The Register's Kettle podcast discussed the reported leak of Anthropic's IPO prospectus, which presenter Brandon Vigliarolo described as full of "doom, gloom, and to the moon promises." Reporter Lindsay Clark noted the oddity of a company selling itself by warning that its technology carries existential risk. The leaked documents, first reported by the Financial Times, also offered a look at Anthropic's finances: an $8 billion loss on $4.6 billion of revenue, plus plans to spend roughly $500 billion on cloud computing over the next few years.
Editor-in-chief Matt Rosoff said the numbers suggest the more money Anthropic earns, the more it has to spend, making margins look poor. He mentioned a report that Anthropic turned profitable on an adjusted basis in Q2, but only after stripping out training costs. Last year's operating loss was almost double its revenue, he said, so unless expenses come under control, the business model is hard to justify.
The panel contrasted Anthropic's IPO push with OpenAI's decision to pause its own IPO. Rosoff speculated that OpenAI's numbers might look even worse, and that Anthropic's safety-focused narrative could be "spackle" over bad financials. Clark and Rosoff agreed that leading with existential risk is a strange way to court investors, though Rosoff stressed that this was his speculation rather than a confirmed conclusion. The source is a single podcast discussion, so no independent reporting was available to verify the{