Google DeepMind AlphaFold 3 Commercial API Reaches 50,000 Research Organisations in 90 Days

## AlphaFold 3 Adoption Sets Records Google DeepMind reported that its AlphaFold 3 commercial API has surpassed 50,000 research organisation sign-ups since its commercial launch 90 days ago, making it the fastest-adopted life sciences software platform by new account creation rate. The service provides high-accuracy predictions of protein structure, protein-ligand binding, and protein-nucleic acid interaction for a fee structure that starts at $0.002 per prediction job — a price point that puts structural biology compute within reach of university labs that previously relied on oversubscribed HPC clusters. ## Impact on Drug Discovery Timelines Early data from pharmaceutical partners suggest that AlphaFold 3-guided hit identification has compressed early-stage drug discovery cycles by an average of 8 months across programmes where the target protein had no resolved crystal structure. Several biotech firms report advancing candidates to preclinical testing based primarily on AF3-predicted binding poses, skipping lengthy crystallography campaigns. DeepMind published a preprint this week documenting 12 cases where AF3 predictions correctly anticipated binding conformations later confirmed by cryo-electron microscopy — an independent validation that strengthens confidence in the model's utility as a primary discovery tool rather than a secondary check. ## Expansion Plans DeepMind confirmed plans to add RNA tertiary structure prediction and multi-protein assembly modelling to the commercial API in Q3 2026. An academic free tier with 20 predictions per month per researcher will launch simultaneously, lowering the barrier for resource-constrained labs.