US Regulatory Landscape
US Robot Regulation: The Fragmented Landscape for Physical AI
No comprehensive federal framework. Eight agencies with overlapping mandates. Zero binding rules for civilian humanoids, home robots, or indoor autonomous mobile robots.
Last updated: July 2026
The state of US robot regulation in one sentence
The US has more commercially deployed humanoid robots than any other country outside China — and no specific federal rules governing most of them.
What Agility's OSHA milestone actually means
In November 2025, Agility Robotics announced that its Digit humanoid had passed a safety field inspection recognized by OSHA — specifically a Nationally Recognized Testing Laboratory (NRTL) inspection. This was widely reported as a milestone.
What it actually means: Digit was voluntarily assessed by a third-party safety testing organization that OSHA recognizes. This is not a legal requirement. OSHA has no humanoid-specific standard. The NRTL certification signals that Digit meets existing general industrial safety standards — which is meaningful for enterprise customers making procurement decisions, but not a legal compliance gate.
The milestone matters because enterprise procurement — especially at large logistics companies like GXO — requires safety justification for insurance and liability purposes. NRTL certification provides that justification. It is market-driven compliance, not regulatory compliance.
Agency-by-agency breakdown
Agility Digit passed OSHA-recognized NRTL safety field inspection Nov 2025 — first humanoid. Voluntary, not mandatory.
Current FDA policy requires human control at all times for AI-assisted surgical systems. IEC 80601-2-77 recognition in May 2025 is the first standardized framework.
Most mature autonomous vehicle framework in the US. Applies only to airborne systems. Frequently cited as the model for future land-based autonomous vehicle regulation.
Wait-and-see mode for robotics. Announced AI pivot for 2026-2027 regulatory agenda. No binding robot-specific rules as of mid-2026.
Potential jurisdiction over deceptive AI claims in robot marketing. No enforcement actions to date targeting physical AI hardware.
Active testing programs for military robot platforms. AI ethics principles require human accountability for lethal force decisions.
Relevant if outdoor robots operate on public roads or sidewalks. No framework for non-vehicle robots in public spaces.
AI RMF is voluntary guidance, not regulation. Used by industry for self-assessment. Often referenced in contracts and procurement.
What has zero binding rules
These are areas where no US federal rule currently applies — despite active commercial deployment.
Legislative outlook
Bars US federal agencies from acquiring AI humanoid robots made by companies in China, Iran, North Korea or Russia. Federal procurement only — does not regulate private sector. Passed committee; status in full Senate uncertain.
Would establish an advisory commission to study robotics and make recommendations. Advisory only — creates no binding rules. Indicates Congress is in information-gathering phase, not regulation-drafting phase.
China has explicit national R&D goals and adoption quotas for robots in manufacturing. The EU has a comprehensive regulatory framework. The US has neither a national strategy nor a regulatory framework as of mid-2026. This is a deliberate policy choice — not an oversight.