AI infrastructure and drug OS converge
The wire is converging on a single pattern: AI infrastructure is getting more efficient while drug development is getting more organized. NVIDIA’s Rubin, Microsoft’s Agent Skills, and Rust’s expanding full-stack tooling all point to a deeper agentic stack; on the cure side, Cheiron’s $8 million seed and the FDA’s review of Revolution Medicines’ daraxonrasib show capital and regulation moving in parallel. The market is not waiting for a single breakthrough — it is building the operating layers that can compress iteration across machines and medicine. That matters because the bottleneck is shifting from ideas to throughput. The window is still open.
NVIDIA’s Rubin is positioned as the next GPU architecture, with up to 10x agentic throughput per unit energy versus Blackwell, while Microsoft released Agent Skills for Python and Tokio unveiled Topcoat for reactive Rust apps.
These moves shift advantage toward builders of agentic and automated systems, while Cheiron’s funding and Revolution Medicines’ FDA review show the cure stack increasingly has the software and regulatory rails to move faster.
Watch whether capital continues to fund infrastructure that multiplies output — in compute and in drug programs — because the next gains will come from systems that reduce the cost of iteration, not isolated wins.