Kevin Weil raising $150M for AI science startup

Share
Kevin Weil raising $150M for AI science startup

The AI talent exodus from OpenAI keeps producing new companies: former chief product officer Kevin Weil is raising a nine-figure round for an AI-for-science startup, and CoreWeave's earnings showed the infrastructure buildout underneath it all is still accelerating.

Kevin Weil, OpenAI's former chief product officer, is seeking to raise $150 million for a new AI science startup at a valuation of at least $750 million, according to three people with knowledge of the discussions. Business Insider reports the details are not finalized and could still change, and Weil did not respond to a request for comment. Sources say he has pitched the company as a way to gather scientific data for AI models — a thesis that extends the work he led at OpenAI, where he oversaw OpenAI for Science and Prism, a workspace designed to put frontier models directly into scientists' workflows before his departure in April.

Weil is the second senior OpenAI leader this week to announce a next act — OpenAI's Brad Lightcap exits after eight years to start something new. The pattern is the story: a growing wave of OpenAI and Google alumni are starting companies that promise to accelerate scientific discovery, from designing drugs to running autonomous experiments, and investors are paying up for the thesis. A $750 million valuation on an unrevealed product is a bet on Weil's track record — product chief at Instagram, Twitter, and Meta — more than on anything he has shown yet. That is how this corner of the market is moving: the next frontier is being priced before it ships.


CoreWeave's Q2 report showed AI infrastructure demand still running hot: revenue rose 112% year over year to $2.58 billion, slightly ahead of the $2.56 billion analysts expected, and the revenue backlog hit $104 billion as of June 30. The company said that figure excludes another $25 billion in Q3 commitments, and it ended the quarter with 1.5 GW of contracted power. Adjusted EBITDA reached $1.51 billion, adjusted operating income of $128 million beat the $66 million consensus, and the net loss per share of $1.14 came in better than the $1.41 expected. Shares jumped as much as 12% after hours.

The print is an answer to the bear case that had dragged the stock down more than 30% since May over spending and financing concerns. CEO Michael Intrator called it "an important inflection point" as scale translates into operating leverage — though adjusted net loss still widened to $567 million from $130 million a year ago, so the leverage story is early. With SpaceX now renting its own data-center capacity to Anthropic and Google, CoreWeave also has a new competitor that was a customer's parent company a year ago. For now, the numbers argue the GPU cloud's economics improve as it grows; the question is whether that holds as the buildout gets more crowded.

What to watch: whether Weil's round closes at that valuation — and what it says about how much money is chasing AI-for-science startups.

Frontier-lab executives keep leaving to automate scientific discovery — is that the next real AI market, or the next hype cycle? Tell us in the comments.

Read more

Nvidia-backed Firmus pulls its $30 billion IPO after demand collapses

Nvidia-backed Firmus pulls its $30 billion IPO after demand collapses

The biggest AI-infrastructure listing of the year didn't get delayed — it got pulled entirely, after two price cuts in two days found no takers. Nvidia-backed data-center operator Firmus withdrew its Australian IPO on Friday, abandoning what would have been the country's second-largest listing ever after slashing the offer price twice in 48 hours. The company had been marketing shares at A$11 each to raise about $5 billion at a valuation near $30 billion — roughly three times the $10.5 billion

AI inference is redrawing the storage hierarchy

AI inference is redrawing the storage hierarchy

The money has followed GPUs for years, but at GMIF 2026 in Shenzhen the argument was about everything behind them — and the numbers backing it are getting hard to ignore. China's 140 trillion daily token calls are forcing a redesign of the storage stack. Now in its fifth year, the Global Memory Innovation Forum gathered storage makers, analysts and chip designers around one theme: inference, not training, is now the workload that dictates hardware. China's daily token calls had already passed

Cloudflare launches Clef-omni, cuts Clef-flash price by 58%

Cloudflare launches Clef-omni, cuts Clef-flash price by 58%

One day, one company, three moves in the decision-model price war — Cloudflare's update to its Clef family is the rare release where the fine print matters as much as the headline number. Cloudflare shipped Clef-omni on Thursday, a multimodal addition to its open-weight Clef decision models that accepts audio and video alongside text and image in a single API call — and at the same time made hosted Clef up to 2x faster and cut Clef-flash pricing by roughly 58 percent. Clef-omni takes WAV or MP

Meta turned down Amodei's personal plea for compute

Meta turned down Amodei's personal plea for compute

The chip hunt is back in the headlines — and today's edition runs from a boardroom ask at the very top of the AI industry to the economics of putting robot drivers in truck cabs. Dario Amodei personally approached Meta earlier this year to source more compute for Anthropic, and Meta declined — according to the Wall Street Journal. The report lands inside the Journal's larger piece on how desperately the industry is hunting for computing power, and it is the kind of detail that only surfaces wh