Is Anthropic losing its Oura?
The AI lab’s IPO prospectus has leaked but it faces a narrowing window to list before the end of the year
The IPO ‘risk factor’ keeping bankers up at night
Wall Street is watching closely as a possible fourth-quarter fee bonanza comes down to the wire. Reporters obtained the details of Anthropic’s IPO prospectus earlier this week but the AI lab still hasn’t officially released the document, even as it’s reportedly aiming for a November listing.
On Tuesday yet another tech group, the smart ring-maker Oura, postponed its own planned listing, blaming “uncertainty in the IPO market”. In recent weeks SoftBank’s US data centre developer SB Energy, the nuclear energy group Holtec International, which hopes to supply power to AI data centres, and OpenAI have all delayed their own IPO plans.
Meanwhile, Silicon Valley leaders and politicians are being forced to respond to growing anxiety over AI algorithms and data centres.
Donald Trump, who has called AI safety concerns a “hoax” and a “sick conspiracy”, hosted tech bosses including Anthropic’s Dario Amodei at the White House on Tuesday to discuss AI risks. (Trump endorsed “tremendous self-regulation” of the sector after Tuesday’s lunch.) Democrats are capitalising on voter anxiety about AI and data centres as they campaign to win back Congress in the midterm elections.
For now Anthropic’s investors are confident the group can list at a valuation of more than $2tn. According to the AI start-up’s S-1 filing, the group said it planned to spend $518bn on cloud, computing and infrastructure obligations in the coming years to support its rapid growth.
Anthropic’s prospectus also revealed details of its prior financial performance and governance arrangements, which were first reported by Reuters and confirmed to the FT by multiple people who have seen the filing.
Last year the company recorded an operating loss of more than $8bn as its revenue jumped 12-fold to almost $4.6bn and mounting costs for computing resources to train and run its models pushed operating expenses to almost $13bn.
This year Anthropic’s spending and revenue have both increased sharply. Revenue in the second quarter of this year was $11.5bn, and it’s on course for its second consecutive quarter of operating profit, albeit on a heavily adjusted basis.
But even more notable than the group’s heavy spending is its rather unconventional “risk factors” section, which takes up nearly a third of the filing.
Anthropic details risks including the potential of increasingly advanced AI models to manipulate, blackmail and exhibit other unpredictable behaviours. It also warns that its technology may pose “existential risks to humanity”.
DD can’t predict whether AI will wipe out the human race. But we can warn with some certainty that if Anthropic and OpenAI don’t go public soon, it sure is going to feel like the end of the world for bankers, tech funds and private capital giants.