Nvidia backs Thinking Machines with compute and capital

The provided text contains only cookie banner and privacy policy information — there is no financial news, data, or market-relevant content. No prices, earnings, guidance, policy actions, or events are reported that would warrant portfolio action.

Analysis

The market is already repricing the plumbing of digital marketing: identity resolution, server-side tracking and clean-room orchestration are the emergent high-margin services that will absorb spend migrating away from ephemeral third‑party signals. Expect revenue mix shifts to favor vendors that can both (a) deterministically tie conversions to persistent IDs and (b) host privacy-safe data collaboration — these two capabilities shorten attribution cycles and raise CPMs by an incremental 10–30% for advertisers who can measure ROI.

Walled gardens will capture a disproportionate share of top‑of‑funnel ad dollars in the near term because they offer the path of least resistance for marketers seeking measurable outcomes; however this creates a two‑year arbitrage where independent identity and clean‑room vendors can charge 2–5x higher per‑connection fees as publishers and agencies build interoperable stacks. Smaller ad tech and supply-side platforms face margin compression from both lower-quality signal and from publishers migrating to direct commerce- and subscription-based monetization, compressing programmatic CPMs by a potential 15–25% in 6–12 months.

Regulatory and state‑law definitions of “sale” or “sharing” of data are the primary tail risk: a stricter interpretation that forces opt‑in will accelerate first‑party ID adoption but could also prompt advertisers to rebalance spend to owned channels faster than infrastructure can monetize, creating a 3–9 month revenue shock for mid‑tier vendors. Conversely, rapid adoption of standardized hashed‑ID frameworks or universal clean‑rooms (within 6–18 months) is the most direct catalyst to re‑rate winners; lack of standardization is the principal path to a broader market reset.

The contrarian angle: the market assumes privacy equals a permanent shrinkage of ad monetization. That understates the willingness of advertisers to pay for quality measurable outcomes — not raw reach — and the capacity of publishers to convert anonymous users into higher‑ARPU relationships via lightweight prompts, paywalls, and commerce integrations. If identity consolidation occurs around a small number of neutral vendors, those vendors will see scalable pricing power and multiple expansion, even as programmatic incumbents reprice down.

Key Decisions for Investors

• Long RAMP (LiveRamp, RAMP) — 6–12 month hold. Size 2–3% book. Rationale: identity graph and data‑connectivity revenues reaccelerate as clients pay up for deterministic matching; expected upside 20–35% if adoption of universal hashed IDs increases. Key risks: loss of neutrality / regulatory interventions; stop loss 12%.
• Long SNOW (Snowflake, SNOW) — 9–18 month hold. Size 1.5–2% book. Rationale: clean‑room adoption drives higher seat growth and upsell to compute; revenue re‑rating potential 25–40% as publishers monetize first‑party data in collaboration with advertisers. Risk: slower enterprise adoption or pricing pressure; hedge with 3–6 month protective puts.
• Pair trade — Long GOOGL (Alphabet, GOOGL) 3–12 months / Short MGNI (Magnite, MGNI) 3–12 months. Size net market‑neutral 2% gross each leg. Rationale: reallocation to walled gardens benefits Alphabet; independent supply‑side platforms face CPM headwinds and client consolidation. Risk: regulatory/open‑market interventions could hurt GOOGL; target 15–25% capture on alpha, trim at 10% adverse move.
• Event/options play — Buy 6–9 month TTD (The Trade Desk) calls or buy MGNI put spread if programmatic downside accelerates. Size small (0.5–1% of book) as asymmetric option exposure to consolidation or shock. Rationale: Trade Desk stands to benefit if the industry standardizes on its or partners’ ID solutions; put spread protects against rapid CPM deterioration across open exchange. Risk: carry and implied vol decay; treat as tactical, not core.