--- title: "Anthropic targets $900bn and leapfrogs OpenAI" date: 2026-05-26T08:00:00+02:00 language: en slug: 2026-05-26-anthropic-30-milliards-900-milliards-valorisation-openai url: https://mathieuhaye.fr/blog/en/2026-05-26-anthropic-30-milliards-900-milliards-valorisation-openai alternate: https://mathieuhaye.fr/blog/2026-05-26-anthropic-30-milliards-900-milliards-valorisation-openai category: Business & Growth description: "Anthropic closes $30bn this week at a $900bn valuation. Sequoia, Altimeter, Greenoaks lead. Why the market is paying that price, and what it means for buyers." --- # Anthropic targets $900bn and leapfrogs OpenAI > Anthropic closes $30bn this week at a $900bn valuation. Sequoia, Altimeter, Greenoaks lead. Why the market is paying that price, and what it means for buyers. According to [Bloomberg](https://www.bloomberg.com/news/articles/2026-05-22/anthropic-to-close-over-30-billion-round-as-soon-as-next-week) on May 22, 2026, Anthropic is set to close a funding round of at least $30bn during the week of May 26. Pre-money valuation comes in above $900bn. Four funds are co-leading the deal: Sequoia Capital, Dragoneer Investment Group, Altimeter Capital and Greenoaks Capital Partners, each writing a roughly $2bn check. Long-time backers Founders Fund (Peter Thiel) and General Catalyst are also taking part. The $900bn mark is symbolic: it pushes Anthropic past OpenAI, whose last post-money valuation hit $852bn in March 2026. The San Francisco-based company is now the most valuable private AI shop on the planet. Three months earlier, in February, Anthropic closed a $30bn Series G at a $380bn valuation, per the company's [official announcement](https://www.anthropic.com/news/anthropic-raises-30-billion-series-g-funding-380-billion-post-money-valuation). That is a 137% jump in four months. The operational numbers behind the deal explain investor appetite. Per projections shared with backers and reported by [CNBC](https://www.cnbc.com/2026/05/20/anthropic-revenue-explosive-growth-ipo-profitable-quarter.html) on May 20, 2026, Anthropic expects $10.9bn in revenue for Q2 alone, versus $4.8bn in Q1. A 2.3x bump in three months. Projected Q2 operating income lands at $559m: that would be the firm's first-ever profitable quarter. The annualized run-rate points to $50bn by end of June. An IPO is planned for October 2026. OpenAI, for its part, confidentially filed with the SEC on May 22, targeting a September 2026 listing. ## When valuation stops being a bet Through 2025, AI lab valuations were justified by the promise. Today they are justified by the cash. The jump from $380bn to $900bn in four months is not a speculative blowoff if you accept the revenue numbers. At $10.9bn for a single quarter, Anthropic trades around 21x annualized run-rate. For context, Snowflake trades near 14x, Palantir near 50x, and Nvidia near 22x forward sales. Anthropic sits in the same corridor as the hyper-growth software names, not in some category disconnected from the fundamentals. The engine of that growth has a specific name: Claude Code. The coding agent, launched in May 2025, hit a $1bn annualized run-rate within six months, then $2.5bn by February 2026. At that pace, it now represents the majority of Anthropic's revenue. Enterprise sales make up roughly 80% of the total. More than 1,000 companies spend over $1m a year with the vendor. The revenue is not ad-driven, not consumer virality, not a regulatory wager: it is signed contracts with a procurement office and a legal team on the other side. That revenue structure shifts the conversation. When an investor values Anthropic at $900bn, they are no longer buying a narrative; they are buying a contractual curve. The firm posted its [first lead over OpenAI](https://www.therundown.ai/p/the-enterprise-shift-openai-saw-coming) on the Ramp enterprise adoption index in April 2026, at 34.4% versus 32.3%. That is a corporate spend signal, not a poll. It matters to a fund partner who has to defend a nine-figure check in front of an investment committee. ## The implicit bet: margin is not guaranteed The quieter memo behind the raise is less flattering. Anthropic is warning investors that Q2 profitability may not repeat in Q3 or Q4, due to scheduled increases in compute infrastructure costs. The firm announced over recent weeks a deal with SpaceX worth $1.25bn per month in compute through 2029, or $15bn per year for access to the Colossus supercomputer alone, per [TechCrunch](https://techcrunch.com/2026/05/20/anthropic-will-pay-xai-1-25-billion-per-month-for-compute/). Add the existing Amazon contracts ($100bn of multi-year compute, see our piece on [Anthropic and Amazon signing a $100bn compute pact](/blog/en/2026-04-21-anthropic-amazon-100-milliards-compute-ia)) and ongoing talks with Microsoft for its Maia 200 chips. Compute is becoming the dominant cost line of a mature AI company. Where a traditional SaaS vendor pays 25 to 30% of gross margin on infrastructure, Anthropic probably pays 50 to 60%. The $30bn raised this week is not a marketing war chest; it is the next three years of compute bills. Sequoia is not betting on a marketing flywheel; it is betting on the company's ability to sign enough enterprise contracts to absorb the compute commitments already on the books. Hence the urgency of the IPO. October 2026 is five months away. Anthropic has to close this private round to give itself time to file a clean prospectus, with a profitable quarter on record. The market window is narrow. A macro downgrade or a bad surprise on compute consumption could shut the listing door for all of 2027. ## What it changes for B2B buyers For CIOs and chief procurement officers currently negotiating their Anthropic or OpenAI contracts, the situation has a direct impact. Vendors racing toward an IPO will push hard for multi-year commitments, even at significant discounts. This is historically the most favourable moment to sign a 3-year framework contract on Claude or GPT: the internal pressure on Anthropic salespeople is to lock in committed revenue ahead of the S-1. The second consequence touches vendor selection. A company valued at $900bn is no longer a startup. It enters the *too big to fail* bucket in the eyes of IT teams. Compliance, SOC 2 Type II, DPAs, contractual SLAs: all become easier to obtain. Vendor risk drops, which unlocks use cases that were previously reserved for self-hosted models (private banking, healthcare, defence). SAP already took the step, embedding Claude at the core of its suite (see [SAP hands Claude the brain of its autonomous enterprise](/blog/en/2026-05-18-sap-anthropic-claude-erp-autonome)). KPMG followed (see [KPMG signs with Anthropic](/blog/en/2026-05-20-kpmg-anthropic-claude-big-four-consulting)). A $900bn valuation makes those choices even more defensible in an investment committee. ## From my desk The episode lines up directly with my personal Bloomberg dashboard. Since February 2026 I have been running a Claude Haiku 4.5 agent every morning that parses my portfolio composition, recomputes sector exposure and suggests rebalancing moves against a risk grid I defined. Monthly inference cost barely reaches $20 for around 250 structured queries. At that price point, Anthropic's unit economics become obvious: the perceived value (reasoning roughly equivalent to a junior analyst) bears no relation to the cost. That gap is what explains the 80% enterprise revenue split. An SMB or mid-market buyer does not think in tokens; they think in headcount avoided. When a portfolio review that would take me forty minutes a week is handled in six seconds for a few cents, the ROI calculation no longer fills a slide. That is exactly the argument I make to my freelance clients when I scope an automation or an AI app: the highest-leverage use cases are rarely the headline ones, but they are the ones that fund Anthropic's $900bn line by line. ## To watch The Q3 2026 test will be decisive. If margin holds despite the SpaceX bill, Anthropic confirms its status. If it deteriorates sharply, private market comps will be repriced within weeks and the October S-1 will be filed under stress. Open question: at $900bn today, what would justify breaking the symbolic trillion-dollar barrier tomorrow? --- Source: [https://mathieuhaye.fr/blog/en/2026-05-26-anthropic-30-milliards-900-milliards-valorisation-openai](https://mathieuhaye.fr/blog/en/2026-05-26-anthropic-30-milliards-900-milliards-valorisation-openai) | Other language: [https://mathieuhaye.fr/blog/2026-05-26-anthropic-30-milliards-900-milliards-valorisation-openai](https://mathieuhaye.fr/blog/2026-05-26-anthropic-30-milliards-900-milliards-valorisation-openai)