AI giant Anthropic files to sell shares to public for first time

Artificial intelligence giant Anthropic filed paperwork for what could be the biggest initial public offering of stock ever, even as its top executive warns that advanced AI could pose an existential threat to humanity.

The filing sets up the San Francisco company for a Wall Street debut next month, potentially eclipsing the record $86 billion raised by SpaceX in its initial public offering in June.

An Anthropic IPO would also put the company ahead of OpenAI in going public, setting the stage for a pair of blockbuster stock debuts by rival AI developers.

An IPO could give Anthropic enormous resources to accelerate its expansion.

The amount the company ultimately raises remains uncertain, as do its final valuation and offering price.

But even at lower estimates, it would turn the five-year-old startup into one of the world's most valuable companies.

The 335-page filing showed Anthropic's revenue is growing rapidly, reaching more than $16 billion in the first half of the year, 14 times its revenue from the same period last year.

But like many other tech companies going public, the company behind the Claude chatbot has yet to turn a profit.

Despite the losses, Anthropic, valued earlier this year at $965 billion by private investors, is reportedly set to see that value rise to $1.5 trillion or possibly $2 trillion.

But while Anthropic has tried to set itself apart from other AI companies as the "good guys," even it admits the risks AI could pose serious risks to the world as we know it.

BERNARD CONDON (The Associated Press): "The other truly frightening thing in the document was a nearly 100-page list of various risks to the company. This is a required disclosure for every publicly traded company, and usually it's boring, quotidian, not particularly surprising. We're not going to be able, we might not be able to raise money selling bonds at a low enough rate, or we could lose some of our key employees to rivals, who will poach them. In this case, the risk factors included things like authoritarian governments could use our AI models to crack down on their citizens. A terrorist might use our AI models to create a chemical or biological weapon to kill us, and the curious and riveting risk of the possibility that all life on planet Earth could get wiped out if one of these AI agents goes rogue and we can't find the kill switch."