Superpower Daily
OpenAI Loses Data Center Chief While Pursuing a Reported $600 B...Alibaba Raises HK$80B for Full-Stack AI as Investors Question t...LLMScholarBench Tests 22 Models and Finds an Accuracy-Represent...Meta Adds Voice-Controlled AI and Navigator Home Screen to Ques...Harvard Uses 7 AI Faculty Avatars in a $699 Founder Bootcamp, W...DeepSeek V4 Pro Lands on Fireworks With $2.50 CyberGym Solves,...OpenAI Loses Data Center Chief While Pursuing a Reported $600 B...Alibaba Raises HK$80B for Full-Stack AI as Investors Question t...LLMScholarBench Tests 22 Models and Finds an Accuracy-Represent...Meta Adds Voice-Controlled AI and Navigator Home Screen to Ques...Harvard Uses 7 AI Faculty Avatars in a $699 Founder Bootcamp, W...DeepSeek V4 Pro Lands on Fireworks With $2.50 CyberGym Solves,...
Businesspublished

Anthropic Hires Builder of Google’s First Seven TPUs for In-House Silicon Push

The Claude maker is adding deep chip-building experience while retaining a compute mix spanning Nvidia, Google and Amazon hardware.

By 2 min read
Anthropic Hires Builder of Google’s First Seven TPUs for In-House Silicon Push
Sign in to save and react

Listen to this story

The audio brief

About 1:38
0:001:38
Read transcript
Anthropic has hired Amir Salek, the engineer who built Google’s custom-chip program and delivered its first seven TPU generations, to help create an in-house silicon operation. It’s a significant addition for the maker of Claude, but not a move away from outside hardware. Salek left Google in 2022. Before that, he founded and led Nvidia’s system-on-a-chip design organization, giving him experience across two of the most important companies in AI hardware. Anthropic had already started recruiting custom-silicon specialists in August, including engineers who know how to take semiconductor designs through production. Salek now joins that effort. The key point is that Anthropic’s compute strategy remains deliberately mixed. Claude runs on Nvidia graphics processors, Google TPUs, and Amazon Trainium chips. Amazon is Anthropic’s primary cloud and training partner, while the company says using several platforms lets it place workloads on different hardware and reduce dependence on any one supplier. Anthropic has already worked at the software and hardware boundary: its engineers wrote low-level kernels for Trainium, contributed to Amazon Web Services’ Neuron software stack, and worked with Amazon’s Annapurna Labs on future chip generations. The scale of the external commitments is substantial. Anthropic has said it could use as many as one million Google TPUs, with well over one gigawatt of capacity expected in 2026. A Google-Broadcom agreement covers multiple gigawatts of next-generation TPU capacity from 2027. The question now is how much Anthropic’s internal silicon effort will add to that already large, supplier-heavy compute mix.

Story brief

3 key points

Anthropic is building internal chip-design capability without abandoning outside accelerators. Amir Salek, who led Google’s custom-chip effort through seven TPU generations and previously founded Nvidia’s SoC organization, will help develop the operation. The strategy remains hybrid: Claude uses Nvidia GPUs, Google TPUs, and Amazon Trainium, with Amazon as its primary cloud and training partner. Anthropic has also...

  1. 01

    Salek left Google in 2022 after delivering the first seven TPU generations.

  2. 02

    Anthropic began recruiting custom-silicon engineers in August, including production and semiconductor-design specialists.

  3. 03

    The company has written Trainium kernels, contributed to AWS Neuron, and worked with Annapurna Labs.

Anthropic has hired Amir Salek, who founded Google’s custom-chip program and delivered its first seven TPU generations, for the compute team building an in-house silicon operation. The move adds a veteran of large-scale accelerator development as Anthropic continues to run Claude on hardware from multiple suppliers.

Salek’s background spans two of the companies central to AI hardware. Before Google, he founded and led Nvidia’s system-on-a-chip design organization. At Google, he established its custom-chip program and delivered the first seven generations of Tensor Processing Units, or TPUs, before leaving in 2022.

Anthropic had already confirmed in August that it was recruiting a custom-silicon team, including engineers experienced in taking semiconductor designs through production. Salek joins that effort to help build the company’s internal silicon operation.

Building alongside suppliers

The internal effort sits beside Anthropic’s existing hardware strategy, rather than replacing it. Anthropic currently trains and runs Claude across Nvidia GPUs, Google TPUs and Amazon Trainium chips. Amazon is its primary cloud and training partner, while Anthropic describes the mix as a way to put workloads on different hardware and reduce reliance on one platform.

Experience at the hardware layer

Anthropic has already worked closely with one specialized accelerator platform. Its engineers wrote low-level software kernels for Trainium, contributed to Amazon Web Services’ Neuron software stack, and worked with Amazon’s Annapurna Labs on future chip generations.

  • Salek brings experience founding a custom-chip organization and delivering multiple TPU generations.
  • Anthropic plans to develop internal silicon while continuing to reserve large amounts of Nvidia, Google and Amazon compute capacity.

The hire is an organizational addition to that mixed approach: Anthropic is assembling an in-house silicon operation while its disclosed commitments keep external accelerators central to its compute supply.

Related coverage

Continue reading