Anthropic is building an internal team to design custom chips for Claude as the company seeks more computing capacity to support growing demand for its AI models.

Business Insider first reported the plans Wednesday, citing a company statement and a new job listing. Anthropic later confirmed that it is hiring engineers with experience across hardware and software to jointly design chips and models that can run Claude faster and more efficiently.

The company described the custom silicon effort as the latest part of its broader chip strategy. Anthropic said it will continue using computing hardware from Amazon Web Services, Google, Nvidia, and AMD as it develops its own technology.

Anthropic has not disclosed when its first chip could be completed or whether it plans to manufacture the processors itself. Designing an advanced AI chip can cost roughly $500 million because of engineering, testing, and manufacturing requirements, according to industry estimates cited by Reuters.

The company is recruiting engineers for what a job listing described as a custom silicon team. Anthropic is seeking candidates with experience in chip architecture, design, verification, and the completion of semiconductor products.

The position offers a salary ranging from $320,000 to $485,000 and requires candidates to demonstrate direct experience contributing to semiconductor designs that reached production.

Anthropic had previously explored working with Samsung on the planned processor. The Information reported in July that the companies had discussed a potential manufacturing partnership, although details including the chip’s purpose, performance, and role within Anthropic’s servers had not been finalized.

Anthropic declined to provide additional details about a possible Samsung partnership at the time. The company said chips supplied by Google, Amazon, and Nvidia would remain central to its computing strategy.

The chip initiative comes as Anthropic expands its infrastructure agreements to secure the computing power needed to train and operate Claude. Developing custom silicon could allow the company to optimize hardware around its own models while maintaining access to processors supplied by its existing partners.

Anthropic joins a growing group of AI companies developing processors designed for their own workloads. The effort gives the company greater control over how its models use computing resources while adding another source of hardware to its existing infrastructure network.