Blomfield joined Anthropic's compute team — the group responsible for building and managing the infrastructure that powers Claude and Anthropic's AI models. He holds the title of member of technical staff on the compute team and works alongside Tom Brown (also known as @NotTomBrown), an Anthropic cofounder and the chief compute officer.
Blomfield served as a Group Partner / General Partner at Y Combinator since around 2023. In that role, he mentored hundreds of early-stage startups, logged more than 1,000 office hours, and oversaw a portfolio valued at $5 billion. His official YC profile lists him as a General Partner.
Blomfield co-founded two billion-dollar companies in the UK fintech space:
He was awarded an OBE in 2019 for services to the banking sector.
Founder-level operating experience for an infra problem. Blomfield scaled two fintech unicorns from zero to millions of users. Bringing that operational and scaling expertise into the compute team signals that Anthropic views infrastructure procurement, data-center logistics, and capacity planning as a founder-caliber strategic function, not just a back-end engineering task.
IPO readiness pressure. Anthropic confidentially filed a draft S-1 with the SEC on June 1, 2026, targeting a potential Nasdaq listing as early as October 2026 with Goldman Sachs, JPMorgan, and Morgan Stanley as lead underwriters. Its Series H round in May 2026 raised $65 billion at a valuation of approximately $965 billion, placing it on the cusp of a trillion-dollar IPO. A hire of Blomfield's stature shows Anthropic is staffing up its compute and operations muscle in anticipation of public-market scrutiny.
Part of a broader talent blitz. Blomfield joins a notable wave of high-profile hires at Anthropic in 2026, including Andrej Karpathy (pretraining team) and John Jumper, underscoring the company's aggressive build-out ahead of a potential public listing.
Blomfield's own framing — that the industry is entering "the early stages of recursive self-improvement" where compute availability is one of the hardest problems — crystallizes a broader shift. Frontier AI labs are no longer treating compute as a simple procurement line item. Hiring a two-time unicorn founder to work on infrastructure signals that: