Anthropic has hired Amir Salek, the engineer credited with founding Google's custom chip programme, according to a report from The Times of India. The move represents one of the most significant hardware-focused appointments the Claude maker has made to date, and it arrives as the broader AI industry races to reduce dependence on third-party silicon suppliers.
Who Is Amir Salek?
Salek built his reputation at Google, where he is widely credited with establishing the foundations of the company's Tensor Processing Unit effort, the custom chip line that became central to Google's AI infrastructure. His work helped Google reduce its reliance on external chip suppliers and gave the company a meaningful performance and cost advantage in running large-scale machine learning workloads. Bringing that expertise to Anthropic is a clear statement of intent about where the company sees its hardware ambitions heading.
Key Facts
- Amir Salek founded Google's custom chip programme, including early Tensor Processing Unit development.
- Anthropic has confirmed the hire, though the specific role has not been publicly detailed.
- The appointment follows a broader pattern of AI labs investing in proprietary silicon.
- Anthropic has previously signaled interest in custom hardware as a long-term infrastructure priority.
The timing is notable. Anthropic confirmed earlier this year that it is developing a custom AI chip, a project that would allow it to train and run its Claude models on hardware optimised for its own workloads rather than relying entirely on Nvidia GPUs or cloud-provider chips. Salek's arrival fits squarely within that roadmap. Custom silicon projects at this scale typically require years of development, significant capital, and leadership with direct experience shipping chips at the hyperscaler level.
Building custom chips is not a shortcut. It requires deep institutional knowledge, long timelines, and a willingness to absorb early losses for long-term infrastructure control.Industry analysis on AI chip strategy
A Broader Hardware Push
Anthropic's hardware ambitions do not exist in isolation. AMD agreed to invest up to $5 billion in Anthropic and struck a chip supply deal as part of an arrangement that gives Anthropic both capital and access to alternative silicon at scale. That deal, combined with the Salek hire, points to a company that is pursuing multiple hardware strategies simultaneously rather than betting everything on a single supplier or a single internal programme.
The competitive pressure here is real. OpenAI, Google DeepMind, and Meta have all invested heavily in custom or semi-custom chip efforts. For Anthropic, which has positioned itself as a safety-focused lab running some of the most computationally intensive frontier models, controlling more of the hardware stack could have meaningful implications for both cost structure and the ability to fine-tune performance for specific workloads. Anthropic's broader hardware push has been building for several months, with Salek's appointment representing its highest-profile chip-related hire yet.
Details about Salek's exact title and scope of responsibilities at Anthropic have not been publicly disclosed. What is clear is that the company is assembling chip expertise at a senior level, suggesting its internal hardware programme has moved well past the exploratory stage. For followers of the latest Claude AI news, this hire is worth watching closely. The trajectory from software-first AI lab to company with serious silicon ambitions is one that tends to reshape competitive dynamics over a multi-year horizon, and Anthropic appears to be moving deliberately in that direction.
