Skip to the story
Business

Anthropic’s $81.31 billion IPO overtakes SpaceX as computing commitments near $900 billion

Claude developer Anthropic priced shares at $83, raising a record $81.31 billion at an approximately $1.5 trillion valuation. Its Nasdaq debut comes as investors scrutinise the cost of AI growth.

Anthropic’s $81.31 billion IPO overtakes SpaceX as computing commitments near $900 billion

Anthropic has raised $81.31 billion in the world’s largest initial public offering, surpassing SpaceX’s $75 billion June listing as investors continue backing artificial intelligence companies despite concerns about their valuations and mounting development costs.

The Claude model developer set its share price at $83 on Wednesday, exceeding its proposed $70 to $78 range and valuing the business at approximately $1.5 trillion on a fully diluted basis. Trading on Nasdaq is expected to start on Thursday under the ticker ANT, making Anthropic the first major frontier AI model developer to go public.

The offering grew nearly 19%, from 825 million shares to 980 million, after demand reportedly outstripped supply several times over. A separate private placement at the same share price is raising another $26 billion.

The fundraising arrives as investors question whether spending on data centres, sophisticated chips and computing infrastructure can produce lasting returns. Anthropic’s disclosures show a business growing rapidly while facing substantial costs.

Its annualised revenue run rate stood at $81 billion on October 31, up $10 billion from the end of September. Demand for Claude models and Claude Code, its coding agent for software developers, accounted for much of that growth.

By September 30, Anthropic had 11,900 customers each contributing at least $100,000 in revenue over the previous 12 months, against more than 1,500 at the end of 2025. Third-quarter revenue reached $16.96 billion, a 47% increase over the second quarter.

However, a $567 million operating loss replaced the preceding quarter’s $974 million operating income. Its net loss declined from $66.3 billion to $23.3 billion; both amounts included non-operating items tied to bonds convertible into shares.

Computing capacity and infrastructure spending approached $9 billion in the first six months of 2026, almost quadruple the corresponding year-earlier amount. Computing costs added another $5 billion in the third quarter.

Anthropic has also committed nearly $900 billion to computing expenditure with partners including Alphabet, Google’s parent, and Amazon, expanding capacity to train and run more sophisticated models. Some providers depend on outside borrowing and credit support to buy chips and build facilities. Anthropic warned in its prospectus that financing difficulties could disrupt contracted capacity’s availability, price and delivery.

Competition is intensifying, with OpenAI potentially seeking a listing as early as 2027. According to media reports citing September data from corporate expense platform Ramp, OpenAI’s GPT-6 Astra represented approximately 13% of enterprise AI spending tracked by Ramp, versus around 8% for Anthropic’s Claude Fable. OpenAI’s annualised revenue run rate exceeded $40 billion in July, though Anthropic was still ahead then.

Open-weight models pose another challenge, allowing businesses to download and operate AI themselves instead of depending entirely on commercial providers. A Capgemini study found 59% of executives at large companies intended to expand their use over the following year as AI costs overshot budgets. Major Anthropic customer Meta Platforms has also been developing internal AI capabilities while seeking to reduce dependence on Anthropic’s technology.

Commercial growth also raises questions about Anthropic’s safety commitments. Chief Executive Officer Dario Amodei has previously cautioned that powerful AI could advance faster than society can manage its risks. The prospectus acknowledged that future models might gain capabilities before suitable evaluations and safeguards exist, increasing the potential for harm.

Anthropic is considering stronger human oversight of some research and development decisions as AI takes on more of that work. Goldman Sachs and Morgan Stanley are leading the flotation with support from a 22-bank underwriting syndicate; both also led SpaceX’s June IPO.

Share this story

Send it to someone who should read it.

Join the conversation

Comments

Comment rules

More in Business

See all