AI Leaders Abandon Waiting List, Build Own Silicon Amid Chip Shortage

Artificial intelligence is driving an unprecedented demand for computing power, outpacing supply and forcing cloud providers and AI developers into a high-stakes game of chip procurement. To mitigate this risk, some of the world's largest AI companies are taking matters into their own hands by designing custom silicon.

Anthropic has become the latest company to join this growing club, confirming it is building an in-house chip design team to create custom silicon for its Claude family of AI models. This move validates a Reuters report from April that revealed Anthropic was exploring the idea.

The shift towards custom silicon reflects the constrained nature of AI hardware, where Nvidia's Blackwell GPUs are difficult to obtain and shipments of its next-generation Vera Rubin platform are already ramping up despite carrying price tags of roughly $55,000 per GPU. A fully populated Vera Rubin NVL72 rack containing 72 GPUs is estimated to cost between $7.8 million and $9.1 million before networking and supporting infrastructure are included.

With a single server rack now costing $9 million, the world's AI leaders are ditching the waiting list to build their own silicon. Anthropic emphasized that this is part of a multi-chip strategy, not a replacement for partners such as Nvidia, AMD, Amazon's AWS, and Google Cloud.