July 13, 2026

Anthropic’s Krishna Rao details compute allocation strategy for AI research

Anthropic has announced that its CFO, Krishna Rao, highlighted the company’s focus on compute allocation as crucial for maintaining its competitive edge in AI development during an interview. The company directs its compute resources primarily towards its research team, which is tasked with designing advanced AI models, while also ensuring the flexibility of its multi-platform data centers that support various chip architectures for both training and inference. This innovative approach allows Anthropic to utilize the same infrastructure for different stages of their AI processes, illustrating their commitment to improving AI capabilities, which is further emphasized by a personal update from a team member who is transitioning to join Anthropic’s compute team to tackle the challenges of recursive self-improvement in powerful AI.

T. Blom: Tom Blomfield, known online as @t_blom, is a technology executive and cofounder of fintech companies Monzo and GoCardless who is taking a leave from Y Combinator to join Anthropic. He will work on the compute team alongside other team members, focusing on challenges related to AI infrastructure and recursive self-improvement.
Anthropic: Anthropic is an AI research and safety company focused on developing advanced language models such as Claude. Its strategy emphasizes flexible compute allocation across training, customer serving, and research to maintain leadership in frontier AI capabilities. The company operates GPU-agnostic data centers that enable training on diverse hardware platforms including Google TPUs, Amazon Trainium, and Nvidia chips, allowing the same infrastructure to support both model development and inference.
Krishna Rao: Krishna Rao serves as the Chief Financial Officer of Anthropic, overseeing procurement and allocation of compute resources central to the company’s operations. In a recent interview, he outlined the company’s approach to prioritizing compute for research teams to accelerate development of new models while balancing other uses.

`json
{
“Compute Flexibility”: “Anthropic maintains multi-platform data centers that support training and inference interchangeably across different chip architectures.”
}
`

AI tools built by Emerald Force

Built and supported by Emerald Force.

You might also like