Anthropic confirms in-house chip design team to accelerate Claude AI workloads

Anthropic has publicly confirmed that it is assembling an in-house chip design team tasked with developing custom AI hardware tailored to its Claude large language model. The announcement validates earlier reports from April 2026 that the AI company was exploring the possibility of designing its own silicon, marking a significant strategic shift for the firm as it seeks greater control over its AI infrastructure.

The confirmation, first reported by Business Insider, follows a wave of industry speculation and comes as Anthropic scales its operations to meet surging demand for generative AI services. A company spokesperson said the new team would focus on co-designing custom hardware that enables Claude to run faster and more efficiently, addressing the growing computational constraints faced by large-scale AI models.

Anthropic has already posted job listings for silicon engineer and hardware system architect roles, signaling a serious commitment to building out the team. Currently, the company relies on a broad array of chips from multiple vendors, including Nvidia and AMD GPUs, Google TPUs, and Amazon Trainium hardware. In response to a request for comment from DCD, a spokesperson emphasized that the move is "the latest step in our multi-chip approach" and that Anthropic will continue deploying a "diversified hardware stack" across its operations.

The new initiative gains further momentum from a key executive hire. In June 2026, DCD reported that Clive Chan, who previously led OpenAI's custom chip program, had joined Anthropic. At the time, Chan wrote on X: "I joined Anthropic this week because I was deeply impressed with the team's talent, values, and ambition, and I'm already energized by the pace and intensity of the past few days. It's time to build." His arrival underscores the competitive talent race in the AI semiconductor space.

Anthropic's move comes at a pivotal moment in the industry. The same month Chan announced his departure, OpenAI and Broadcom debuted "Jalapeño," their first co-designed Intelligence Processor for LLM inference. The chip reportedly went from initial design to manufacturing tape-out in just nine months, which the companies claimed was likely the fastest ASIC development cycle ever achieved in high-performance advanced semiconductors, with first shipments expected by the end of 2026.

Industry analysts view Anthropic's foray into custom silicon as part of a broader trend among leading AI labs to reduce dependence on merchant chip suppliers and optimize hardware for their specific model architectures. By investing in in-house chip design, Anthropic aims to improve performance, lower inference costs, and maintain a competitive edge in the rapidly evolving AI landscape. The company has not disclosed further details on the chip team's structure or timeline, but the public confirmation signals a clear strategic direction as it scales its AI infrastructure.