Anthropic Builds Custom AI Chip Team for Claude Models

Anthropic confirmed Wednesday it is building an internal team to design custom chips specifically for its Claude AI models, marking the company's entry into a competitive race among AI leaders to control their hardware destiny.

The company is recruiting engineers with expertise spanning both hardware and software development to work on chips and AI models simultaneously. Job postings describe the initiative as a "custom silicon team," according to TechCrunch, though Anthropic provided no timeline for when the chips might enter production or whether it plans to manage manufacturing internally.

Why it matters

Custom chip development represents a strategic bet that proprietary silicon will deliver meaningful advantages in performance and cost efficiency compared to off-the-shelf accelerators. For Anthropic, which competes directly with OpenAI and Google in the foundation model market, controlling the hardware stack could prove essential to maintaining competitive inference speeds and economics as Claude scales to meet enterprise demand.

The economics of custom silicon

Developing advanced AI chips carries a substantial price tag. Industry sources cited by Reuters estimate costs can approach $500 million, driven by the need for specialized engineering talent and the expense of achieving reliable, defect-free chip fabrication at scale.

Despite these costs, Anthropic is following a path already taken by its competitors. OpenAI unveiled its Broadcom-built Jalapeño chip in June, designed specifically for inference workloads. Alphabet's tensor processing units power Google DeepMind's AI systems, while Meta has developed its own MTIA accelerators for AI workloads.

Existing hardware partnerships remain in place

Anthropic has secured access to AI computing resources through multiple channels, including arrangements with AWS, Google, Nvidia, and AMD. Last month, The Information reported the company was exploring Samsung as a potential manufacturing partner for its custom chips.

The chip initiative forms part of a larger infrastructure expansion. Anthropic sits at the center of a $15 billion financing deal for an AI data center campus in Hubbard, Texas, being developed by Nexus Data Centers. Google is backstopping Anthropic's financial obligations for that project.

For the Texas facility, Anthropic plans to deploy tensor processing units co-developed by Google and Broadcom, with chip costs covered through a vendor-financing arrangement with Broadcom. Google, which agreed in April to invest up to $40 billion in Anthropic, is expected to receive approximately 20% equity in the data center and power project in exchange for its backing.

These details were first reported by Quartz, citing information from TechCrunch, Reuters, and The Information.