# Agents Get a Hardware Control Standard for Physical Instruments

The interface between AI agents and physical laboratory, factory and scientific equipment is being standardized the way tool-calling and MCP standardized software access, so expect recurring protocol proposals, instrument-vendor adoption, and control of the agent-to-hardware layer becoming a competitive position — with safety and liability questions arriving after deployment rather than before.

- Conviction: 67 / 100 (weakening)
- 7-day move: +19
- Horizon: Medium term (3-9 months)
- Tracking since: 2026-09-06T00:00:00.000Z
- Last updated: 2026-09-14T14:04:09.672Z
- Canonical: https://polylog.news/ai/trends/agent-instrument-control-standards
- Publisher: Polylog
- Affected regions: Global

## Recent score history

- 2026-09-13: 68
- 2026-09-14: 67

## Recent evidence

- [confirms] Anthropic's Hardware Standard Lets Agents Drive Lab Instruments, With Early Results From Genentech and QuEra (2026-09-10): Anthropic published a hardware standard letting agents drive laboratory instruments, with testers at Genentech and QuEra reporting laser stabilization on a quantum computer improving from 58 percent to 99.3 percent and an imaging experiment compressed from weeks to a day. A frontier lab shipping the protocol with named pharmaceutical and quantum-hardware design partners and measured instrument outcomes moves the agent-to-hardware layer from proposal to a contested standard with incumbent backing.
- [confirms] AI Agents Are Taking Direct Control of Laboratory Instruments (2026-09-09): OpenAI describes Codex calibrating qubits at MIT without constant supervision, weeks after Anthropic opened a specification for agents to operate physical hardware. A working deployment on real instruments following a published control specification is the protocol-then-adoption sequence the thesis tracks, and qubit calibration is a high-consequence first target with no established liability framework.

3 more evidence entries, the full score history, the conviction-driver timeline, and affected assets are for subscribers: https://polylog.news/pricing
