On July 28, 2026, the FCC added foreign-produced “advanced robotic devices” and connected power inverters to its Covered List, following national security determinations issued by an executive branch interagency body the day before. In practice, this means new models in these categories can no longer receive the FCC equipment authorizations required to be imported, marketed, or sold in the U.S. The plain-language version of that story, “foreign robots are banned,” is a reasonable starting point, but the actual mechanics reach further than that framing suggests. It puts a hard regulatory edge on a question the robotics industry has been circling for a while: what happens to your automation strategy when the hardware underneath it becomes a liability? 

The definition behind the rule is broad. An advanced robotic device is any mechanical, mobile device capable of locomotion, obstacle avoidance, or navigation — including humanoids and quadrupeds — that operates away from a human operator, weighs more than 4.4 pounds (including any dock or charging station), and carries both environmental sensors and network connectivity. Fixed industrial robots, traditional stationary manufacturing arms, fall outside that definition. Mobile warehouse, inspection, and material-handling robots don’t get the same automatic pass; whether a given model qualifies as “advanced” depends on the same mobility, sensing, and connectivity criteria, which means most of what’s actually on a warehouse floor today needs its own case-by-case read. 

Buried in that scope is the detail that matters most for how companies should read this: “foreign-produced” isn’t defined by a manufacturer’s nationality or headquarters. It turns on whether a product qualifies as a “domestic end product” under existing Buy American trade rules. That single distinction changes who’s actually exposed, and it’s a big part of why this rule is more consequential than it might look at first glance. Robots that already hold FCC equipment authorization are treated differently under the rule: they can continue to be sold and used, though they generally can’t be modified or updated beyond certain software and firmware changes. For new foreign-produced devices, companies can also apply for a Conditional Approval through the Department of War, with an application deadline of January 1, 2028. (Full details are available in the FCC’s public notice.) 

The timing puts this squarely in front of an industry that’s been growing fast. Anyone who’s walked the floor at Automate, ProMat, or MODEX over the last few years has watched the humanoid and mobile robot sections expand at every show. That growth is exactly why this rule is generating so much conversation right now, and why it’s worth working through what it actually means, not just for compliance, but for how companies build their automation strategies going forward. 

A Pattern, Not an Incident 

This isn’t an isolated action. It’s the third or fourth categorywide, place-of-production restriction the FCC has issued in eight months — foreign-produced drones in December 2025, foreign-produced routers in March 2026, and now advanced robotic devices and power inverters in July 2026. Treated as a sequence rather than a single event, the pattern is the more useful signal: once a product category draws a national security determination, the FCC has moved to restrict it categorically, not case by case. 

The robotics industry is still working through exactly how far this particular rule reaches, particularly across warehousing and manufacturing. What’s already clear is the direction: anything that can be controlled remotely, or that’s intelligent enough to collect and act on data from its environment, is now under regulatory scrutiny. That’s not a narrow carve-out. It touches a meaningful share of what’s on the floor at any major automation trade show today. 

The specific rule matters less than the dynamic underneath it, and it’s one worth naming directly: automation strategies built entirely around a single robot brand carry an exposure that goes beyond any one regulation. Hardware evolves, and regulation moves independently of it. A strategy anchored to one brand is a bet that neither will change in a way that matters, and this rule is simply the most recent — and most visible — example yet of that bet not paying off. 

Made in the USA Isn’t the Safety Net Most Companies Think It Is 

That same dynamic shows up again in the rule’s most overlooked detail: “foreign-produced” is defined by whether a product qualifies as a domestic end product, not by where a company is headquartered. That distinction changes who’s actually exposed. 

Most U.S.-headquartered robotics companies manufacture overseas, largely because the domestic supply chain never developed the scale needed to build robots — sensors, lidars, cameras, motors, batteries — at the quality and cost the market demands. As a result, a robot assembled in the U.S. can still be built almost entirely from subcomponents sourced abroad, including the control systems, power converters, and connectivity that make remote access possible in the first place. Assembly location alone doesn’t resolve that exposure. 

One likely workaround already being discussed in the industry is importing components or subassemblies and completing final assembly in the U.S. before seeking FCC authorization domestically, effectively building toward the “domestic end product” standard rather than assuming assembly location alone satisfies it. That also means companies marketed as “U.S.-built” aren’t necessarily exempt if their bill of materials still leans heavily on imported parts; the label and the regulatory standard aren’t the same thing. Read together, the rule’s direction points toward a higher bar for what counts as truly U.S.-manufactured than the industry has typically used, and that bar looks likely to keep tightening as more guidance comes out. 

The exposure could also run in the other direction. Global manufacturers, logistics companies, and retailers operating across regions may see other countries respond in kind against U.S.-made robots, which would push multinational buyers toward deploying different hardware, potentially different software stacks entirely, for the same workflow depending on the region. It’s exactly the kind of fragmentation an automation strategy should be built to withstand, not one it should be vulnerable to. 

The Real Shift: From “Which Robot” to “What Workflow” 

Zoom out, and this rule is one data point in a longer-running trend. For decades, the robotics industry asked which actuator or which sensor was better. That’s no longer the most useful question, because most hardware in a given category, arms, AMRs, AGVs, mobile manipulators, has converged on similar core capabilities. Pick up a pallet, and there are thousands of ways to do it, but the underlying capability is commoditized; humanoids are newer and less settled, but the same pattern is already emerging there too. 

Once hardware looks that similar, the value shifts to what the system actually does, not what brand built it. That reframes what an automation strategy needs to be resilient to. It’s no longer just about picking the right robot today. It’s about building a strategy flexible enough to absorb hardware changes and regulatory changes alike, without disrupting the workflow the robot was deployed to solve in the first place. 

Where Destro Fits In 

This is the exact problem Destro was built to solve, and it’s why this rule lands differently for us than it might for a company selling a single robot brand. 

Our view has always been that companies aren’t really buying a robot. They’re buying a solution to a workflow. The workflow is what needs to survive hardware and policy changes, not any particular chassis. That’s why we built a robot-agnostic intelligence layer with two components: 

• Mothership OS — the off-robot shared brain that persists across hardware generations, so that as robots are swapped out over three or four years, the workflow logic and throughput gains carry forward regardless of which chassis is doing the work.
  
  

• Vision OS — the on-body intelligence layer, purpose-built for precise manipulation at the carton and each level. Running at the edge, it autonomously identifies SKUs, counts inventory, and executes pick-and-place tasks without barcode scanning, orchestrating the actuators directly on the robot to turn high-level task instructions from Mothership OS into reliable physical execution. 

Robots may change, and in this new regulatory environment, they may have to. But the brain running them shouldn't have to. 

Destro builds a robot-agnostic intelligence layer that lets manufacturing, logistics, and retail operators switch robot hardware without rebuilding their automation strategy from scratch. Learn more.