Anthropic had a day of two halves on Monday. In the morning it unveiled two new artificial intelligence models, Claude Fable 5.2 and Claude Mythos 5.2, aimed at strengthening its position against OpenAI, Google and Meta ahead of a planned stock market listing. By the evening it was telling bankers that the pricing details in its IPO filing were wrong.
The company circulated a copy of a securities filing to banks that showed it planned to sell about 867 million shares at $70 to $80 each, a range that would have raised up to $69.3 billion, according to emails seen by Reuters. A follow-up email hours later said the pricing details were inaccurate and that an updated filing would follow. Anthropic declined to comment, and Reuters said it could not determine what was incorrect.
The draft filing also showed $26 billion of investment lined up alongside the listing from some of the world’s biggest technology companies. Google, Nvidia, AMD, Broadcom, Micron, Microsoft and TSMC had agreed to take part in a private placement of Anthropic’s Class N shares at the IPO price, closing immediately after the offering.
Anthropic is still expected to launch its roadshow on Tuesday, with pricing expected the following week, and to list on Nasdaq under the ticker ANT. Until a corrected filing appears, however, there is no confirmed price range, share count or valuation for one of the most closely watched listings of the year.
The confusion comes as expectations for the deal have already cooled. Investors had previously discussed a valuation of as much as $2 trillion for Anthropic, but Reuters has reported that $1.5 trillion is now seen as the high end of the range. The company revealed in an amended filing earlier this month that revenue reached almost $17 billion in the three months to the end of September, although its quarterly growth rate has fallen sharply.
The listing also arrives at a difficult moment for AI infrastructure. Other companies in the sector have delayed or withdrawn planned flotations as rising interest rates fuel a sell-off in shares and bonds of AI operators such as CoreWeave, and as investors ask harder questions about the sustainability of AI spending.
Monday’s product launch, at least, went to plan. Claude Fable 5.2 is aimed at coding, reasoning and AI agent work, while Claude Mythos 5.2 is designed for approved cybersecurity and life-sciences research. Access to Mythos is restricted to organisations approved through Anthropic’s verification programmes, while Fable is available more broadly.
Anthropic said the new models generate outputs up to 30% faster than their Claude 5.1 predecessors and communicate more clearly with users, and priced them at $8 per million input tokens and $40 per million output tokens — lower than previous Fable and Mythos releases. The launch follows the rollout of Claude Haiku 5.5 in October, as the company broadens its product range before going public.
Founded in 2021 by former OpenAI executives including Dario Amodei and his sister Daniela, Anthropic has become one of the leading developers of generative AI. Its rapid growth has been underpinned by Google and Amazon, which are simultaneously investors, suppliers of computing power and important distribution partners — as well as AI competitors. The company has nearly $900 billion in computing commitments with partners including Amazon and Google, underscoring how much of its future capacity depends on financing beyond its control.
Sources: Reuters — “Anthropic unveils new Fable, Mythos models as competition heats up ahead of IPO” (https://www.reuters.com/technology/anthropic-unveils-new-fable-mythos-models-competition-heats-up-ahead-ipo-2026-11-09/) and “Anthropic’s plan for nearly $70 billion IPO briefly circulated to banks, then retracted”, via The Hindu BusinessLine (https://www.thehindubusinessline.com/info-tech/anthropics-plan-for-nearly-70-billion-ipo-briefly-circulated-to-banks-then-retracted/article71685044.ece/amp/)
