Anthropic raises $65B at $965B valuation, surpassing OpenAI to claim the title of the world's most valuable AI company.
The AMW Read
Resolves the open debate about frontier AI profitability with the first profit milestone; overturns valuation rankings and introduces a new trust-builder pattern (Pentagon refusal), updating the capital-cycle dynamics for the entire segment.
Anthropic raises $65B at $965B valuation, surpassing OpenAI to claim the title of the world's most valuable AI company.
Anthropic has raised $65 billion in new funding at a $965 billion valuation, making it the highest-valued AI company globally. The round arrives amid explosive revenue growth: ARR surged from $9 billion at end-2025 to $45 billion by May 2026, overtaking OpenAI's $33 billion. Key drivers include the Claude Code developer tool, which alone contributed $2.5 billion ARR by late February after the Claude Sonnet 4.6 and Opus 4.6 releases expanded context windows to 1 million tokens and introduced Agent Teams for multi-agent parallel task execution. A pivotal moment came in February 2026 when Anthropic declined a $200 million Pentagon contract over surveillance and autonomous weapons clauses, triggering the "QuitGPT" movement that drove enterprise users to migrate from OpenAI. The company subsequently inked three major compute deals: a five-year, 5GW TPU agreement with Google, a 10-year ~$100 billion compute partnership with Amazon, and a deal covering 220,000 GPUs from Elon Musk's Colossus 1 project. Anthropic projects Q2 2026 operating profit of $559 million, marking the first profitable quarter for a leading AI lab, with unit economics improving from $0.71 compute cost per $1 revenue to $0.56, driving gross margin from 29% to 44%.
The significance of this event extends far beyond a single funding round. Anthropic's trajectory exemplifies the 'hyperscaler distribution moat' and 'fastest-ARR-ramp' patterns, while its Pentagon refusal represents a counterintuitive trust-building strategy that resonates with privacy-sensitive enterprises. The company's compute deals — securing TPU, Trainium, and GPU capacity simultaneously — mirror the 'multi-architecture arbitrage' structural force, enabling margin expansion independent of a single chip vendor. The QuitGPT migration validates the 'context-engineering moat' as a switching-cost mechanism, with Anthropic offering direct ChatGPT import tools. At the capital-cycle level, the mega-round and $965B valuation update the 'capital-compression arc' for foundation model labs: the market now appears to be rewarding capital-efficient deployment over raw scale, with Anthropic's profitability milestone potentially resolving the open debate about whether frontier AI can achieve sustainable unit economics.
Industry analyst Tang Hui noted to Leiphone: 'When every dollar of revenue consumed 71 cents of compute, the business model was questionable. Now at 56 cents, with margins expanding 15 points, the flywheel is spinning — Anthropic has 60 billion more dollars annually to reinvest.' The article also highlights a structural divergence from OpenAI: Anthropic avoids subsidizing hundreds of millions of free-tier inference requests, a cost burden that eroded OpenAI's margins even as it built brand awareness. If Anthropic sustains its profitability trajectory, it may force a re-evaluation of how Wall Street values AI labs — shifting the metric from user growth alone to gross-margin-per-revenue-dollar.
#Anthropic #AIEconomy #ClaudeCode #FoundationModels #EnterpriseAI #ComputeEconomics
