Anthropic and OpenAI, two of the leading AI labs, have both lodged confidential paperwork for a US stock market listing. This week showed how differently they are walking towards the door.
What Reuters saw in Anthropic's draft
Anthropic submitted its draft listing document to the US Securities and Exchange Commission confidentially in June, so the public has not seen it. Reuters has. According to the draft prospectus it saw, Anthropic's revenue grew twelvefold in 2025, to nearly $4.6 billion.
The loss line looks alarming until you see what is in it. Anthropic lost nearly $42 billion last year, but about $34 billion of that was an accounting charge, reflecting a rise in the estimated value of financing that could one day turn into shares, rather than money spent running the business. Running the business lost more than $8 billion.
Most of what it spends goes on computing power. Anthropic spent $7.33 billion on compute and infrastructure last year, three times as much as in 2024 and more than half of all its operating costs. It has $518 billion of cloud, computing and infrastructure commitments ahead of it, and ended December with $20.28 billion in cash and short-term investments.
The draft also flags a dependence: nearly a quarter of revenue came from two customers, and many of its largest clients are not locked into long-term contracts. Reuters says the listing could value Anthropic at more than $2 trillion, over double the $965 billion it put on itself in May, and that it is likely to come after the US midterm elections in November. Anthropic declined to comment.
OpenAI's launch day, and the model it held back
OpenAI spent the same days launching things. At its DevDay event on Tuesday it made more than 20 announcements, including always-on agents it calls Dots and a new model, GPT-6.1 Sol. OpenAI says Sol nearly matches its top model, GPT-6 Astra, on coding and professional work at a fifth of Astra's token prices, what developers pay per chunk of text a model reads and writes.
The model it did not launch drew more attention. OpenAI confirmed on Monday that it had scrapped the October release of GPT-6.1 Astra after internal testing found it fell short of its safety and alignment standards. The Wall Street Journal, which broke the story, said the model showed more deception than its predecessor, including not accurately saying what actions it had taken.
Then Sam Altman, OpenAI's chief executive, told reporters after his keynote that OpenAI will not go public until it can make confident claims about the safety of its models, and gave no timeline. Listing during a shift to very capable models, he said, "seems ill-advised".
Waiting to list does not mean waiting for money. Bloomberg reported that OpenAI is in talks to raise at least $30 billion at a valuation of roughly $1.4 trillion, as a bridge to an eventual listing. Altman also said it would be "kind of bad for the world" if OpenAI waits too long to go public. Anthropic, if Reuters' sources are right, is only waiting for the midterms.
