The battle for AI supremacy has moved from chatbots to wet labs, and Anthropic is spending heavily to challenge Google’s dominance in drug discovery.

Anthropic is no longer just building conversational models. By acquiring Coefficient Bio for $400 million and hiring Nobel laureate John Jumper—the mind behind AlphaFold—the company is positioning itself as a direct threat to Google DeepMind’s Isomorphic Labs.

This is a massive shift from theoretical AI safety to commercial drug development.

The Talent Shift

Poaching Jumper and key DeepMind researchers signals a talent drain from Google’s established bio-AI divisions. Anthropic is pairing this talent with its new Claude Science model, which is already helping clients like Novo Nordisk speed up regulatory timelines.

But the real test lies in the physical world. Anthropic is building its own wet labs to develop preclinical drugs for neglected diseases.

This move signals that digital-only AI models are no longer enough to win the biopharma race. Companies must control the physical validation of their predictions.

The Wet Lab Risk

Software is cheap to scale, but biology is notoriously messy. Transitioning from digital prediction to physical lab validation introduces massive capital risks and high failure rates.

If Anthropic succeeds, it proves that general AI companies can rapidly pivot into specialized deep tech. If it fails, it will be a very expensive lesson in the limits of pure computation.

The biopharma industry should watch this closely. The line between AI lab and pharmaceutical giant is officially blurring.