TL;DR

Anthropic has hired Eric Boyd, a 16-year Microsoft veteran who most recently ran the company’s AI platform and oversaw around 1,500 staff, as its new head of infrastructure. The appointment follows repeated capacity strain on Claude services and a previously announced $50 billion commitment to build AI data centres in the US.

Compute as the bottleneck

Boyd reported to Microsoft executive vice president Jay Parikh and was responsible for deployment of large language models to Microsoft customers and internal teams. Moving to Anthropic places him against the company’s most stubborn constraint: delivering enough cloud capacity to meet what it has described as “unprecedented demand”, particularly from Claude Code and enterprise API customers.

Anthropic’s $50 billion US data centre plan is substantial but sits alongside rival OpenAI’s roughly $600 billion infrastructure target through 2030. The scale gap matters because inference capacity — not model quality alone — increasingly determines which frontier labs can keep customer commitments when usage spikes.

Context for UK enterprise buyers

The hire lands in the same week as two other Anthropic stories: the unveiling of Project Glasswing and the Mythos cyber model, and the company’s ongoing Google and Broadcom compute partnership. Read together, they describe a lab trying to do three things at once — ship more capable models, pre-empt misuse, and scale the infrastructure underneath. Boyd’s hire is the least glamorous of the three but arguably the most load-bearing.

UK enterprise customers running production workloads on Claude have direct exposure to this: recent outages and rate-limit tightening have made Anthropic’s compute position a live procurement question, not an investor one. A credible infrastructure lead reduces the discount vendors like Anthropic need to offer to win risk-averse buyers.

Looking forward

Boyd’s first visible test will be whether the Google/Broadcom partnership translates into available capacity before Claude’s next model release, or whether Anthropic hits another round of rationing. For customers evaluating frontier model vendors, infrastructure governance is quietly becoming as important as benchmark scores — and this appointment suggests Anthropic agrees.