Anthropic forms team to design its own AI chips for Claude models

AI company Anthropic has confirmed that it has established a new internal hardware team to design custom AI chips for its Claude models. The company's goal is to meet the high-scale processing power required by advanced AI systems more quickly and efficiently.

In Anthropic's new structure, hardware architecture and software components are planned to be developed in tandem. In this context, the company announced that it has recruited engineers experienced in both hardware and software fields.

The move comes at a time when global demand for AI processors is rising rapidly and supply shortages have become a significant pressure factor in the sector. Since training and running large language models require high-capacity processing infrastructure, technology companies are seeking greater control on the hardware side.

MULTI-CHIP STRATEGY TO CONTINUE

Anthropic establishing its own chip design team does not mean that its current hardware partnerships will end. The company views this step as part of a "multi-chip strategy." In line with this, it is stated that a diversified infrastructure including technologies from companies such as Amazon Web Services (AWS), Google, Nvidia, and AMD will continue to be used.

The company did not share a timeline regarding when the chips to be developed will be made available. Furthermore, no statement was made regarding whether the production and foundry processes will be carried out by Anthropic or if they will proceed with external partners.

According to industry sources, the cost of designing an advanced AI processor from scratch can reach up to 500 million dollars. The employment of qualified hardware engineers and the prevention of potential errors in the production process are cited as the primary reasons for this cost.

News Source: 12punto