From AI safety startup to Wall Street: Anthropic’s five-year race towards an IPO
Anthropic’s journey from a small AI safety startup to a company valued at $965 billion in private fundraising rounds sets the stage for a closely watched IPO. But its financial commitments and shareholder structure present important questions for investors.
Just five years ago, Anthropic was raising money to make artificial intelligence safer and more reliable. Now, the company behind Claude is preparing to ask public investors to back a business whose ambitions, revenues and computing bills have grown at extraordinary speed. Anthropic publicly filed for an initial public offering on Monday (October 26) following a confidential submission in June. It plans to list on Nasdaq under “ANT”, according to its official announcement.
The share count and price range remain undecided. The filing moves the startup towards a listing that could help establish how public markets value the companies building the world’s most advanced AI models. A startup built around safety Led by siblings Dario and Daniela Amodei, Anthropic announced a $124 million Series A round in May 2021, led by Skype co-founder Jaan Tallinn. Its stated purpose was to develop AI systems that people could better understand and control.
A $580 million round followed in April 2022. Anthropic’s announcement described a team of around 40 people working from a San Francisco office, with funding earmarked for infrastructure to study increasingly powerful models. That combination of safety research and expensive computing would shape its growth. Anthropic presented the assistant as a tool for writing, summarising, answering questions and coding, with early partners including Notion, Quora and DuckDuckGo.
In May, it announced another $450 million funding round, led by Spark Capital with participation from Google and others. Amazon then committed up to $4 billion in September 2023 and announced another $4 billion in November 2024. Amazon’s statements show the partnership extended beyond financing: AWS supplied infrastructure and distribution, becoming Anthropic’s primary training partner. Products advanced alongside the funding.
Claude 3 arrived in March 2024. Claude Code became generally available in May 2025, while September’s Claude Sonnet 4.5 release emphasised longer coding tasks and improved reasoning, according to Anthropic. The valuation accelerates The fundraising numbers soon changed scale. Anthropic’s official announcements put its valuation at $61.5 billion after a $3.5 billion round in March 2025, then $183 billion after raising $13 billion that September.
In February 2026, it raised $30 billion at a $380 billion valuation. By May, another $65 billion round valued the company at $965 billion. Those are private fundraising valuations, rather than a settled IPO price. Anthropic’s June 1 confidential filing opened the path to public markets; Monday’s disclosure gives prospective shareholders a closer look at the business behind those numbers. Safety meets commercial pressure The company’s safety commitments have also shaped what it sells.
In April 2026, Anthropic introduced Mythos Preview through Project Glasswing, initially restricting access to selected organisations for cybersecurity work because of the model’s capabilities, its filing said. Those commitments also contributed to a confrontation with the US government. According to the filing, the administration moved to restrict Anthropic’s government business after it opposed mass domestic surveillance and fully autonomous weapons.
The document described mixed court outcomes: an August ruling largely in Anthropic’s favour, followed by a September appeals-court decision upholding a separate supply-chain risk designation. Copyright disputes brought another cost. The filing confirmed that a court approved its $1.5 billion settlement with authors in July 2026, while other claims remain contested. What the financials reveal Anthropic’s filing reports revenue of $16.28 billion in the first half of 2026, against $1.16 billion a year earlier.
The April-June quarter generated a $974 million operating profit. Yet the first-half net loss reached $80.74 billion. Much of that reflected non-cash valuation changes involving convertible financing and other financial instruments. It should not be confused with cash spent running the business: operating activities generated $3.92 billion during the half. The infrastructure bill remains substantial.
At June-end, expected future fixed or minimum guaranteed purchase commitments stood at about $265.6 billion, spread over one to 20 years, according to the filing. 3 billion revenue and $80.7 billion net loss An unusual proposition for shareholders Anthropic is approaching the market as a public benefit corporation. Its prospectus said directors must balance shareholders’ financial interests with the interests of people affected by the business and its mission of developing AI for humanity’s long-term benefit.
Its Long-Term Benefit Trust generally has authority to select a majority of directors, while a founders’ entity is set to retain majority voting power on matters submitted to shareholders generally. For public investors, that makes the proposed IPO a test of more than demand for Claude. They will be assessing whether extraordinary growth can support enormous commitments and how much influence they will have over a company built to pursue goals beyond profit.
