Hello,
The AI boom hit new limits this week: who owns the chips, where the data centers can go and who keeps AI agents in check. The Information’s reporters are inside each of these stories. Subscribe for $299 and save 25% on your first year to follow where they lead.
Here are five of this week’s top stories.
Anthropic’s AI is designing its own chips
Anthropic has told IPO investors that a small team is already using Claude to design its custom chips, and the news sent shares of Synopsys and Cadence lower. Our exclusive on Claude’s chip design skills raises the question of whether Anthropic will ever sell that ability to others.
OpenAI’s new communications chief
Uber’s Jill Hazelbaker is joining OpenAI as corporate affairs chief, a job that has sat empty since January. She arrives as Anthropic pulls ahead and OpenAI’s own IPO slips to next year.
Governors vs. the data center boom
Election night put governors who want to pause new data centers in office in states central to the AI build-out, including Texas and Ohio. Our reporting on the blue wave and data centers looks at how far those pledges could go, and what they mean for the companies counting on those sites.
AI agents nobody knew were there
Security firms keep finding AI agents running inside big companies that no one knew about, sitting in sensitive data for months. That has turned the effort to keep AI agents under control into a hot area for deals and startups.
The investor outspending the neoclouds
Apollo is lining up deals to buy chips for SpaceX and Anthropic and lease them back, which would give it a bigger stockpile than any neocloud. What that means for the startups that built this market is the open question.
Chips, politics and security are all setting new limits on how fast AI can grow. How those limits play out is still open.
Before it’s everywhere, it’s in The Information. Subscribe for $299 and save 25% on your first year to follow where the AI build-out goes next.