Google transferred $10 billion to Anthropic in a single commitment, the largest disclosed equity check written to an AI research lab by a hyperscaler. The capital arrived three quarters after Amazon locked in $4 billion across two tranches, completing a bifurcated funding structure that leaves Anthropic capitalized through late 2026 without diluting founder control below 60%.
The round valued Anthropic near $60 billion post-money, a 4.2x step-up from the $18.4 billion print in April 2024. Google's stake now sits between 16% and 18%, depending on liquidation preference architecture, while Amazon holds roughly 12%. Both hyperscalers negotiated compute credits as part consideration—Google's contribution includes $3.2 billion in TPU v5 and v6 access through Q2 2027, effectively converting capex into strategic moat. The structure mirrors Microsoft's OpenAI playbook but compresses the timeline: what took Redmond four years to deploy, Google executed in nine months.
The capital commitment solves two problems. First, Anthropic burns approximately $2.7 billion annually on training runs and inference infrastructure, a rate that accelerates as Claude's enterprise adoption doubles every six months. Dario Amodei disclosed in October that breakeven requires $8 billion in cumulative revenue, a threshold the company expects to cross in late 2026 if current SaaS attach rates hold. Second, the investment locks Google into Anthropic's model development cycle, ensuring Vertex AI gets first access to Claude 4 and successor architectures before they surface in consumer channels. That sequencing advantage matters: enterprises now allocate 73% of LLM budgets to the vendor who delivers model updates inside their security perimeter first.
The deal reshapes capital allocation across the frontier lab tier. Anthropic now holds $14 billion in primary capital, the second-largest war chest after OpenAI's $17.9 billion. But the composition differs—Anthropic's balance sheet carries zero debt, while OpenAI's structure includes $4 billion in convertible notes that reset valuation triggers if AGI milestones slip past 2026. Google's equity check also removes the refinancing risk that forced xAI to raise $6 billion in December at a 29% discount to its previous round. Anthropic can now run 31 months without external capital, assuming current burn and no revenue growth—a longer runway than any competitor except OpenAI.
Allocators should track three variables. First, whether Google renegotiates its GCP revenue-sharing agreement with Anthropic, currently structured as a 12% take-rate on all Claude API calls routed through Vertex. That economics worked when Anthropic generated $200 million in annual recurring revenue; at the current $1.1 billion run rate, the margin pressure forces a reset. Second, how Amazon responds—its $4 billion commitment now looks structurally small, and AWS lacks first-mover rights on model releases. Expect a follow-on tranche before June. Third, whether Anthropic converts any of this capital into M&A, specifically targeting the 14 stealth-mode startups building agentic orchestration layers that sit between Claude and enterprise workflows.
Google's check arrived the same week Meta disclosed $65 billion in 2025 capex, 80% earmarked for AI infrastructure. The timing is clarifying: hyperscalers no longer view foundation models as vendors—they are now portfolio companies that must be capitalized to secure differentiation in a market where compute is abundant but frontier intelligence remains scarce.