OpenAI Co-Founder Andrej Karpathy Joins Anthropic to Use Claude on Claude

Anthropic has hired Andrej Karpathy, one of the original co-founders of OpenAI and the former director of AI at Tesla, in what is being read as the most consequential AI talent move of 2026 so far. Karpathy confirmed the news himself on X on Tuesday, writing that "the next few years at the frontier of LLMs will be especially formative" and that he was "very excited to join the team here and get back to R&D." He starts inside Anthropic's pre-training organization, reporting to team lead Nick Joseph, according to techcrunch.com.

Crucially, Karpathy is not just slotting into an existing role. He is building a new group whose explicit mandate is to use Claude to accelerate Claude's own pre-training research — the large, expensive runs that bake in a model's foundational knowledge. That is a bet on AI-assisted research as the next compounding advantage, and a notable departure from the prevailing "just buy more chips" doctrine. Anthropic is signalling that smarter experimentation loops, not raw FLOPs, are where it intends to outpace OpenAI and Google.

Karpathy's resume is unusually load-bearing for the field. He was a founding member of openai.com in 2015, ran Tesla's Full Self-Driving and Autopilot programs from 2017 to 2022, briefly returned to OpenAI in 2023, and then left in 2024 to start Eureka Labs, an AI-native education company. He is also one of the most-followed AI researchers on the open web, with viral lectures, the "Neural Networks: Zero to Hero" course on his YouTube channel, and the now-canonical "vibe coding" post that helped name a generation of AI-assisted developer workflows. Karpathy said he remains "deeply passionate about education" and plans to return to that work in time.

The hire lands in the middle of an unusually busy month for Anthropic. The company is closing a roughly $30 billion fundraise at a valuation north of $900 billion, just acquired SDK toolmaker Stainless for around $300 million, expanded Claude Cowork rollouts at PwC and KPMG, and committed $200 million with the Gates Foundation to deploy Claude in global health. Pulling a co-founder out of a rival lab to lead recursive self-improvement research on Claude is the through-line: Anthropic is buying not just compute and customers, but the people most capable of changing how frontier models are built in the first place.