Anthropic hires ex-Stripe CTO for AI infra

Just a heads up, if you buy something through our links, we may get a small share of the sale. It’s one of the ways we keep the lights on here. Click here for more.

Anthropic just hired a new chief technical officer, and it’s someone with serious cloud cred. 

Rahul Patil, who most recently wore the CTO badge at Stripe, is taking over the role from company co-founder Sam McCandlish. 

Don’t worry, McCandlish isn’t going anywhere. He’s moving into the newly created role of chief architect, where he’ll focus on the glamorous world of pre-training and large-scale model training. Both will report to Anthropic president Daniela Amodei. 

The shuffle isn’t just about new business cards. 

Anthropic is also rejiggering its core tech structure to bring product engineering, infrastructure, and inference closer together, basically trying to get its builders, maintainers, and model whisperers all rowing in the same direction. 

As CTO, Patil gets the unenviable job of keeping Anthropic’s infrastructure humming while demand for Claude keeps spiking. (Via: TechCrunch)

McCandlish, meanwhile, gets to dream bigger about how to make future Claudes smarter.

The timing matters. OpenAI and Meta are dumping staggering sums into compute: Mark Zuckerberg casually promised $600 billion in US infrastructure by 2028, while OpenAI’s Stargate project with Oracle is said to cost around the same. 

Anthropic’s budget isn’t public, but it’s safe to assume they won’t be outspending Zuck. That means squeezing more speed and efficiency out of every GPU rack.

They’ll need it. Claude’s growing popularity already forced Anthropic to slap rate limits on power users this summer, because some folks apparently thought running Claude Code like a 24/7 background process was a great idea. 

New limits now cap Sonnet usage between 240 and 480 hours a week and Opus 4 between 24 and 40 hours. (Still plenty of time to automate your side hustle.)

Patil, with 20 years in the trenches at Stripe, Oracle, Amazon, and Microsoft, is being brought in as the infrastructure adult in the room. 

Amodei praised his “proven track record” in keeping enterprise systems stable. For his part, Patil sounded downright starry-eyed about the mission: “I can’t think of a greater calling,” he said.

Is Anthropic’s leadership reshuffle a smart move to scale infrastructure, or a sign they’re already struggling to keep up with OpenAI and Meta’s spending power? Should AI companies focus more on efficiency and smarter architecture, or is throwing billions at compute the only way to stay competitive? Drop us a line below in the comments, or carry the discussion to our Twitter or Facebook.