Anthropic is offering 801.4 million shares, while its backers are selling 23.6 million. The fundraising includes a $26 billion private placement to Advanced Micro Devices, Broadcom, Google, Micron Technology, Microsoft, Nvidia and Taiwan Semiconductor Manufacturing Company.

At the top of the price range, Anthropic would have a market value of about $1 trillion based on outstanding shares. Including employee stock options, restricted shares and stock reserved for its charitable donation-matching programme, its fully diluted value would be about $1.4 trillion.

The fundraising could reach roughly $100 billion after including an over-allotment option, or greenshoe, Bloomberg calculations showed. That would exceed SpaceX's $75 billion base offering and the $86.2 billion it raised after its greenshoe was exercised.

The target comes after a sharp sell-off in AI-related stocks last week. The decline was driven by expectations that politicians promising tighter regulation of data centres and related infrastructure could increase costs and restrict the technology's development. Investors must assess whether AI valuations reflect its expected economic and social impact or excessive confidence in an investment theme yet to prove its staying power.

The adoption of Claude Code and rival tools among software developers indicates their potential use in areas ranging from drug discovery and legal and accounting work to factory operations. More than 90 per cent of contributions to Anthropic's code base were written by Claude, the filing showed.

Anthropic has also introduced industry-specific products, particularly for healthcare and life sciences. Novartis Chief Executive Officer Vas Narasimhan is a board member. The IPO is expected to be priced on Wednesday, 18 November, according to deal terms reviewed by Bloomberg.

Third-quarter revenue rose to nearly $17 billion as of 30 September from $1.4 billion a year earlier. Anthropic on Monday unveiled faster versions of its Mythos and Fable models, with Fable also priced lower as some companies reduce spending on AI.

The company has $895 billion in commitments, mainly for computing power and chips. About $70 billion is due in the year through September 2027, an amount larger than the individual market capitalisations of roughly two-thirds of S&P 500 companies.

Unlike technology companies that can rely on income from businesses such as advertising, Anthropic must convince investors that its AI-only focus can generate profits exceeding its rising infrastructure costs. SpaceX linked its ambitions to AI during its IPO pitch, though most of its revenue at the time came from other businesses.

Anthropic is backed by funds across Fidelity Investments, Capital Group and BlackRock, and venture capital firms including Iconiq, Spark Capital and Sequoia Capital. Google, Amazon and Microsoft are investors as well as customers and major suppliers of computing power.

The pace of AI development has also raised safety concerns. Anthropic Chief Executive Officer Dario Amodei has called on leading laboratories to “pace the frontier” and prioritise safety. The company warned in its filing that AI could pose “catastrophic or existential risks to humanity” without adequate safeguards and oversight.

Regulators, investors and Anthropic employees have raised the risk of humans losing control of AI. One employee put the chance of AI eliminating humanity within the next decade at more than 10 per cent, while Bill Gates said it was powerful enough to drive events causing “a billion deaths”.

Rival OpenAI is targeting a 2027 listing, while Chief Executive Officer Sam Altman called 2026 an “ill-advised moment” for an IPO because of AI safety concerns. OpenAI is raising funds at a $1.25 trillion valuation, people familiar with the matter said.

Anthropic was founded nearly six years ago by former OpenAI executives who wanted to prioritise safety. It is structured as a public benefit corporation, with a Long-Term Benefit Trust empowered to elect and remove a majority of its board. Its seven co-founders each hold stakes of about 1.4 per cent and would remain billionaires several times over even if they fulfil pledges to donate 80 per cent of their personal equity to philanthropic causes.

Operating expenses rose 292 per cent from a year earlier to $17.5 billion in the third quarter. After recording a $974 million operating profit in the second quarter, its first following at least five quarters of losses, Anthropic posted a $567 million operating loss in the three months ended 30 September.

The loss was mainly driven by $5 billion in computing costs under short-term commitments and $2 billion in non-cash charges linked to its charitable donation-matching programme. Lower model prices and customer discounts also reduced third-quarter revenue growth by about $600 million.

Anthropic cut prices for newer models, including Opus 5.5 in September, while OpenAI has also released cheaper models. Both face competition from lower-cost open-weight models, including those developed in China, which customers can download and customise.

Goldman Sachs and Morgan Stanley are leading the IPO, with 20 other banks participating. Anthropic's shares are expected to trade on the Nasdaq Global Select Market under the symbol ANT, as the company seeks a record fundraising while managing rising costs, pricing pressure and safety concerns.