Anthropic Made More in Q2 Alone Than It Did in All of Last Year, and Turned a Profit

For years, the standard line on frontier AI companies was that they were burning money at extraordinary rates — necessary investments in a race whose economics would eventually make sense, but not yet. Last week, Anthropic changed that narrative.

Bloomberg reported Thursday, based on documents shown to prospective investors, that Anthropic’s preliminary second-quarter revenue topped $11.5 billion — a more than 14-fold jump from the prior-year quarter. More importantly: Anthropic recorded positive adjusted operating income in the quarter. It is the first signal of a frontier AI lab posting an operating profit on a quarterly basis.

That’s a milestone the entire industry has been waiting for.

The Numbers in Context

The growth rate is almost hard to absorb. Revenue jumped from $787 million in the second quarter of 2025. It more than doubled the $4.73 billion recorded in the first quarter of 2026. Put differently: Anthropic’s revenue grew more in three months than most enterprise software companies grow in three years.

First-half revenue totals roughly $16.2 billion. Anthropic’s full-year 2025 revenue was about $10 billion. In six months this year it earned 1.6 times what it made in all of last year.

And the growth shows no sign of slowing. In May, Anthropic said its annualized revenue run rate had climbed above $47 billion. If Q2 came in at $11.5 billion and that run rate held into Q3, the full-year figure would substantially exceed the $47 billion figure Anthropic shared with investors three months ago.

What’s Driving the Growth

The headline revenue growth comes from two reinforcing trends: enterprise adoption of Claude for coding and software development, and the broader shift of professional workflows onto AI assistants.

Anthropic has gained traction among professionals using its software for tasks including coding — and the scale of that traction is now visible in the financials. Claude’s coding capabilities, particularly through the Claude Code command-line tool and API integrations into developer environments, have made it the default AI for a growing share of professional software development.

Enterprise is where the real money is. OpenAI CFO Sarah Friar also told investors this week that enterprise business is now bigger than consumer by revenue for the first time — suggesting this is an industry-wide shift, not unique to Anthropic. Businesses are embedding AI into production workflows in ways that generate recurring, growing revenue — not just one-off subscriptions from individual users.

The Profitability Question

The positive adjusted operating income is significant — but the word “adjusted” deserves attention. The documents do not specify what was adjusted, and the company has said nothing officially. Treatment of long-term compute commitments and stock-based compensation could change the picture considerably.

Adjusted operating income is a common metric in tech that excludes certain costs — typically stock-based compensation, amortisation of acquired intangibles, and sometimes restructuring charges. The “adjusted” figure can look considerably healthier than the GAAP equivalent. When Anthropic goes public, investors will scrutinise the gap between adjusted and GAAP profitability closely.

That said, even adjusted profitability at this scale matters. It demonstrates that Anthropic’s unit economics are improving — that the cost of generating each dollar of revenue is declining as the business scales. That’s the fundamental requirement for eventual sustainable profitability, and it’s not guaranteed just because revenue is growing fast.

The IPO Is Getting Closer

Anthropic confidentially filed listing paperwork with the US Securities and Exchange Commission in June and is working with Morgan Stanley, Goldman Sachs, and JPMorgan Chase on the offering. The Q2 numbers — released to prospective investors as part of IPO preparation — give the market its clearest picture yet of what Anthropic’s financials actually look like.

The early meetings have been described as high-level, focusing on the company’s Claude model series, market positioning, and leadership team, without delving into specific financial details or valuation figures. That’s standard for early investor conversations — the specific valuation discussion comes later, once the S-1 is filed and the roadshow begins.

Anthropic’s last private valuation was approximately $965 billion post-money after its $65 billion Series H round. With $11.5 billion in Q2 revenue and positive adjusted operating income, a public listing above $1 trillion — potentially significantly above it — is now in scope. The timing puts Anthropic on track for a fall 2026 debut, following OpenAI’s planned listing.

What It Means for the Industry

Anthropic’s Q2 results matter well beyond the company itself. They’re evidence that frontier AI has crossed a financial threshold that many in the industry doubted would arrive this quickly.

The argument against AI companies at trillion-dollar valuations has always been: the compute costs are too high, the unit economics don’t work, and the revenue will plateau as the novelty wears off. Anthropic’s numbers challenge all three assumptions simultaneously. Revenue is accelerating, not plateauing. Operating income turned positive. And the compute costs, while still substantial, are now being absorbed by a revenue base that’s growing faster.

If OpenAI’s S-1 shows a similar trajectory — and early indications suggest it will — the AI industry’s financial story changes fundamentally. Not “promising technology losing money” but “high-growth, increasingly profitable platform businesses at extraordinary scale.” That’s a different investment thesis. And it’s one that public markets will price very differently. For more on Anthropic’s competitive position and Claude’s role in the AI market, see our coverage of the OpenAI IPO filing and AMD’s $5 billion investment in Anthropic.

Read more tech related articles here.