If there is a limit to how much AI infrastructure the world needs, we haven't hit it yet.

Neocloud provider CoreWeave said yesterday that its revenue more than doubled in the second quarter, totaling $2.58 billion. Coreweave's revenue backlog—deals that it has not billed yet— rose 246% year-over-year to $104.2 billion in the second quarter. And that figure doesn’t include roughly $25 billion of net new customer commitments already added in early Q3.

“Our near-term capacity remains effectively sold out,” CEO Michael Intrator said on a call with analysts on Tuesday as shares rose nearly 15% in after hours trading.

Fears of excessive capital expenditures and potential over investment in AI infrastructure have flared throughout the past year. And the debate intensified earlier this week as chipmaker Nvidia (which owns close to 13% of CoreWeave) announced a plan to mobilize $500 billion in financing for AI infrastructure in partnership with Apollo Global Management Inc., Blackstone Inc., BlackRock Inc. and Brookfield Asset Management. 

To judge by investors' reaction to Coreweave's results, the debate may be on pause once again.—Amanda Gerut

 

 

Brad Lightcap, one of the most recognizable faces of OpenAI’s senior leadership team, is leaving the company to “start something new,” he announced on Tuesday, representing the latest in a string of executive departures at the ChatGPT maker.

Lightcap, who has served stints as OpenAI's CFO and its COO, did not say what the new venture is, but said he’s “not going far,” and described it as involving “a few important new things the world will need to get right” as artificial intelligence becomes more powerful and capable.

In April, three OpenAI executives announced their departures, including Bill Peebles, who led OpenAI’s now-dead Sora video app, Kevin Weil, the vice president of OpenAI for Science, and Srinivas Narayanan, CTO of B2B applications. And in June, Barret Zoph, the head of enterprise sales, left OpenAI for a second time. Read the full story here. —Emily Forlini

Anthropic is planning to introduce a subtle watermark that can be used to identify AI-generated text.

Starting with models released on or after August 2, Anthropic says it’s weaving an imperceptible, machine-readable signal directly into Claude-generated text. People won’t be able to see it, and the company says it won’t affect quality or readability.

While other AI companies have attempted this kind of watermarking before, it’s been mostly aimed at images. Text has traditionally been more difficult to identify in this way because it gets copied, paraphrased, translated, chopped up and folded into someone else’s writing all the time. And Anthropic acknowledged that its new text watermark is not perfect: The company says the mark only shows that Claude had a hand in something, not that a Claude model generated the entire thing.—Bea Nolan

 

—Spotify asks artists to identify as human or AI. We are the music makers.

—Ex Czar wants a $1 billion. David Sacks' Craft Ventures is raising a new fund.

—Gemini hits 1 billion users. Google's latest product to join the billion-user club.

—SpaceX and Cursor's new app. Meet Grok Bot.

—Why Steve and Connie Ballmer are splitting their $8 billion philanthropy into three.