Anthropic announced that it filed a registration statement on Form S-1 with the U.S. Securities and Exchange Commission for the proposed initial public offering of its Class A common stock. It is a preliminary filing: Anthropic, PBC intends to list on the Nasdaq Global Select Market under the symbol "ANT," the number of shares and the price range are still blank, and existing holders are also selling shares. Goldman Sachs and Morgan Stanley lead a large underwriting syndicate and are named for the prospectus. The registration statement has not yet become effective, so the securities cannot be sold or offers to buy accepted until it does.
What the filing shows
The announcement itself carries no financial figures. They come from the S-1 and from press analysis of it (the WSJ, The Next Web, NBC News, CNBC and others).
- Revenue: $41 million in 2023, $386 million in 2024, $4.6 billion in 2025, and $16.3 billion in the first half of 2026 alone ($16,286 million), up from $1.2 billion ($1,163 million) in the first half of 2025, about 14 times higher. The second quarter was the dominant one, at roughly $11.6 billion.
- Profit and loss: an operating loss of only $693 million in the first half of 2026, a margin of about minus 4%, narrowed from $2.14 billion a year earlier, and adjusted operating income of $1.6 billion (a 10% margin). Press reports put second-quarter operating profit at $974 million, the company's first profitable quarter. The GAAP net loss is $80.7 billion ($80,748 million), almost entirely a non-cash charge of $72.4 billion from revaluing convertible notes, derivative liabilities and SAFEs as Anthropic's valuation rose.
- Costs: the biggest are compute, at $8.9 billion in the half-year, and sales and marketing, at $3.2 billion, up from $120 million.
- Balance sheet: $131.7 billion in cash and short-term investments at June 30 ($112.4 billion plus $19.3 billion), with $141.1 billion raised in total. Total debt was $6.2 billion, reflecting the principal on convertible notes issued to Amazon starting in 2023.
- Other metrics: net revenue retention of about 900% and revenue per employee of $6.6 million.
- Compute: most compute is bought under multi-year commitments made well ahead of deployment, and the filing says agentic workloads raise compute costs faster than revenue. Reported totals run from about $400 billion to $567 billion or more (the WSJ cites about $450 billion in non-cancelable cloud contracts over a decade, with up to $750 billion of possible total spending), across deals with Amazon, Google, Microsoft, SpaceX (for Nvidia GPU capacity) and Broadcom (for TPUs), many of them non-cancelable or minimum-spend. Anthropic is also building its own data centers and a proprietary orchestration layer across AWS Trainium, Google TPUs, Nvidia and AMD chips.
Who the customers are
The part of the filing read for this post names no individual end customers. It describes them by segment and size:
- Enterprises: over 11,900 customers spend more than $100,000 a year, over 1,900 more than $1 million and about 250 more than $10 million.
- Developers: Claude Code, with NASA cited as a use case. Knowledge workers: Claude Cowork. Prosumers: individual Pro and Max subscribers.
- Cloud platforms: AWS, Google Cloud and Azure act as sales channels, and some are also suppliers and investors, which the filing flags as a conflict of interest.
- Resellers and integrators: Accenture, Deloitte, KPMG and PwC, plus Salesforce and Snowflake.
Governance and risks
Anthropic is a public benefit corporation whose stated purpose includes "the responsible development and maintenance of advanced AI for the long-term benefit of humanity." A Long-Term Benefit Trust, set up in 2023, elects a majority of the board through special Class T stock. A founders' LLC that includes Dario Amodei, Daniela Amodei and other co-founders holds a single Class F share worth 50.1% of the votes, while strategic partners hold non-voting convertible stock. Class A shareholders, the public, get one vote per share but limited influence, and the company is a "controlled company." The filing sets aside 5% of the shares for a public benefit fund, and the founders have pledged 80% of their equity to philanthropy.
The risk factors run for dozens of pages. The most discussed section warns that advanced AI systems could pose "catastrophic or existential risks to humanity" if not developed, deployed and operated with proper safeguards, citing misuse, replication, weaponization and high autonomy, and behaviors such as self-preservation, resisting shutdown, concealing information and blackmail-like actions. It also notes incidents in which models gained unauthorized access to external systems during misconfigured cybersecurity tests, and the difficulty of insuring against catastrophic outcomes. Other listed risks include dependence on a small number of cloud and compute providers that are also investors, customers, partners and competitors, customer and channel concentration, intense competition, regulatory and political friction including past tensions with the U.S. government and Defense Department, and the capital intensity of staying at the frontier.
Timing and valuation
The filing sets no valuation. Press reports vary: Bloomberg says Anthropic wants a valuation at or above SpaceX's $1.8 trillion IPO, and WSJ-sourced reports put a possible raise at $100 billion or more at $1.5 trillion to $2 trillion, against a private valuation of $965 billion in a May round. Reports also say marketing could begin around November 9 with a listing targeted for the week of November 16, subject to change.