Dear Trader,
Anthropic wants to go public at a valuation north of $2 trillion, and it handed investors about 80 pages on how its own product could go wrong. I'd bet most buyers skim right past that section on the way to the revenue growth.
Then yesterday, the CEOs of the biggest AI companies signed a set of safety standards at the White House. The internet spent the rest of the day roasting their handwriting, and hardly anyone seemed curious about what they'd actually agreed to.
That's why I'm glad Hans Albrecht's in the room today. He's been all over the AI trade this year, from the memory shortage to Anthropic's run at the public market, and he doesn't skip the fine print. He's joining Mark Sebastian live in the State of the Market room at 10 AM ET to dig into the AI stories below.
Mark will also walk through this morning's inflation number, so bring your questions on the Fed too.
This link adds State of the Market to your calendar.
The Daily News Breakdown
SIREN: The Fed's favorite inflation gauge came in cooler than expected for August, and that should take some pressure off an October rate hike. Headline PCE (Personal Consumption Expenditures, the inflation number the Fed actually steers by) rose 3.4% when economists expected 3.7%. Core PCE, which tosses out food and energy, slipped to 3% from 3.3% in July.
Don't get too excited. Core at 3% still sits a full point above the Fed's 2% target. New York Fed president John Williams said Tuesday he sees no rush to hike, but he still wants one more increase "late this year." Most people read that as December, the meeting after October's.
SIGN: Anthropic's leaked IPO prospectus spends roughly 80 of its 261 pages on risk factors, about double the space it gives to describing the actual business. The company warns its own AI could pose catastrophic or even existential risks to humanity, and it flags model behavior that resembles blackmail. SpaceX, whose whole business involves strapping humans to rockets, needed about 38 pages.
The financials come with their own warning label. Anthropic lost more than $8 billion on an operating basis in 2025, even as revenue grew twelvefold.
And yet the deal could still value the company at more than $2 trillion. Yahoo's read: the only risk anybody's pricing in is sitting this one out.
SNEAKER: Trump and the heads of the biggest AI companies sat down for lunch at the White House on Tuesday and signed a set of voluntary AI safety standards. Elon Musk, Mark Zuckerberg, Anthropic's Dario Amodei, Nvidia's Jensen Huang, OpenAI's Greg Brockman and Google's Sundar Pichai all put pen to paper. Trump said the standards work almost like a constitution.
Keep the word "voluntary" in mind, though. You'd figure the internet would tear into what the document actually says, especially after the recent Hugging Face mess and Amodei's own risk warnings. Instead, X spent the day roasting the CEOs' handwriting.