Anthropic hires Karpathy, SpaceX compute ties reported

- Andrej Karpathy said on May 19 he joined Anthropic, as reports this week linked the Claude maker to fresh pre-training work and new compute supply. - Anthropic said on May 6 it signed for all capacity at SpaceX's Colossus 1 site, more than 300 megawatts and 220,000 Nvidia GPUs. - Anthropic's May 6 post and May 19 hiring reports remain the clearest public markers; investor updates are the next place for numbers.

Andrej Karpathy said on May 19 that he had joined Anthropic, giving the Claude maker a high-profile research hire as it expands model training and infrastructure. Anthropic had already disclosed a new compute partnership with SpaceX on May 6, saying the deal would give it all of the capacity at SpaceX's Colossus 1 data center in the near term. Separate media reports this week said Anthropic had told investors it expected its first profitable quarter, adding a financial datapoint to a run of research, product and infrastructure announcements. A May 23 episode of *The AI Daily Brief* grouped those developments together as evidence of a broader push by Anthropic across talent, economics and compute. ### Who confirmed Karpathy's move, and what is he expected to do? Karpathy said on X on May 19, "I've joined Anthropic," according to reports that cited his post. CNBC and TechCrunch reported that he joined Anthropic's pre-training team, after earlier roles as an OpenAI co-founder and Tesla's former AI lead. Nicholas Joseph, Anthropic's head of pretraining, said in a post cited by VentureBeat that Karpathy would build a team focused on using Claude to accelerate pretraining research itself. (techcrunch.com) That description is narrower than a general research role: it ties Karpathy directly to the stage of model development where labs run the large, expensive training jobs that shape a model's core capabilities. (cnbc.com) ### What did Anthropic actually say about SpaceX? Anthropic said on May 6 that it had "agreed to a partnership with SpaceX" that would "substantially increase" its compute capacity. In the same announcement, the company said it had signed an agreement to use all of the compute capacity at SpaceX's Colossus 1 data center. Anthropic said that meant access to more than 300 megawatts of new capacity, or more than 220,000 Nvidia GPUs, within a month. (venturebeat.com) CNBC reported that the site is in Memphis, Tennessee, and said the deal pairs Anthropic with a company owned by Elon Musk, who also controls rival AI company xAI. TechCrunch later reported, citing SpaceX's S-1 filing, that Anthropic would pay $1.25 billion per month through May 2029, with either side able to terminate on 90 days' notice. ### Where did the profitability claim come from? (anthropic.com) TechCrunch reported on May 20 that Anthropic had told investors it was about to have its first profitable quarter. CNBC, citing a person familiar with the matter, reported that Anthropic generated $4.8 billion in the first quarter and was projecting about $10.9 billion in second-quarter revenue, which would put it on track for its first profitable quarter if achieved. (cnbc.com) Those figures have not been published by Anthropic in a public earnings release, and the company is privately held. The public basis for the claim is therefore media reporting on investor materials and people familiar with the matter, not a company filing. ### Why did these three items get bundled together this week? (techcrunch.com) The AI Daily Brief said on May 23 that Anthropic's week stood out because it combined a senior research hire, improving economics and a new compute relationship. A social post from The AI Executive, linking to that discussion, summarized the same cluster of themes around Anthropic's talent, profitability signals and SpaceX ties. (techcrunch.com) Anthropic's own public record supports two of those three points directly: the SpaceX compute deal and the company's continuing emphasis on research, including recent work posted on its research pages. The profitability point remains based on outside reporting rather than a formal Anthropic disclosure. ### What should readers watch next? May 19 and May 6 are the key dates in the public timeline so far: Karpathy's move was reported on May 19, and Anthropic published the SpaceX announcement on May 6. (anthropic.com) The next concrete updates are likely to come through Anthropic announcements, additional reporting on the SpaceX contract terms, or investor-related disclosures that confirm whether the company reached the profitable quarter described in reports this week. (techcrunch.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.