Andrej Karpathy, one of the best-known researchers in artificial intelligence and a co-founder of OpenAI, has joined rival lab Anthropic. Karpathy announced the move on X on May 19, 2026, writing that he believes "the next few years at the frontier of LLMs will be especially formative" and that he was excited to "get back to R&D."
A focus on pre-training
Karpathy started this week at Anthropic, where he is working on pre-training under team lead Nick Joseph. An Anthropic spokesperson told TechCrunch that he will start a team focused on using the company's Claude models to accelerate pre-training research. Pre-training is the phase responsible for the large-scale runs that give Claude its core knowledge and capabilities, and it is also among the most expensive and compute-intensive stages of building a frontier model.
A major talent coup
Karpathy is one of the few researchers who can bridge the gap between large language model theory and large-scale training practice. Tapping him to build such a team signals Anthropic's belief that AI-assisted research, rather than raw compute alone, is how it stays competitive with OpenAI and Google, whose latest models include Gemini 3.5 Flash, unveiled at I/O 2026. The hire is the latest sign that Anthropic is emerging as a magnet for elite technical talent. The company separately added cybersecurity veteran Chris Rohlf to its frontier red team, which stress-tests advanced models against severe threats.
From OpenAI to Tesla and back
While at OpenAI, Karpathy focused on deep learning and computer vision before departing in 2017 to join Tesla, where he led the Full Self-Driving and Autopilot programs until 2022. He then returned to OpenAI for about a year before leaving again in 2024 to start Eureka Labs, a startup applying AI assistants to education. He is also widely known for his free "Neural Networks: Zero to Hero" course and lectures on building models from scratch. Karpathy said he "remains deeply passionate about education" and plans to resume that work in time.
The intensifying AI talent race
The move underscores how fiercely leading labs are competing for top researchers as capital floods the sector, following milestones such as OpenAI's record AI funding round. As frontier developers race to ship ever more capable systems, from reasoning models to consumer features like Gemini's new third-party editing integrations, the people who can design and run the largest training runs have become some of the industry's most sought-after assets.
Reporting based on coverage from TechCrunch and Andrej Karpathy's public statements.