Agility Robotics' Humanoid Robot Policy Recommendations

What Agility Robotics Proposed

Agility Robotics published six policy recommendations for United States humanoid-robot regulation on July 15, 2026. Peggy Johnson, Agility's chief executive, laid out the proposals in a post on Agility's own site titled "Industrial Humanoid Policy." Agility builds Digit, a warehouse-and-logistics humanoid already running paid deployments for commercial customers including GXO Logistics.

That commercial position gives Agility a direct stake in how Washington regulates the category it is trying to scale. The recommendations follow a July 2026 U.S. restriction on new foreign-made humanoid robot imports, covered in a separate explainer on the US humanoid robot import ban. Johnson's pitch is not to undo that restriction. Instead, it asks for a fuller industrial policy built around it, one that treats sourcing, safety, labor, and procurement as connected problems rather than four separate fights.

The Six Policy Recommendations

Johnson organized the proposal around six numbered pillars, each aimed at a different piece of how humanoid robots get built, sold, and put to work in the United States.

Read together, the six points add up to a single ask: treat humanoid robotics as one domestic industrial-policy question, not six separate problems spread across different agencies and different timelines.

• Close the gap on internationally sourced components. Actuators, precision gearboxes, and rare-earth magnets used in humanoid joints are still sourced largely from Asia. Agility wants federal incentives that make domestic or allied-country sourcing of those parts price-competitive with what a robot maker can buy today from overseas suppliers.
• Adopt a coordinated national robotics strategy. Agility wants one federal strategy, built around trade tools such as tariffs and anti-dumping measures, allied-country coordination, and limits on where U.S. robotics data can be processed, instead of a patchwork of state rules and one-off trade actions.
• Prioritize standards and an industry-led safety framework. The proposal asks regulators to build voluntary safety and capability benchmarks together with standards bodies and manufacturers before writing prescriptive rules, the same approach industrial robotics used to develop earlier human-robot collaboration standards.
• Create federal policies to support workforce development. Agility wants funding for technician-training pipelines, apprenticeships, and community-college partnerships that teach people to install, maintain, and supervise humanoid robots, plus incentives for employers whose robot deployments measurably cut workplace injuries.
• Lower the barrier to entry for small manufacturers to automate. Early-adopter tax credits, equipment-financing programs, and other incentives aimed specifically at small and midsize manufacturers would offset the six-figure upfront cost that currently keeps humanoid automation mostly inside large enterprises with dedicated capital budgets.
• Establish clear procurement pathways. Agility wants federal agencies and defense logistics commands to have defined channels for testing and buying humanoid robots, including non-weaponized, dual-use applications that build domestic manufacturing capacity without turning the machines into weapons.

Agility Is Co-Writing the Safety Standard It Wants Regulators to Adopt

Agility Robotics is not asking regulators to build the third recommendation's safety framework from nothing. Agility is one of three organizations leading the international working group drafting ISO 25785-1, the first safety standard written specifically for legged and other dynamically stable industrial robots. Its partners on the project are the Association for Advancing Automation and Boston Dynamics.

Kevin Reese, an Agility Robotics engineer, leads the drafting project. The working group published its first draft in May 2025, with final publication expected sometime in 2026 or 2027. That head start matters for recommendation three specifically: Agility is not asking Washington to invent a safety framework, only to formally recognize a voluntary one the industry is already writing.

What Lowering the Barrier to Entry Would Change for a Small Manufacturer

The recommendation to lower the barrier to entry for small manufacturers has the clearest effect on an ordinary shop floor. A single Digit unit costs in the low six figures to buy outright. Agility's existing leasing option is built around large enterprise customers, not a twenty-person contract shop with no in-house finance team.

Agility's fifth recommendation targets that math directly. It asks Washington to fund the tools that would change it.

A smaller shop could then automate one station on its own terms, instead of competing for the same enterprise-scale leasing deal as a warehouse operator many times its size. Whether that funding actually materializes decides whether a 20-person machine shop in Ohio can afford the same technology a national logistics company already uses.

Recommendation one carries a similar consequence upstream, for the manufacturer buying the robot rather than the shop trying to afford one. Rare-earth magnets and precision gearboxes used in humanoid actuators are still sourced largely from Asia, and Agility wants federal incentives that shift that sourcing domestically or to allied countries. A buyer feels that gap directly.

A supply chain concentrated outside the U.S. carries the same exposure that produced the July 2026 import restriction. The next trade action could move lead times and prices again, with little warning.

Recommendation four intersects with that same shop floor a few steps later, once the financing question is solved. A shop that finances its first humanoid still needs someone to service it. Agility's fourth recommendation, funding technician-training pipelines and community-college partnerships, targets exactly that gap, since a small manufacturer's new robot otherwise waits for a specialist to fly in.

The Broader 2026 Policy Push

Agility's proposal is not happening in isolation. Congress has a parallel effort underway: the National Commission on Robotics Act, introduced in the Senate in June 2026 by Senators Dave McCormick and John Hickenlooper. The bill would create an independent federal commission to study U.S. competitiveness in robotics and recommend policy changes.

A companion bill followed in the House the same year, and Johnson publicly backed the Senate version. She pointed to Agility's own manufacturing investment in Pittsburgh as evidence the industrial base needs coordinated federal support, not scattered state rules. She also urged the Senate to move quickly and pass the bill on a bipartisan basis.

The debate extends beyond any single bill or company statement. The Association for Advancing Automation, the trade group representing the wider robotics and automation industry, has published its own policy recommendations. Those include a national robotics strategy, deployment-linked tax incentives, and modernized manufacturing-extension programs, echoing several of the same points in Agility's own six.

Where the Proposal Could Run Into Resistance

The proposal's own internal logic creates at least one obvious tension. Recommendation two calls for tariffs and anti-dumping measures aimed at foreign robotics competitors, while recommendation five asks for programs that make automation cheaper for small manufacturers. Tariffs on imported components can raise the landed cost of a robot before any domestic-sourcing incentive has had years to mature a U.S. supply chain.

That leaves the same small manufacturers recommendation five is meant to help exposed to higher prices first, with relief arriving only once domestic suppliers scale up production. Recommendation three carries a different kind of friction. Agility is asking for voluntary, industry-led standards rather than binding regulation.

Worker-safety advocates have historically preferred enforceable rules with real penalties over a framework written mostly by the companies it would govern. The ISO working group gives Agility a real answer to that criticism: a documented, multi-company drafting process, not a single firm's internal guidelines. Even so, the standard is not law, and nothing requires an employer to follow it once published.

Procurement Pathways Look Most Likely to Move First

Procurement pathways are the recommendation most likely to move before the other five. Building a federal buying channel for humanoid robots is largely an agency decision, not new legislation. That makes it faster to act on than a tariff fight or a domestic-sourcing incentive, both of which have to survive a divided Congress.

The safety-standards recommendation has a head start for a different reason: industrial robotics already built the collaborative-standards playbook Agility wants to reuse. Extending that playbook to humanoids is a continuation of existing work, not a new institution. The sourcing and workforce recommendations sit at the other end, needing sustained federal budget commitments that are easy to stall whenever the coalition in Congress shifts.