Anthropic's Model Hardware Standard

Anthropic's Model Hardware Standard (MHS) is in research preview as of August 27, 2026, giving scientific research labs and advanced manufacturers a standardized way for AI agents to control physical equipment. MHS provides shared drivers with read/write primitives and reference files describing device capabilities, physical properties, and enforced safety limits; agents access these via MCP, command-line tools, or code. The framework is model-agnostic and works with any device that has a programmable interface. Hardware vendors Tecan, Universal Robots, and Doosan Robotics are adding native support; launch partners include AWS, Hugging Face LeRobot, QIAGEN, and Raspberry Pi. Access is through a waitlist at modelhardwarestandard.com. Early partner results include automated protein assays, dose-response experiments three times faster, and laser stabilization at QuEra Computing recovering lock 99.3% of the time autonomously. Both Anthropic and early projects report the standard cuts hardware integration time from weeks or months to hours or minutes. Anthropic plans to open-source MHS after the research preview, using the interim to develop safety evaluations and a physical safety roadmap.