Topline
Andrej Karpathy, a founding member of OpenAI and former AI chief at Tesla, has joined Anthropic, he announced Tuesday—a high-profile hire of one of the most recognizable names in AI research as the lab races against tech companies including OpenAI, Google and Meta for elite talent.
Key Facts
Karpathy started this week at Anthropic, where he is working on pre-training—the foundational stage of developing large language models—under team lead Nick Joseph, the company said.
He will help launch a new team focused on using Claude to develop its models, not relying solely on expensive raw compute and hardware alone, the company added.
Karpathy posted on X Tuesday that he believes “the next few years at the frontier of LLMs will be especially formative” and is excited to get back into research and development—after running AI education startup Eureka Labs for the past two years.
Karpathy famously coined the term “vibe coding” to describe using AI to write code and is a widely regarded voice on AI developments and the industry through his social media posts online.
Chris Rohlf, a 20-year cybersecurity veteran most recently at Meta, also announced Tuesday he joined Anthropic’s frontier red team, which stress-tests models against severe threats.
Key Background
Karpathy is among the most influential figures in modern AI. He focused on deep learning and computer vision at OpenAI before leaving in 2017 for Tesla, where he led the Full Self-Driving and Autopilot programs until 2022. He returned to OpenAI for a year before departing again in 2024 to launch Eureka Labs, a startup applying AI to education. He said education still matters to him and that he plans to return to that work when the time is right. Anthropic, founded in 2021 by former OpenAI executives Dario and Daniela Amodei, has positioned itself as the safety-focused frontier lab and has been on an aggressive hiring run as it scales Claude against ChatGPT. In April, Anthropic hired Microsoft’s Eric Boyd—most recently president of AI platforms—as its head of infrastructure, a role tied to the company’s $50 billion U.S. data center buildout. The traffic hasn’t been one-way: OpenAI poached Anthropic safety researcher Dylan Scandinaro in February as its new head of preparedness, a role reportedly paying up to $555,000.
Big Number
$900 billion. That’s Anthropic’s latest reported valuation as of last week thanks to a $30 billion funding round that the company has agreed to, according to sources familiar with the matter. The Claude maker’s valuation has more than doubled in just a few months from its $380 billion valuation from February’s funding round and has for the first time surpassed OpenAI’s valuation, which reached $852 billion in March.
Tangent
The Karpathy news landed the same morning Anthropic was in federal appeals court fighting the Pentagon. In March, the Defense Department designated Anthropic a “supply-chain risk”—a label historically reserved for foreign adversaries—after the company refused to drop its red lines against mass surveillance and fully autonomous weapons. The Pentagon argues Anthropic is uniquely untrustworthy because it won’t agree to the military’s “all lawful use” standard for AI; Anthropic argues the designation is unconstitutional retaliation. A three-judge panel in Washington heard arguments Tuesday and could rule within weeks. In the meantime, Anthropic was shut out of a recent Pentagon deal in which Google, OpenAI, Microsoft, and Amazon Web Services agreed to bring their AI onto the Defense Department’s classified networks.
