Anthropic Joins AI Chip Race, Building Custom Silicon to Power Claude's Soaring Growth

Context mode is active. Hover over any highlighted term to see its definition. Click a nested term to go deeper.
In a big move, AI powerhouse Anthropic is now hiring top engineers to create its own special AI chips, marking its official entry into the custom silicon race. This decision comes as the company struggles with 'compute constraints' due to the unbelievably fast growth of its AI model, Claude, which has seen its usage explode by as much as 80 times recently. By designing its own hardware, Anthropic hopes to make Claude run much faster and cheaper, ensuring it can keep up with the massive demand from users. This isn't just about making Claude better; it's a critical step in the fierce battle among top AI companies. Giants like Google with its TPUs, OpenAI with its Jalapeño chip, and Microsoft with Maia have already been building their own custom chips to reduce their reliance on general-purpose hardware from companies like Nvidia. Anthropic's 'multi-chip strategy' means they will still work with partners like Amazon Web Services, Google, Nvidia, and AMD, but having their own chips will give them more control and help lower costs significantly as AI models become even more complex and widely used. The company is looking for experienced engineers who know how to build and deliver actual semiconductor designs, showing they are serious about this venture. While Anthropic will design these chips in-house, reports suggest they are in talks with Samsung Electronics to handle the actual manufacturing. This strategic shift is crucial for Anthropic, which recently became the world's most valuable private AI company with a valuation near $1 trillion, as it aims to maintain its lead and support Claude's future growth amidst intense competition and a global shortage of AI-specific hardware.