NVIDIA BioNeMo

Type: Cloud platform (AI compute for life sciences) — subsidiary of NVIDIA (NVDA)

BioNeMo is NVIDIA's cloud platform for biological foundation models — it hosts, fine-tunes, and deploys specialized AI models for drug discovery, protein design, and molecular simulation. It is not a drug discovery company but rather the compute backbone on which the AI-bio sector runs.

Strategic position: NVIDIA occupies the same position in AI-bio that it occupies in general AI — it sells the infrastructure layer (GPUs + cloud APIs) to all players regardless of who wins the platform/asset race. Every company building protein models (ESM3, RFdiffusion, Chai), doing generative chemistry, or running lab-in-the-loop systems needs GPU compute, and BioNeMo provides managed access to pre-trained biology-specific models on that compute.

Key relationships: Recursion, Lila Sciences, and EvolutionaryScale all have explicit NVIDIA partnerships. NVIDIA's NVentures invested in both EvolutionaryScale and Lila Sciences.

Key announcement (2026): NVIDIA Newsroom confirmed BioNeMo platform adoption by multiple life sciences leaders, positioning it as the sector's shared inference infrastructure. (NVIDIA Newsroom)

Investment thesis

NVDA is the cleanest "picks and shovels" exposure to AI-bio: benefits from the sector's growth regardless of which specific drug or platform wins; moat is the CUDA ecosystem which is harder to displace in biology than in general AI due to specialized workflows.

Cross-topic link

See also kb/hardware/wiki/entities/nvidia.md for the primary NVIDIA dossier.