Anthropic Forms Chip Development Team for Claude AI

AI company Anthropic has formed a dedicated team to develop its own AI chips for operating its model, Claude. The company confirmed it plans to develop hardware and AI models in tandem to allow its technology to run faster and more efficiently. Anthropic decided to develop its own chips amid rising demand for Claude, while AI companies race to secure computing infrastructure capacity to meet development and service needs. Anthropic was previously reported to be exploring a partnership with Samsung for chip development. However, the company believes that relying solely on other firms is insufficient to meet the ever-increasing computational demands. Anthropic is not the first AI company to develop its own chips. In June, OpenAI introduced its own AI chip called Jalapeño, developed with Broadcom and designed specifically for inference workloads. Google DeepMind has long relied on Alphabet’s Tensor Processing Units (TPUs) to run its AI models, while Meta is also developing its own AI accelerators through its Meta Training and Inference Accelerator (MTIA) line. To support the chip development, Anthropic is currently recruiting engineers with experience in semiconductor design.