Anthropic’s Q2 revenue may hit $10.9 billion

Anthropic is on track for $10.9 billion in Q2 revenue, its first profitable quarter.

Anthropic may be moving from fast-growing AI startup to serious business faster than most of its rivals. CNBC reported on May 20, 2026 that the company is on pace to generate $10.9 billion in second-quarter revenue, up from $4.8 billion in the first quarter and above its total sales for all of 2025.

If that number holds, it would mark a rare milestone in frontier AI: a company with massive model demand, rising enterprise adoption, and enough scale to turn a profit in the same quarter. That matters because the AI sector has spent the last two years proving it can raise money and spend it quickly; Anthropic is now showing it can also convert usage into revenue at a pace that changes the math.

Why this revenue number matters

Revenue at this scale changes how investors read Anthropic. A company pulling in $10.9 billion in one quarter is no longer being judged only on model quality or product hype. It is being judged on unit economics, infrastructure spending, and whether it can keep demand ahead of compute constraints.

The jump from $4.8 billion in Q1 to $10.9 billion in Q2 is especially striking because it suggests the company’s enterprise and developer business is expanding quickly, while consumer usage is also rising. CNBC said the company’s revenue in 2026 has already reached $10 billion, which means this one quarter could nearly match the entire prior year in a single stroke.

• Q2 revenue target: $10.9 billion
• Q1 revenue: $4.8 billion
• 2025 revenue total: $10 billion
• Reported monthly payment to SpaceX for compute: $1.25 billion

That kind of growth does not happen without a very specific product mix. Anthropic’s Claude family has become a favorite for developers, and Claude Code has helped turn that preference into direct usage revenue. The company also sells into large organizations, where contracts can scale quickly if teams standardize on a model for coding, support, and internal workflows.

The product mix behind the surge

Anthropic was founded in 2021 by former OpenAI executives and researchers who wanted a different path for building AI systems. That origin story matters less than the current business, but it explains why the company has leaned hard into model safety, enterprise sales, and high-trust deployments.

Its biggest public-facing product is Claude, and the company has tried to make that family useful in places where people actually pay. Coding is the clearest example. A strong coding assistant can drive daily usage, which drives subscriptions, which drives enterprise contracts. That is a much better business loop than a chatbot that gets occasional attention and little retained spend.

“We think Claude Code is the best coding tool out there.” — Dario Amodei, Anthropic co-founder and CEO, in a May 2025 interview with Y Combinator

Amodei’s point is easy to dismiss as CEO optimism, but the revenue numbers give it weight. If developers keep paying for Claude Code and enterprises keep rolling out Claude across teams, Anthropic does not need a consumer app to dominate the headlines. It needs repeat usage, high retention, and enough compute to keep response times and model quality acceptable.

Compute is now the bottleneck, not demand

Anthropic’s biggest problem may be the one most startups would love to have: demand is outrunning infrastructure. The company said last month that enterprise and developer demand, plus a sharp rise in consumer usage, created “inevitable strain” on its systems. That is a polite way of saying growth is expensive.

To keep up, Anthropic struck a deal with SpaceX earlier this month to use all of the compute capacity at the Colossus 1 data center in Memphis, Tennessee. According to SpaceX’s IPO prospectus, Anthropic will pay $1.25 billion per month through May 2029. That is a staggering commitment, even for a company printing revenue at this speed.

• Anthropic founded: 2021
• SpaceX compute commitment: through May 2029
• Colossus 1 location: Memphis, Tennessee
• Potential IPO timing: 2026

The compute deal also shows how the AI market has changed. The winners are no longer just model labs; they are companies that can secure chips, power, and data center capacity at industrial scale. Anthropic is paying for that access because it has to. If it does not, growth slows and customers feel it first.

How Anthropic stacks up against OpenAI

The comparison with OpenAI is unavoidable. OpenAI is valued at more than $850 billion by private investors, and CNBC reported that it could confidentially file a draft IPO prospectus as soon as Friday. Anthropic is also eyeing a 2026 IPO, which means both companies may end up asking public markets to judge their growth, margins, and infrastructure bills.

On paper, Anthropic’s revenue run rate looks absurdly strong. On the other hand, OpenAI has broader consumer recognition, a larger product footprint, and a much louder public presence. Anthropic’s edge is that it may be proving something investors care about even more: that enterprise AI can produce real revenue fast enough to justify the capex.

• OpenAI private valuation: more than $850 billion
• Anthropic reported investor talks: $900 billion valuation
• OpenAI IPO filing could come: as soon as Friday, per CNBC
• Anthropic IPO target: 2026

There is one more detail worth watching. Anthropic’s growth has come with political friction, including a public clash with the Pentagon that briefly sent its Claude app to the top of Apple’s App Store in February. The company was blacklisted by the Defense Department after refusing to grant broad military access, then later appeared to regain ground after launching Claude Mythos Preview, a model with advanced cybersecurity capabilities.

What to watch next

Anthropic’s Q2 number is more than a headline. It is a signal that the company has crossed into a new phase where revenue, profit, and infrastructure all matter at once. If the $10.9 billion figure holds, the real question is whether Anthropic can preserve this pace without turning every new customer into a compute bidding war.

My read: the next meaningful test is not whether Anthropic can grow again, but whether it can keep margins healthy while paying for enough capacity to satisfy enterprise demand. If it can do that, a 2026 IPO will look a lot more credible. If it cannot, the market may still love the growth story, but it will start asking harder questions about how expensive that growth really is.