The Headline Is $965B. The Story Is the Revenue.

Anthropic closed a $65 billion Series H on May 28, 2026, valuing the company at $965 billion post-money — overtaking OpenAI as the most valuable AI startup in the world. The round was led by Altimeter Capital, Dragoneer, Greenoaks, and Sequoia Capital, with participation from Blackstone, D.E. Shaw, Fidelity, and strategic partners including Samsung, SK Hynix, and Micron.

The valuation number is remarkable. But the numbers underneath it explain why sophisticated investors at this scale are writing checks, and what the trajectory means for the developers and enterprises building on top of Claude.

From $9 Billion to $47 Billion ARR in Five Months

At the end of 2025, Anthropic’s annualized revenue was roughly $9 billion. By the time of the Series H announcement, run-rate revenue had crossed $47 billion — a more than fivefold increase in five months. For reference, OpenAI’s estimated ARR sits around $33 billion, meaning Anthropic has now passed its closest rival on revenue as well as valuation.

The Q2 2026 projection is $10.9 billion in revenue and approximately $559 million in operating profit — the company’s first-ever profitable quarter. That shift from cash-burning to profitable operation changes the IPO story significantly. A company burning billions to sustain growth faces different public-market scrutiny than one that has crossed into profitability while still growing at triple-digit rates.

CEO Dario Amodei has publicly acknowledged the growth “far exceeded internal expectations.” Claude’s paid subscriber base had been hitting five-hour rate limits and capping code generation earlier this year. Both constraints have since been lifted. That context matters: the revenue surge wasn’t just commercial execution — it was constrained by capacity.

The SpaceX Compute Bet: $15 Billion a Year

The capacity constraint is being resolved at extraordinary scale. SpaceX’s IPO filing, submitted to the SEC in May 2026, revealed that Anthropic has committed to paying SpaceX $1.25 billion per month for AI compute through May 2029 — $15 billion per year, approximately $45 billion over the life of the contract. The deal covers 220,000+ NVIDIA GPUs spread across two data centers, Colossus and Colossus II.

For context: Anthropic is spending more on compute per year than most large software companies generate in total annual revenue. This is not an edge case in the AI infrastructure build-out; it is the defining economic characteristic of frontier AI development in 2026. The bet is that the capacity will generate revenue well in excess of its cost — and at current run rates, that math is beginning to work.

Practical effects for developers are already visible. Since the SpaceX contract came online, Anthropic has doubled rate limits for subscription plans, removed peak-hour restrictions on Pro and Max accounts, and raised API rate limits by as much as 17x for certain tiers.

Claude Code Is Carrying Its Weight

Among the individual drivers of Anthropic’s revenue growth, Claude Code stands out. The agentic coding tool — which launched in October 2025 and topped developer charts within weeks — now holds an estimated 54% of the enterprise coding market, compared to OpenAI Codex at 21%. It represents roughly 20% of Anthropic’s total revenue, or approximately $2.5 billion in annualized revenue from a single product launched less than eight months ago.

The May 2026 “Code with Claude” developer conference reinforced the product direction. Managed agents, proactive workflows, and the dynamic workflow system — which allows Claude to spawn multiple subagents concurrently — are all pushing Claude Code into territory that goes beyond autocomplete and toward autonomous software development. If you want a detailed breakdown of where Claude Code sits relative to Cursor and GitHub Copilot, the May 2026 comparison is worth revisiting in light of these numbers.

The Model Context Protocol (MCP) is another quiet growth driver. MCP now has more than 10,000 published servers connecting Claude to developer tools, databases, and enterprise applications — a number that makes it increasingly difficult for organizations to standardize on a different platform without migration friction.

The IPO Window and What It Discloses

Anthropic is eyeing an October 2026 NASDAQ listing. The trigger will be the S-1 filing — Anthropic’s first comprehensive public financial disclosure. That document will reveal things private investors already know but the market does not: precise gross margins, the structure of the compute contracts and whether they can be renegotiated, the concentration of revenue among top customers, and the equity structure governing a company where Amazon has invested $8 billion under terms that include cloud deployment commitments.

The valuation multiple tells part of the story. At $47 billion ARR and a $965 billion valuation, Anthropic trades at roughly 20.5x revenue — down from 27x when the Series G closed in February 2026. That compression reflects both the revenue growth and investor expectation that growth rates will normalize. A company growing from $9B to $47B ARR in five months cannot sustain that rate; the question for IPO investors is what the steady-state growth looks like when the capacity-constrained backlog is cleared.

Alongside the funding announcement, Anthropic shipped Claude Opus 4.8 — an upgrade to its flagship model with improvements in agentic coding, reasoning, and financial analysis, at unchanged pricing ($5 per million input tokens, $25 per million output tokens). The company also teased “Mythos-class” models as coming in the next few weeks. Mythos, Anthropic’s most capable model class, has drawn attention for its advanced cybersecurity capabilities and is being made available to major tech partners — Amazon, Microsoft, Apple — under Project Glasswing. More on what Opus 4.8 ships is covered in our earlier review.

What Changes for Builders

Three things matter most for developers and enterprises building on top of Claude right now. First, rate limits are up and will continue to rise as the SpaceX capacity comes fully online — the months of constrained access appear to be ending. Second, an IPO creates investor pressure for margin discipline, which historically translates to pricing stability or moderate increases on API tiers; it is worth locking in enterprise contracts before the S-1 makes those conversations more complicated. Third, the gap between Anthropic and its nearest competition on revenue has widened faster than most analysts predicted in January 2026, which has real implications for which platform is worth standardizing on for long-term projects.

The $965 billion valuation is a milestone, but it is the revenue velocity — quintupled in five months, profitable in Q2, with a capacity constraint now resolving — that explains why the milestone happened, and what comes next.

Further Reading

• Anthropic Series H announcement — the official release with investor list and fund use breakdown
• TechCrunch: Anthropic raises $65 billion, nears $1T valuation ahead of IPO — good coverage of the competitive dynamics with OpenAI
• Anthropic Projects $559M Q2 Operating Profit on $10.9B Revenue — detailed breakdown of the Q2 financial projections and what profitability means for the IPO narrative