Key Takeaways:
Anthropic's first quarterly profit has pulled the AI sector's profitability timeline forward by two years — and set up a $2 trillion IPO test.
Anthropic posted preliminary second-quarter revenue above $11.5 billion, up 14-fold from a year earlier, and recorded its first adjusted operating profit — a milestone that has moved the AI sector's profitability expectations forward by two years. The maker of the Claude model said the quarterly figure compares with $787 million in the same period last year and $4.73 billion in the first quarter, a sequential jump of at least 143 percent.
"Could they get a $2 trillion valuation, yeah they could, and I just wonder if it would stay there over time," David Merkel, a principal at Aleph Investments, said.
The results set up Anthropic's planned October IPO, which investors expect to value the company at $2 trillion or more — a listing that would surpass SpaceX as the largest in history. Underwriters Morgan Stanley, Goldman Sachs and JPMorgan Chase are leading the deal, with the company projecting revenue of $190 billion to $200 billion by 2028.
Anthropic's annualized revenue run rate reached $65 billion by the end of July, up from $47 billion in May and about $9 billion at the end of 2025 — a sevenfold increase in roughly seven months. About 85 percent of revenue comes from enterprise APIs and developers, generating an average of $211 per user per month, more than eight times OpenAI's figure. The company counts more than 300,000 enterprise clients, with over 1,000 paying more than $1 million a year, and eight of the top 10 Fortune companies use Claude.
Claude Code, the programming agent commercialized in May 2025, reached $1 billion in annualized revenue within six months and $2.5 billion by February 2026. Around 4 percent of all public code submissions on GitHub are completed by it, a share that doubled within one month. Shopify engineers using the tool cut the time to launch new features to five days from 24, according to the company.
The enterprise mix explains the profitability gap with rivals. OpenAI, whose annualized revenue run rate is about $40 billion, posted an operating loss rate of 122 percent in the same quarter and expects to burn about $27 billion in cash this year. Google Cloud, by contrast, grew 82 percent year-over-year to $24.77 billion in second-quarter revenue with a 32.9 percent operating margin, though that figure reflects the full cloud platform rather than the model layer alone.
Anthropic's profit comes from improving unit economics rather than cost-cutting. Inference gross margin rose to 85 percent from 38 percent, the compute cost required to earn $1 fell to about $0.56 from $0.71 in the first quarter, and the cache hit rate exceeds 90 percent. Training costs run about a quarter of OpenAI's.
The sustainability of that margin is the open question. Anthropic signed computing power procurement commitments of more than $300 billion over four months, spending over $160 million a day on compute — roughly $14.4 billion in the quarter against $11.5 billion in revenue. The company's 2028 forecast of $190 billion to $200 billion assumes enterprise AI spending keeps compounding while inference efficiency outpaces rising training costs.
At a $965 billion valuation from its latest funding round, Anthropic trades at about 20 times annualized revenue, below Palantir at roughly 55 times and Cloudflare at 41.6 times expected 2026 revenue. OpenAI's $852 billion valuation implies about 21 times its $40 billion run rate, but without the profit. The divergence will be tested in the public market, where investors are shifting from narrative-driven multiples to questions about free cash flow, customer retention and the slope of the inference cost curve.
This article is for informational purposes only and does not constitute investment advice.
