Anthropic previews Model Hardware Standard for AI control of lab instruments
Anthropic opened a research preview of the Model Hardware Standard, a shared specification and driver set that lets AI agents operate lab and manufacturing instruments across vendors, with Carnegie Mellon integrating an instrument in eight hours instead of weeks and QuEra lifting quantum laser lock recovery from 58% to 99.3%. Genentech, HHMI Janelia and Universal Robots are among the preview partners, and the standard is model-agnostic — reachable through MCP, a CLI, or code files — so the integration tax that keeps instruments siloed becomes a solved layer rather than per-lab custom work once it goes open source.
Source: anthropic.com ↗
It typically takes a lab or manufacturing facility weeks, if not months, to set up and integrate their hardware. MHS reduces this integration work to hours or minutes.
Why this matters
- → Lab integration drops from weeks to hours, unblocking experiments at scale.
- → AI agents can now operate physical instruments autonomously with safety guardrails.
- → Shared standard dissolves vendor lock-in, turning custom work into reusable drivers.