Anthropic Revenue Hit $16.8 Billion in Q3, First Profit Sets Up IPO

Anthropic told prospective investors on Wednesday that its third-quarter revenue reached $16.8 billion, up about 45 percent from the prior quarter, and that the company was operating profitably for the first time. The disclosure, first reported by the Financial Times and confirmed to MarketIntelLabs by Reuters and Axios, gives the AI market its strongest bull case yet in a week dominated by debt repayment fears and falling chip stocks.

The number matters because it sits on the revenue side of a debate the market has only been pricing on the cost side. CoreWeave debt yields above 14 percent, a 10-year Treasury near 5 percent and a GlobalFoundries-led pullback in chip stocks have dominated the AI narrative for three days. Those are financing and equipment stories. This one is the opposite half of the ledger: whether the companies spending the most actually make money when the models ship.

Related reading: An $8 Billion Loss and a $2 Trillion Ask, as the Federal Door on Anthropic Faces Its September 30 Deadline.

Anthropic, which operates the Claude models, is a private company and does not file public statements, so the figure is management's own presentation to prospective investors rather than an audited print. Reuters reported that two people who attended the investor day confirmed the figures. The company told that room it had turned an adjusted operating profit in the third quarter, a message designed to soften what investors have called the sector's biggest structural risk: billions of dollars of compute spend with no visible return.

The comparison matters as much as the level. Reuters reported the third quarter grew about 45 percent sequentially, which implies second-quarter revenue of roughly $11.6 billion and a single-quarter jump of about $5.2 billion, an annualized run rate near $67 billion. Axios reported the growth was explained in part by Claude Code, an agentic coding product Bloomberg Intelligence estimates reached $6 billion to $7 billion in annualized revenue by late summer. Anthropic also underwrites its margin with pricing power: Bloomberg Intelligence figures it charges 20 percent to 30 percent more than rivals and still holds customers, and SemiAnalysis estimates the API business generates about 90 percent of revenue using roughly 58 percent of the compute needed to run its models.

Related reading: TSMC August Revenue Hits NT$514.81 Billion, Up 53.3% Year Over Year.

The timing is deliberate. Anthropic has been preparing what could be the largest initial public offering in history, and this investor-day disclosure is the first hard look at the unit economics that a listing will turn into a market multiple. Investors interviewed by Axios said some expect a $2 trillion valuation. Reuters, however, reported that several prospective investors are tempering expectations toward $1.5 trillion because revenue growth is decelerating from the first to second quarter, when sales more than doubled. A $2 trillion ask against $67 billion in annualized revenue is roughly 30 times sales; at $1.5 trillion it is still about 22 times. Both are demanding multiples for a company whose largest cost, frontier-model training, does not scale down.

The bull case the market got this week was not confined to Anthropic. Alphabet, one of the world's most profitable companies, reported free cash flow negative in its second quarter as AI infrastructure investment surged, which Axios flagged as sobering context. Against that backdrop, a young frontier lab saying it was cash-flow positive in the third quarter is a meaningful data point: it suggests the demand for frontier intelligence can materialize as operating cash before the buildout fully matures.

Related reading: TSMC Q3 Earnings: Record Profit Up 68.7% as AI Demand Lifts 2026 Outlook.

The caveats are real and numerical. Anthropic still faces enormous training costs and stock-based compensation that will weigh on reported profit, and the company has publicized more than $500 billion in planned infrastructure and computing spending over the next decade. Its premium pricing also sits under pressure from increasingly capable, cheaper open-weight models. And one profitable quarter is not an earnings trend; management's own framing conceded growth is slowing off a doubling pace.

What to watch next is the other scheduled print on this beat. The IPO filing will carry the first audited, dated revenue and operating figures, which will settle the $1.5 trillion versus $2 trillion valuation debate and, with it, a large slice of how the market prices frontier-model revenue generally. Anticipation around that filing is why this investor-day disclosure is being scoured: it is the first real book on whether the biggest spenders in AI can end the trade's defining fear, the gap between capex and cash. So far, for one quarter and for one company, the answer is yes.

Related reading: Samsung Profit Estimate Surges 782%, but AI Memory Mix Is Unclear.

This content is for informational purposes only and does not constitute financial advice. Past performance is not indicative of future results. Consult a qualified financial advisor before making investment decisions.