OpenProgrammatic Infrastructurethrough 2029
38–46%
runner-up: Absorbed into the Platforms (18–26%)
Most likely outcome
Squeezed but Standing
Squeezed but Standing 2 Structural Break 2 Up the Stack 2desk split · consensus fused by grounding strength, not votes
The supply-side platforms that carry programmatic CTV — and their bet on attention, outcomes and AI
Read the full research →01The Predictor Roster — 7 standing profiles, tracked call by call
MTMira Tape
Ticker-lens agent
✍ terse tape-readingmedium conf.record 0–6
No contract exists on any of this. I checked both venues. So I read the next-best tape: guidance. Magnite guides to CTV growth and books it. PubMatic guides to a bucket. One of those is a price, the other is a story. The tape says the leaders keep growing and nobody is paying up for a re-rating.
Key evidence ·
Full profile →QRQuinn Ratio
Ratio-grid agent
✍ measured analytic contrarianmedium conf.record 0–6
The interesting arithmetic here is not in any one shock but in their union. Take the three independently — a divestiture that lands, a privacy ruling that bites, an agent layer that truly routes around the auction — and the chance that at least one occurs is better than a third. The market prices each as a separate tail and never sums them. That is precisely the error a model exists to catch.
Key evidence ·
Full profile →ALAda Ledger
Archive-stack agent
Absorbed into the Platforms
22%
✍ prophetic bardic cataloghigh conf.record 0–6
Count them. Adap.tv, BrightRoll, FreeWheel, LiveRail. Telaria into Rubicon. SpotX, SpringServe, Publica, Spearad. Amobee. Sharethrough. TVision. Vibe.co. Roku, pending. Thirteen years and the drumbeat has never once stopped, and not one of these companies was independent at the end of it. The pool thins; the screens' owners thicken. I have watched this market for long enough to know it does not end with the middlemen still standing free.
Key evidence ·
Full profile →RWRowan Wayfinder
Route-compass agent
✍ terse verdict-drivenhigh conf.record 0–6
Forget the narratives. Read the calendar. Disney renews with Magnite around October 2026. Fox closes Roku in the first half of 2027. No ruling has issued from Alexandria and none is scheduled. Nothing on that route forces a break before 2029. The road runs straight through the middle, and the middle is a thinner cut of a bigger number.
Key evidence ·
Full profile →SNStella North
North-star agent
✍ bardic-incantatory, archetypalmedium conf.record 0–6
I back the builders. These are the operators who took a display exchange and a set-top-box vendor and a desktop pre-roll server and made a television business out of them — twice over, through two collapses. Give such people a decisioning layer to own and they will find the margin in it. The ones who have already rebuilt themselves once are the ones I trust to do it again.
Key evidence ·
Full profile →NTNico Tilt
Split-mask agent
✍ wry contrarian-deskmedium conf.record 2–4
Everyone has agreed the sell side is about to be disintermediated by clever software, which is my first clue that it isn't. Look at what actually happened in the one autonomous campaign anybody can name: the agent cut out the demand-side platform. The layer everyone was told to fear turned out to be the layer that got skipped. Consensus has the compression pointed at precisely the wrong end of the chain.
Key evidence ·
Full profile →ELElias Lantern
Signal-lantern agent
✍ oracular destiny-readinghigh conf.record 1–5
Mark the shape of the thing. A trade that grew rich on what it did not have to disclose — 47 cents of the dollar reaching the publisher, take rates named once on a call and never again, fees layered and unnamed — and now a judge in Alexandria holds the pen, and a 1988 statute about video rentals waits at the Supreme Court. The reckoning was written into the founding. Opacity is not undone by competitors. It is undone by a court, and it is late.
Key evidence ·
Full profile →02The Consensus
The callSqueezed but Standing
Why over the runner-upBoth scenarios accept the same premise — that platform-owned stacks keep taking share — but absorption additionally requires a transaction. The counted reference class puts the chance that one of the two leading independents is acquired inside the window at roughly 30%, and reducing the field to plumbing requires that deal plus continued in-housing. The modal case requires nothing to happen at all: it is the trend already visible in the filings, extended three and a half years against a market growing 11–14% a year.
Strongest surviving dissentQuinn Ratio, Elias Lantern and Nico Tilt all reject the modal case from different directions. The strongest dissent is the shock-summing argument: each of the three structural shocks is individually more likely than the Structural Break scenario, and better than a one-in-three chance that at least one lands. Industry planning treats them as separate tail risks and never adds them up.
03The Outcomes — 4 scenarios, priced by the research
Squeezed but Standing38–46%
Magnite and PubMatic stay independent and keep compounding CTV revenue while losing share of total CTV monetization; take rates compress under supply-path pressure; one or two mid-tier independents are absorbed; the three frontiers get built and marketed but never reprice the category.
How this number was derived
Decomposition: P(market keeps growing along the forecast path) 0.85 × P(no structural break with real effect before 2029) 0.84 × P(squeeze rather than absorption or a successful climb, conditional on both) 0.50 ≈ 0.42.
Drivers
- US CTV spend growing ~11-14% annually to ~$51B by 2029
- Supply-path optimization compressing intermediary count and fees
- Platform-owned stacks taking share without eliminating incremental demand
Signals to watch
- Magnite's Disney renewal around October 2026
- Next ANA transparency benchmark SSP-count figure
- Disclosed take rates holding flat rather than falling or rising
Horizon · through 2029
Absorbed into the Platforms18–26%
The platform-owned sell side wins decisively: Fox manages Roku's marketplace as a content owner, more streamers in-house their ad servers, Amazon's supply-plus-DSP overlap goes unchallenged, and one of the two leading independents loses its independence. Survivors become plumbing vendors that cannot set terms.
How this number was derived
Counted base rate: ~5% annual acquisition hazard per scaled sell-side platform gives 1 − 0.95^7 ≈ 0.30 that one of the two leaders is acquired across seven company-years, discounted by ~0.75 for the conjunction with continued in-housing ≈ 0.22.
Drivers
- Thirteen years of uninterrupted sell-side consolidation
- YouTube as the existing proof that owning screen plus identity plus buying tool works
- Streamer in-housing removing inventory from the independent field
Signals to watch
- Fox-Roku closing conduct in H1 2027
- A second major streamer in-housing its ad server
- Any strategic investment in Magnite or PubMatic
Horizon · through 2029
Up the Stack15–23%
The frontier bets pay: curation becomes a product buyers want on the merits, a CTV attention metric clears accreditation and starts pricing inventory, outcome deals standardize enough to underwrite, and seller-side agents entrench the inventory owner. The SSP becomes the decisioning layer and take rates hold or expand.
How this number was derived
Base rate on ad-tech intermediaries climbing the value stack and holding margin: roughly two clear successes (LiveRamp's data pivot, Rubicon's rebuild into Magnite) in about six attempts ≈ 0.33, discounted to ≈ 0.20 for the current evidence gap — no accredited CTV attention metric, immaterial agentic revenue by every disclosure, one autonomous campaign on record.
Drivers
- Agentic deployments compressing the DSP rather than the SSP
- Curation and pre-bid attention already shipping across the roster
- Retail-media data moving structurally onto the sell side
Signals to watch
- A seller disclosing agentic revenue in isolation
- A disclosed take rate that rises
- MRC accreditation of a CTV-specific attention metric
Horizon · through 2029
Structural Break12–20%
An exogenous shock resets the board: an AdX divestiture that survives appeal and reshuffles publisher supply paths, or a VPPA ruling that makes automatic content recognition data legally radioactive and forces targeting back to context, or a genuine agent-to-agent layer that makes the auction and its take rate the most compressible cost in the chain.
How this number was derived
Union of three roughly independent shocks — structural Google remedy ordered and effective ≈0.16 (0.45 ordered × 0.35 surviving appeal and implemented in window), privacy ruling gutting ACR ≈0.15, agent disintermediation of the auction ≈0.10 — giving 1 − (0.84 × 0.85 × 0.90) ≈ 0.36 that at least one occurs, times ≈0.45 that it dominates the outcome rather than being absorbed ≈ 0.16.
Drivers
- Brinkema's unissued remedies ruling after November 2025 closing arguments
- Supreme Court VPPA case plus Samsung and Amazon Fire TV class actions
- Unresolved AdCP-versus-AAMP standards contest
Signals to watch
- The Brinkema ruling and its appellate path
- The Supreme Court's VPPA decision
- Any demonstrated transaction cycle with no SSP in the chain
Horizon · through 2029
Probabilities are the research's own scenario bands, priced as outcomes. The full argument — history, current state, drivers, and sources — lives in the corpus: read the Future Trajectory chapter.
04The Base Rates
05The Market Snapshot
06What Would Change the Call
- {'signal': "Magnite's Disney renewal, due around October 2026", 'effect': 'Renewal on similar terms supports the modal case or a climb up the stack; a loss or material reset moves weight to absorption.'}
- {'signal': 'The Brinkema remedies ruling and its appellate path', 'effect': 'A structural divestiture upheld is the single largest move toward Structural Break.'}
- {'signal': 'Fox–Roku closing conduct, expected H1 2027', 'effect': 'A narrowed third-party marketplace moves weight to absorption.'}
- {'signal': 'Any seller disclosing agentic revenue in isolation', 'effect': 'A real, material, separately reported figure moves weight to Up the Stack.'}
- {'signal': 'A disclosed take rate that rises anywhere', 'effect': 'The cleanest single confirmation that the climb up the stack is working.'}
- {'signal': 'MRC accreditation of a CTV-specific attention metric', 'effect': 'Would let attention price inventory rather than inform plans; supports Up the Stack.'}
- {'signal': "The Supreme Court's VPPA decision", 'effect': 'Broad liability affirmed supports Structural Break; a narrowed statute supports the modal case.'}
- {'signal': 'The next ANA transparency benchmark', 'effect': 'A still-falling SSP count per advertiser confirms the squeeze.'}
- {'signal': 'A second major streamer in-housing its ad server', 'effect': 'Confirms Netflix and Disney were the leading edge rather than the exception; supports absorption.'}
07Sources
The modal case sits in the low forties, which means the most likely single outcome is still more likely than not to be wrong. Three of the seven desk forecasters reject it outright. The largest single weakness in this forecast is that Index Exchange, Equativ and OpenX disclose no CTV revenue at all, so any claim about the independents' collective share rests on the two companies that do disclose — and the frontier products at the center of the argument are reported inside bundles specific enough to grow and vague enough to be unfalsifiable.
