD FY2025 10-K; MGNI FY2025 10-K; PUBM FY2025 10-K; DV FY2025 earnings; IAS Q3 2025 10-Q; market data via yfinance Jun 2, 2026.
| Ticker | FY2022 | FY2023 | FY2024 | FY2025 | Q1 2026 | Q1 YoY |
|---|---|---|---|---|---|---|
| TTD | 1,578 | 1,946 | 2,445 | 2,896 | 689 | +12% |
| MGNI | 577 | 620 | 668 | 714 | 164 | +6% |
| PUBM | 256 | 267 | 291 | 283 | 63 | -12% |
| DV | 452 | 572 | 656 | 748 | — | — |
PUBM Q1 YoY decline reflects loss of a top DSP buyer and political spend comparisons. DV Q1 2026 not yet reported at time of writing.
| Ticker | Operating CF | Capex | FCF | SBC | FCF − SBC |
|---|---|---|---|---|---|
| TTD | ~739 | ~107 | ~632 | 495 | ~137 |
| MGNI | 236 | ~60 | ~175 | ~90 | ~85 |
| PUBM | 81 | 35 | 46 | ~47 | ~(1) |
| DV | ~235 | ~62 | 173 | ~119 | ~54 |
Source: FY2025 10-K filings. TTD FY2024 OCF/capex used where FY2025 annual figures not yet available. SBC from income statements. DV figures from FY2025 earnings. est. on DV OCF/capex split.
| Metric | Value |
|---|---|
| Share price (Jun 2, 2026) | $20.56 |
| Shares outstanding (basic) | ~475M |
| Market cap | $9.7B |
| Net cash (cash + ST investments − debt) | ~$982M |
| Enterprise value | ~$8.7B |
| TTM revenue (Q2'25–Q1'26) | $2,969M |
| EV / TTM revenue | 2.9x |
| TTM GAAP net income | $433M |
| P/E (trailing) | ~22x |
| Q2 2026 guided revenue | ≥$750M |
| Annualized Q2 guidance | $3,000M |
| Buyback remaining | $327M |
| FY2024 SBC as % of revenue | 20% |
TTD processed $12B of gross ad spend in FY2024 at a ~20% take rate. At $9.7B market cap, the market values TTD at roughly $0.81 per dollar of annual gross ad spend flowing through the platform — or 2.9x revenue net of the publisher pass-through.
Source: TTD Q1 2026 earnings; yfinance Jun 2, 2026.
| Metric | Value |
|---|---|
| Share price (Jun 2, 2026) | $14.45 |
| Shares outstanding (basic) | ~143M |
| Market cap | $2.1B |
| Net debt (debt − cash) | ~$245M |
| Enterprise value | ~$2.3B |
| FY2025 contribution ex-TAC | $670M |
| EV / contribution ex-TAC | 3.5x |
| FY2025 adj. EBITDA | $232M |
| EV / adj. EBITDA | ~10x |
| FY2025 GAAP net income | $145M (includes $90M one-time tax benefit) |
| Goodwill | $984M (47% of total assets) |
| $200M buyback authorized | Through Feb 2028 |
| FY2026 guided CxT growth | ≥11% |
MGNI's $984M of goodwill comes from its 2020 acquisition of SpotX (CTV SSP) for ~$560M. Tangible book value is roughly negative.
Source: MGNI FY2025 10-K, Q1 2026 8-K; yfinance Jun 2, 2026.
| Metric | Value |
|---|---|
| Share price (Jun 2, 2026) | $11.38 |
| Shares outstanding (basic) | ~38M |
| Market cap | $529M |
| Net cash (cash − debt) | ~$103M |
| Enterprise value | ~$426M |
| FY2025 revenue | $283M |
| EV / revenue | 1.5x |
| FY2025 adj. EBITDA | $62M |
| EV / adj. EBITDA | ~7x |
| FY2025 FCF | $46M |
| EV / FCF | ~9x |
| FY2025 GAAP net income | $(14.5)M (loss) |
| Goodwill | $30M (5% of assets) |
| Net dollar retention | 96% |
| Buyback remaining | $94M (~18% of mkt cap) |
PUBM's net cash of $103M is 19% of market cap. The company repurchased 8.1% of its diluted shares in FY2025. Revenue declined 3% in FY2025 (partly from losing volume from a top DSP buyer), and Q1 2026 revenue fell 12% YoY. Management guided to a return to double-digit growth in H2 2026. CTV revenue grew 50%+ (ex-political) in FY2025 but is not broken out in dollars.
Source: PUBM FY2025 10-K; Q1 2026 preliminary results; yfinance Jun 2, 2026.
| Metric | Value |
|---|---|
| Share price (Jun 2, 2026) | ~$25 |
| Market cap | ~$4.4B |
| Net cash | ~$259M |
| Enterprise value | ~$4.1B |
| FY2025 revenue | $748M |
| EV / revenue | 5.5x |
| FY2025 adj. EBITDA | $246M (33% margin) |
| EV / adj. EBITDA | ~17x |
| FY2025 FCF | $173M |
| EV / FCF | ~24x |
| FY2026 guided revenue | $810–826M |
| Net revenue retention | 109% |
IAS was acquired at ~3.2x revenue / ~8.9x adj. EBITDA. DV trades at a premium to that private-market comp on every metric.
Source: DV FY2025 earnings; yfinance Jun 2, 2026.
| Fact | Source | Confidence |
|---|---|---|
| TTD Q1 2026 financials, Q2 guidance, $12B gross spend FY2024 | TTD earnings releases (Q1 2026, FY2024) | filed |
| MGNI FY2025 and Q1 2026 financials, CxT breakdown, 2026 guidance | MGNI 10-K, Q1 2026 8-K | filed |
| PUBM FY2025 financials, Q1 2026 preliminary revenue | PUBM 10-K, preliminary Q1 2026 release | filed |
| IAS TTM revenue, acquisition price, delisting | IAS Q3 2025 10-Q, Novacap press release | filed |
| DV FY2025 financials, 2026 guidance, transactions measured | DV FY2025 earnings release | filed |
| US programmatic spend ~$180B (2025), >$200B (2026) | eMarketer forecasts (H1 2025, H1 2026) | est. |
| US CTV ad spend ~$30B (2024) | eMarketer via Xenoss | est. |
| US retail media ~$69B (2026), 84% Amazon+Walmart | Adtelligent citing eMarketer | est. |
| Programmatic = ~92% of digital display | eMarketer H1 2025 forecast | est. |
| CTV CPMs $25–45, display CPMs $2–12, ad fraud 10–15% | Industry estimates, not live-verified | est. |
| Demand growth 10–15% annually, CTV/video 20–30% | eMarketer directional forecasts | est. |
| Share prices, market caps | yfinance, Jun 2, 2026 | live |