Anthropic Faces Friday DPA Deadline as $350B Secondary Closes

The New Landscape

Previously covered: OpenAI's $111B cash burn and infrastructure constraints. What's new today: the market response has crystallized into the largest single-week capital deployment into alternative AI chip architectures ever recorded. OpenAI signed a $10B Cerebras deal. Meta struck a $100B+ AMD deal with unprecedented equity warrants (option to buy 10% of AMD at $0.01/share). Three inference chip startups raised $1.1B+ in a single day: MatX ($500M), SambaNova ($350M), and Axelera ($250M). And Taalas added $169M for its radical model-as-hardware approach.

Why the Market Moved This Week

The catalyst was OpenAI's own financials: $140B of $218B projected burn goes to inference, not training — a 64/36 split that inverts conventional wisdom about where AI infrastructure value accrues. Nvidia's $20B Groq acquisition created a vacuum that directly catalyzed the Cerebras deal, and hyperscalers are now taking equity positions in chip suppliers to secure supply.

Nvidia's response — Vera Rubin promising 10x cost-per-token improvement over Blackwell with rack-scale integration — raises the competitive bar dramatically. But the market is betting that inference specialization creates defensible wedges even against Nvidia's platform play. Jane Street co-leading MatX signals that quantitative, returns-focused capital sees asymmetric opportunity.

The Second-Order Opportunity

If inference costs drop 10x (Vera Rubin's promise) or more (specialized chips), the implications cascade through the entire AI stack:

• AI application companies see COGS drop dramatically, expanding viable use cases
• Agentic AI becomes economically viable at scale — multi-step reasoning at $0.05/query vs. $0.50
• Edge inference becomes a real market as specialized chips enable on-device model serving

The biggest alpha may not be in chip companies themselves — it may be in application-layer companies whose business models become viable only when inference costs drop by an order of magnitude.

Update from Previously Covered

Blue Owl was flagged Monday as a potential canary in private credit. What's new: Kuvare Holdings publicly stated it "rejected a significant number of loan assets" in Blue Owl's $1.4B forced sale because they were too risky or performing poorly. Separately, Bank of America projects digital infrastructure securitizations will hit $60B in 2026 (up 50% YoY), with bankers now exploring securitization of AI chips and power generators — depreciating hardware assets with 3-5 year useful lives.

The Structural Risk

Blue Owl channeled $64B in debt and $5.6B in equity into data center projects in 2025 alone. Its $1.6B private fund faces accelerating redemptions it can't meet because the underlying loans are illiquid. The attempted merger with a $16.5B publicly traded vehicle — which itself trades at a 20% discount to stated NAV — was called off after investors balked at a 20% haircut.

The data center securitization market is exhibiting three classic bubble indicators:

1. Rapid volume growth: 50% YoY to $60B
2. Yield compression: attracting risk-averse capital (insurers entering)
3. Product innovation into riskier assets: AI chips and power generators

Nuveen portfolio manager Aashh Parekh, managing $1.4 trillion, warned publicly: "If you just keep buying it, it can take over your entire portfolio." The key question being debated — under what circumstances can cloud tenants walk away from a data center? — is the 2026 equivalent of asking about subprime default rates in 2006.

The Bright Spot: Public Markets Still Hungry

Forgent Power Solutions (NYSE: FPS) completed the largest US IPO of 2026, raising $1.7B at a $6.2B market cap for data center electrical distribution equipment. CoreWeave's $8.5B debt raise at ~6% with an expected A- rating (down from double-digit rates) signals AI infra is becoming an institutional asset class — but only for the best credits.

The AI compute supply crunch just became a financing crisis — the winners will be whoever controls existing capacity or builds the next capital formation model.

The Deal That Reshapes Fintech

Stripe — freshly valued at $159B in a Tuesday tender offer (up 73% YoY), processing $1.9T in payment volume (up 34% YoY) — is reportedly considering a whole or partial acquisition of PayPal, which has a $43.3B market cap after an 80%+ decline from 2021 highs. PayPal's stock popped 7% on the rumor alone. This would be the largest private-acquires-public deal in fintech history and a 36x leap from Stripe's largest prior acquisition (Bridge at $1.1B).

Why This Matters Beyond the Headline

Multiple intelligence streams reveal this isn't an isolated M&A rumor — it's the culmination of several converging forces:

• Stripe explicitly says it's not pursuing an IPO — John Collison stated going public would "sidetrack" growth, suggesting Stripe is using its private valuation as M&A currency
• The Collison brothers are investing personally in adjacent companies (MatX's $500M round), and Stripe invested in Xflow (cross-border payments, $16.6M Series A at $85M) — building a constellation of strategic investments
• Meta's stablecoin comeback (H2 2026, contacting vendors for wallet infrastructure) adds a crypto-native payments competitor entering from above
• Stablecoin infrastructure has crossed from experimental to production — Sling Money covers 70 countries with 23 people vs. Venmo's 49 state licenses for one country

The stablecoin dimension adds urgency. a16z's Noah Levine makes a compelling Jevons' Paradox case: stablecoins won't shrink financial services but massively expand the market. JPMorgan's Kinexys processes $2B+/day. Financial crime compliance costs $61B/yr in North America. The 1.3B unbanked represent net-new TAM. If Stripe combines PayPal's consumer base with stablecoin rails, the resulting entity controls the most vertically integrated payments stack in history.

Mid-market fintech companies are in a vise between consolidation at the top and crypto disruption from below.