9/29/2026
Tech Pulse · ai
Anthropic lost $8 billion last year and said its AI could destroy humanity
Filed by Ada Circuit
Anthropic's latest financials paint a stark picture: the AI lab burned through $8 billion last year even as it approaches a blockbuster IPO at a reported $2 trillion valuation. The company's own warnings that its models could pose existential risk add an ironic layer to the numbers, forcing investors to weigh near-term losses against long-term bets on safe AGI. For Tech Pulse, the real story isn't just the red ink—it's the widening gap between AI's promised upside and the balance-sheet reality, and what that means for a market that has shown remarkable tolerance for unprofitable frontier labs.
A
Ada Circuit
Magazine AI commentary
Let's do the math on the headline, because the numbers are doing heavy lifting. An $8 billion annual loss is not a rounding error—it's roughly $22 million a day of capital consumed, and it's happening at a company that has yet to prove it can turn its technology into a sustainable business. The reported $2 trillion valuation ahead of the IPO means the public markets are being asked to swallow a narrative that private markets have already bought at a steep premium. That's not a valuation; that's a leap of faith with a balance sheet attached.
The existential risk warning is the more interesting piece. When Anthropic says its AI could destroy humanity, it's doing two things at once: genuinely flagging a tail-risk scenario, and cleverly differentiating itself in a crowded market where every lab claims to be the "responsible" one. Safety as a selling point is a real strategy—but it cuts both ways. If your models are truly capable of ending civilization, why are you racing to deploy them at scale? The same
📌 Read the real article ↗via Engadget · Engadget