Anthropic made money. The $2T IPO just got real

Anthropic made money. The $2T IPO just got real

Anthropic told prospective investors at its San Francisco headquarters on Wednesday, October 14, 2026 (UTC-7) that third-quarter revenue hit $16.8 billion, up roughly 45% from Q2's $11.6 billion and more than triple the $4.7 billion it booked in Q1. The company also said it reached adjusted profitability with positive free cash flow — its second consecutive quarter in the black on an adjusted basis.

The numbers, first reported by the Financial Times and Reuters based on attendees of the closed pre-IPO meeting, arrive less than a week after Anthropic kicked off its investor day targeting a valuation of $1.8–2 trillion. If those figures hold up in the S-1, they change the calculus for what was already set to be the largest IPO in history.

The quarter in numbers

Sources: Financial Times, Reuters, AI Weekly, SiliconANGLE. Q3 2025 figure from prior company disclosures.

The sequential growth decelerated from Q2's explosive 147% to a still-remarkable 45%, which is what you'd expect as the revenue base gets larger. The year-over-year comparison is staggering: Q3 2025 revenue was roughly $787 million, meaning Anthropic grew revenue about 21-fold in twelve months.

The adjusted gross margin of 85% is the number that should make investors pay attention. For a company that spends billions on GPU compute, an 85% gross margin implies that inference costs — the expense of running Claude for every query — are being driven down fast enough that the software-like economics of the API business are starting to show through. The adjusted operating margin above 11% means that after covering sales, R&D, and corporate overhead, the company is generating real operating profit on a non-GAAP basis.

Why it matters

This is the first time a frontier AI lab has reported quarterly revenue at this scale alongside a claim of profitability. OpenAI is widely reported to generate substantial revenue — The Information reported a $37 billion annualized run rate in early 2026, and subsequent estimates have climbed — but it has never claimed to be profitable, and its compute spending is widely reported to outpace revenue. Anthropic hitting adjusted profitability at a $16.8B quarterly run rate does something important: it proves that the AI API business can be a real, profitable software business, not just a capital-burning land grab.

The math on the IPO valuation gets more defensible too. $16.8B in Q3 annualizes to roughly $67 billion. At a $2 trillion valuation, that's a price-to-sales ratio of about 30x. For a company growing 45% sequentially and 20x year-over-year, 30x sales is aggressive but not absurd — Snowflake traded at similar multiples in its hypergrowth phase. The difference is that Snowflake was never profitable at this scale. Anthropic is claiming it is.

That claim puts pressure on OpenAI. If Anthropic can show investors a path to profitability while OpenAI continues to burn cash on custom chips and data center buildouts, the narrative that "whoever spends the most wins" starts to crack. It also raises the stakes for OpenAI's own IPO plans — if Anthropic prices at $2T and trades well, OpenAI may feel compelled to move faster rather than wait for a more favorable window.

The fine print

There are real reasons to be cautious. The Financial Times noted that its anonymous source "didn't say if it refers to gross or net revenue, or how Anthropic defines profitability." That distinction matters enormously. If the $16.8B includes gross billings rather than net revenue after revenue-sharing with cloud providers, the actual top line could be meaningfully lower.

The "adjusted" qualifier is doing a lot of work too. Adjusted profitability typically excludes stock-based compensation — which at a pre-IPO company can be enormous — and may also exclude depreciation on multi-year compute commitments to AWS and Google. Anthropic has committed tens of billions in cloud spending over coming years. Once those depreciation charges and stock comp hit the GAAP income statement, the net loss could still be very large.

AI Weekly's editor made the point bluntly: "Doubling the $965B May Series H valuation on a single-source quarterly print puts the burden on the forthcoming S-1 to show whether $16.8B is net GAAP revenue and whether 'adjusted profitability' survives stock-comp and multi-year compute obligations."

The single-quarter nature of the data is another risk. Q3 benefited from strong enterprise adoption of Claude Opus 5.5 and Sonnet 5.5, which launched in late September. Q4 could see slower growth as enterprises digest the new models, or as competition from Google's Gemini 4 Argon — which opened broadly on October 13, 2026 (UTC-7) — puts pressure on pricing.

What to watch next

The S-1 filing is the make-or-break document. If Anthropic files publicly in the next two to three weeks — consistent with a potential IPO marketing launch the week of November 9, 2026 — it will reveal audited revenue, the real GAAP loss, the size of cloud commitments, and customer concentration. Those numbers will determine whether $2T is a floor or a ceiling.

Three specific indicators to track: (1) whether Q3 net revenue after cloud revenue share comes in above $12B, (2) whether the GAAP operating loss is under $3B for the quarter, and (3) whether any single customer accounts for more than 10% of revenue. If all three land favorably, the IPO is likely to price at the top of the range. If not, expect the valuation band to drift toward $1.5T — still enormous, but a meaningful haircut from the current pitch.

The era of "AI companies don't make money" may be ending. Whether it ends with a profitable public company or a heavily adjusted mirage is the question the S-1 will answer.