MW Anthropic's latest hire led AI at Tesla, worked at OpenAI and received high praise from Elon Musk

By William Gavin

Meanwhile, ahead of SpaceX's major IPO, Musk is tying that company more tightly to the AI lab

Andrej Karpathy, who Elon Musk once called "arguably the #2 guy" in computer vision, is joining Anthropic.

Tesla's former head of artificial intelligence, Andrej Karpathy, is joining Anthropic - while his old boss, Elon Musk, ties SpaceX more tightly to the AI startup.

"I think the next few years at the frontier of LLMs will be especially formative," Karpathy, who was also a founding team member at OpenAI, said in a statement on Tuesday. "I am very excited to join the team here and get back to [research and development]."

Karpathy has joined Nick Joseph's pretraining team, responsible for the large-scale training runs that gives the Claude large language model its core knowledge and capabilities, according to Anthropic. He'll be starting a team focused on using Claude to accelerate pretraining research.

While at OpenAI, Karpathy worked on deep learning for generative-AI models and deep reinforcement learning. In June 2017, he was hired by Tesla $(TSLA)$ CEO and OpenAI co-founder Elon Musk to help develop the electric-vehicle company's Autopilot driver-assistance feature.

At the time, Musk called Karpathy "arguably the No. 2 guy" in the world in computer vision, according to email correspondence made public as part of Musk's recent lawsuit against OpenAI. In addition to leading the computer-vision team at Tesla, Karpathy also briefly worked on Tesla's Optimus humanoid robot program.

See more: Anthropic and OpenAI are following Palantir's playbook as they seek to grow AI usage

Karpathy remained at Tesla until 2022, shortly after layoffs hit the Autopilot unit. He later returned to OpenAI for a short stint and then launched an AI education startup called Eureka Labs. On Tuesday, he said he plans to resume his work on education "in time," without providing additional details.

Karpathy's move to Anthropic comes shortly after Musk's SpaceX announced a deal to provide Anthropic's AI labs with computing power from its Colossus 1 data center in Tennessee. Anthropic also said it was interesting in working with SpaceX to "develop multiple gigawatts of orbital AI compute capacity."

The deal marked a change of heart for Musk, who had earlier claimed that Anthropic's AI was racist and accused the company of hypocrisy. He said this month that he now thinks the Claude chatbot will "probably be good" and that no one he met at Anthropic set off his "evil detector."

Anthropic, recently valued at $380 billion, on Monday said it had acquired the popular developer-tool startup Stainless. On Tuesday, Anthropic also announced a "global alliance" with KPMG that gives the consulting firm's more than 276,000 employees access to Claude.

See more: Anthropic has 'unprecedented' demand - and it's leaning on Amazon for support

Both Anthropic and SpaceX, along with OpenAI, are expected to pursue initial public offerings in the near future. SpaceX is expected to file paperwork for its IPO as soon as this week.

-William Gavin

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

May 19, 2026 21:11 ET (01:11 GMT)

Copyright (c) 2026 Dow Jones & Company, Inc.