Anthropic Revises IPO Target To $64.4 Billion, Total Fundraising Could Hit $90.4 Billion

Anthropic's proposed initial public offering (IPO) is now expected to raise as much as $64.4 billion, according to its latest draft prospectus, revising an earlier estimate of $69.3 billion. Combined with a separate $26 billion private placement backed by major technology companies, the artificial intelligence firm's total fundraising could reach $90.4 billion, making it one of the biggest capital-raising exercises in technology industry history.

The latest prospectus, seen by Reuters on November 10, indicates that Anthropic and some existing shareholders plan to sell shares at an indicative price range of $70 to $78 apiece. An earlier filing had outlined a 80 price range and a potential $69.3 billion offering. Anthropic subsequently withdrew that document after informing bankers that its pricing details were inaccurate.

The $26 billion private placement will involve major technology companies, including Alphabet's Google, Nvidia, Microsoft, Advanced Micro Devices (AMD), Broadcom, Micron Technology and Taiwan Semiconductor Manufacturing Company (TSMC). These companies are expected to purchase shares at the IPO price. Their participation highlights the close financial and infrastructure relationships shaping the global AI industry, as several investors also supply the chips, cloud services and computing capacity needed to operate advanced models.

The fundraising push comes as Anthropic strengthens its position in generative AI. On Monday, the company launched Mythos 5.2 and Fable 5.2, introducing improvements in coding, reasoning, visual understanding and response times. It also reduced Fable 5.2's input token price from $10 to $8 per million tokens and promised responses around 30 per cent faster, intensifying competition with OpenAI's GPT-6 Astra.

Anthropic also introduced accelerated inference at twice the standard price and expanded enterprise data protections through zero data retention policies under its Enterprise Frontier Safeguards system.

The latest models arrive amid growing scrutiny of AI safety and evaluation. Anthropic's testing suggests that its systems are increasingly capable of recognising when they are being assessed, raising questions about whether controlled evaluations accurately reflect their behaviour in real-world environments.

Founded in 2021 by former OpenAI executives Dario Amodei and Daniela Amodei, Anthropic has emerged as a major AI developer through its Claude family of models. Investors had previously discussed a valuation of as much as $2 trillion, although expectations have moderated, with $1.5 trillion now considered the upper end of the range, according to earlier Reuters reporting.

Google and Amazon have supported its expansion through investments and infrastructure partnerships, even as both compete in the AI market. Anthropic had nearly $900 billion in computing commitments with partners, including the two companies, according to an earlier report.

The company reported revenue of nearly $17 billion for the three months ended September 30. However, its quarterly growth rate had slowed sharply, according to an amended filing earlier this month.