8 Experts Weigh In On The FCC Foreign Robot Ban: Good Or Bad?

Cost is an issue

One thing it’s going to do for sure is add cost. Think 3X the cost, according to one brand-new study, making a $50,000 robot cost $150,000. And it’s doing to make finding raw materials, components, and suppliers a lot more interesting, shall we say.

OpenMind founder and CEO Jan Liphardt walked me through his new Physical AI Readiness Index on my podcast this week. (The index will be released next week; I will be covering that in full in a separate post.) He puts the cost of building a humanoid robot without Chinese suppliers at 2.85X more than it would be without the ban. It makes sense: Chinese vendors shipped 97% of the world’s humanoid robots in the first half of 2026, and China has built up the world’s best and most cohesive robotic supply chain.

Plus, whatever badge ends up on the chassis, the bill of materials behind it still needs a lot of China: rare earth metals, magnets, gears, actuators, bearings, PCBs, sensors. Alternatively, it needs a lot of creative sourcing and building, like from humanoid robot makers 1X and Figure, which are vertically integrating as much as possible to build almost everything in-house.

But supply chain is a bigger issue

Perhaps the hardest part isn’t any single missing component. It’s supply chain – or supply web – density. For most of what goes into a robot, there are at least some US suppliers. The problem is that they’re mostly scattered across a continent. In Shenzhen, Liphardt points out, everything you need sits inside a roughly 10-kilometer radius, which makes product iteration and innovation incredibly fast.

You can literally walk across the street to the person who makes your gears.

Nine of the top ten robot-building metro clusters in OpenMind’s index are in East Asia. The Bay Area is ninth. Boston is 12th, Pittsburgh follows, Detroit is 21st, and Austin — home of Apptronik — doesn’t crack the list at all. The overall index has China first at 83 out of 100, Japan second at 69.6, and the US third at 69.

The domestic industry response to the FCC rule has been broadly supportive. That’s unsurprising, given that it largely protects their business prospects in the United States. Gavin Kenneally, CEO of Ghost Robotics — which competes head-to-head with Unitree on quadruped robots — makes the sharpest security case, arguing that spyware on Chinese robots inside the US is an active fact rather than a hypothetical, and that this innovation battle is not really company-versus-company but private American firms against a coordinated national strategy, similar to what has already happened with drones, EVs, and solar.

Samuel Reeves of FORT Robotics agrees on the threat, but objects to the blast radius of the ban: if the concern is national security, he says, the policy should target Chinese companies specifically rather than sweeping in allies like Canada, Europe and Israel. Nic Radford of Persona AI says it’s more about market validation — enterprise buyers of humanoid and legged robots were already asking about trust, security, parts and long-term service before the FCC weighed in.

Perhaps the most nuanced response came from Brad Porter, the CEO of Cobot. He previously scaled Amazon’s robotics fleet past 500,000 units.

Porter calls the rule a sensible and carefully designed first step, but his argument is that protection and leadership are different things. And that America keeps confusing them. What took Amazon from 50,000 robots to 500,000 in about four years wasn’t regulation, he says.: it was having the most ambitious robotics customer in the world as a customer. His prescription is demand: he wants the US government to be the most ambitious early buyer in factories, hospitals, labs, shipyards, disaster response and munitions logistics. Which, of course, is a very similar national strategy to the one that China has followed.

Porter’s warning is that failure won’t be dramatic.

“Failure looks quiet.”

Will the FCC ban hurt robot sales growth?

It might be pretty loud, however, if the IDC’s predictions come true. IDC modeled the FCC ban through 2030 across multiple robot categories such as household cleaning, professional and commercial, and humanoid and found that the category the rule was ostensibly written to protect takes the deepest hit.

According to IDC projections, US humanoid volumes run 41% below baseline in 2027 and 58% below by 2030 thanks to the FCC ban. That’s nearly four of every five projected 2030 revenue dollars evaporating, mostly because US builders like Figure, Apptronik and Boston Dynamics remain dependent on Chinese motors, actuators, batteries and rare earths, and because IDC doesn’t expect scaled US mass production until the end of 2028. Household cleaning, which is already foreign-dominated, is the most resilient, trailing baseline by just 4% in 2027 and 18% in 2030.

According to those IDC projections, the FCC ban results in total foregone US robotics revenue across those categories of more than $6 billion between 2026 and 2030.

Part of the problem: this doesn’t just hurt US robot manufacturers at home. It also impacts their global sales potential. If they’re 3X as expensive as the Chinese option, few other nations are going to choose an American solution. For example, Europe is already the largest consumer of these tracked robot categories, at more than double America’s share, and IDC’s model sends the displaced hardware there.

If there’s a reason to think even IDC’s numbers are optimistic, it comes from the component layer.

Michael Murray, CEO of the microdisplay maker Kopin (KOPN), is watching the same reshoring push play out one tier down in drones — the industry Reeves cites as the cautionary tale — and his warning is that a lot of American “de-risking” is essentially just theater.

US defense firms buying microdisplays from Japanese or Korean vendors instead of Chinese ones believe they’ve removed their China exposure. They mostly haven’t: those vendors build for consumer volumes and source OLED deposition material, wire bonds and digital wafers from China to hit consumer price points. Murray calls that position “false and irresponsible.” He also notes that a microdisplay isn’t a passive part but an active one, with a processor, memory and the ability to talk to other electronics, which makes it a system and therefore an attack surface.

Chinese smart TVs are already barred from government buildings on exactly that logic.

Murray is the most optimistic voice here on timing, and notably the only one giving dates. Kopin will be ready to produce both OLED and MicroLED domestically in 2027, a process he says started 18 months ago. (Which gives you a sense of the time frames involved in rebuilding supply chains domestically.) He thinks the US drone supply chain can be “fairly independent” in 12 to 18 months and fully independent in 24. He does expect tariffs to raise prices, though.

Here’s what the people building these things had to say, in their own words.

Brad Porter, CEO, Cobot

Porter was previously VP and Distinguished Engineer at Amazon Robotics.

John Koetsier: The FCC says this is “country-neutral,” but it hits US firms that assemble abroad or use Chinese parts. Does the rule help American robotics or handicap it?

Porter: I think it’s a sensible first step, and it appears to have been designed carefully. It sets expectations around how much of a robot should be sourced domestically rather than banning foreign components outright, so US companies can keep accessing the best sensors, motors, and computing in the world while domestic supply grows. That said, protectionist measures can be double-edged. The question is how you bring actuators, motors, and other elements of the supply chain back, and how you make sure you have a resilient base. That takes investment not just restrictions.

John Koetsier: You say protection won’t create leadership. What has to happen in the next 18 months to make this an advantage, not just a wall?

Porter: Government has a real role to play here beyond restrictions like supporting investment in domestic manufacturing, with loans and other market-compatible incentives for reindustrialization. Making sure robotics companies have access to capital and encouraging adoption of American robotics technology, because robots get better by doing real production work at scale. America’s industrial base needs more robots, faster. 18 months from now, the measure of success is whether more American robots are doing real work in American facilities than today.

John Koetsier: Which is the harder gap to close: the AI and software stack, or the physical supply chain — motors, actuators, batteries?

Porter: The supply chain, but I don’t see a fundamental capability gap. The United States has extremely capable engineers, manufacturing expertise and a lot of determination. The idea that the U.S. doesn’t manufacture anything is a little overblown. The U.S. manufactures a lot. What China has built is concentration. I’ve been to trade shows there where virtually every component you need to build a robot is available immediately and cheaply. Here, you order the same part online and wait a few days. That density took decades, and rebuilding it takes volume, which comes, in large part, from demand.

John Koetsier: You scaled Amazon past 500,000 robots. Could you have built that under today’s rules?

Porter: We would have navigated the rules. What actually let us go from 50,000 to 500,000 robots in about four years was demand. Amazon was the most ambitious robotics customer in the world, and it happened to be our own company, willing to redesign buildings and retrain teams to deploy automation at scale. That ambition and drive are what has been missing in American robotics far more than any regulation.

John Koetsier: You want the US to be an “ambitious early customer.” Who specifically?

Porter: Anywhere materials move at human scale. Factories, hospitals, labs, office buildings, logistics facilities. People are already moving materials around in all of those places, and a lot of that work is repetitive, physically punishing, or puts people near real hazards. Those jobs are hard to fill, and the government operates these environments at enormous scale. I’m particularly proud of the work Cobot is doing to support the Department of Defense on munitions logistics, where getting people out of harm’s way really matters. There’s room to be much more ambitious. Naval shipyards. Disaster response, where the environment is dangerous by definition. Eventually somebody is going to be moving materials around on the Moon, and it’s not going to be a person pushing a cart. Somebody is going to be the first customer at scale for this new generation of robots, and America should want that customer to be American.

John Koetsier: If protectionism buys a few years of runway and it gets wasted, what does failure look like in 2030?

Porter: Failure looks quiet. American robotics is still boutique in 2030, other countries are deploying robots by the millions into their own factories, and the restrictions are still standing with nothing built behind them. What would make that so frustrating is that there’s no fundamental capability gap here. We have the engineers, we have the capital, we have the determination. If it doesn’t happen, it’s because we didn’t follow through on the investment and the adoption.

Gavin Kenneally, CEO, Ghost Robotics

“People tend to talk about the risks from Chinese robotic companies as if they’re hypothetical, but they’re really not. Active and purposeful spyware is deployed inside the U.S. on Chinese robots. Examples of predatory pricing abound. And this isn’t just a competition between U.S. and Chinese robotics companies; it’s between private U.S. companies and China’s coordinated national strategy. We’ve seen China run this playbook before with drones, EVs, and solar panels. There’s more that needs to be done, but I support the FCC’s decision. If today’s announcement encourages stronger cybersecurity and a more level competitive environment, that’s good for customers and good for the robotics industry.”

Samuel Reeves, CEO, FORT Robotics

“The drone industry has spent years rebuilding itself at enormous expense after DJI effectively wiped out much of the domestic market. I think it’s foolhardy to allow China to heavily subsidize strategic technology industries and then enable those companies to undermine nascent American industries. The traditional concept of comparative advantage was developed long before today’s technology landscape and doesn’t account for competition from a state-backed adversary that doesn’t always play by the same rules.

That said, I think the ban goes too far if it sweeps in companies from trusted allies such as Canada, Europe, and Israel. If the concern is national security, the policy should be targeted specifically at Chinese companies.

The safety and security issues are absolutely significant. As robots become increasingly connected and autonomous, they introduce real cybersecurity and operational risks. We have roughly 650 customers, and perhaps only one is in China. Personally, I wouldn’t put a Unitree robot in my home or office.”

Nic Radford, CEO, Persona AI

“We’re entering a new phase for robotics. Humanoid and legged robots are becoming platforms that businesses depend on every day. The FCC’s position underscores something enterprise customers have known all along: when you’re evaluating humanoid robots for critical operations, trust, security, and long-term support have to come first. Customers making real investments in these systems need to know they can scrutinize the technology, maintain it over time, and get parts and service quickly without relying on a fragile overseas supply chain.”

Michael Murray, CEO, Kopin (KOPN)

Kopin makes microdisplays and optical systems, including for FPV drone goggles, aviation helmets, night vision and weapon sights. Murray’s comments address the parallel push to onshore the drone supply chain.

John Koetsier: Are tariffs enough to break U.S. dependence on China?

Murray: They are a great start and the government is being considerate of the time it takes to transition to newer and U.S. made technologies. A further ban on ‘active’ components in drones is also required since any microprocessor (in a MicroDisplay, for example) is a threat to our defense networks.

John Koetsier: What drone components are most vulnerable today?

Murray: In my view, drone motors and displays are critical components for our supply chain since they are the most heavily purchased items from China. This is why UMAC, ONDAS and Kopin are so important, as they are major manufacturers of several of these components.

The reason being, most, if not all, of the FPV goggles are manufactured in China with either Chinese or Japanese displays which are active components. An active component has a processor, memory, and the ability to communicate with other electronics which makes a display a system and, therefore, a vulnerability which most do not think about.

To underscore the point, Chinese ‘smart’ TVs are not allowed in any Government building for this same reason — a MicroDisplay is no different albeit the wireless component is not included in our displays.

John Koetsier: Why are displays a national-security issue?

Murray: 1. All MicroDisplays are active components. An active component has a processor, memory, and the ability to communicate with other electronics which makes a display a system and, therefore, a vulnerability to access the Defense Network and certainly to the control of the system it resides in.

1. Security of Supply. If China were to stop shipping displays to the USA, our FPV goggles would cease production and we would not have ANY drone fleet to field in the battle.
2. The major issue we are seeing now in the market which is not understood by most, is several U.S. Companies believe that buying from Japan or Korea is ‘good enough’ to state that they do not have exposure to China, however, that is false and irresponsible. Japanese and Korean vendors build their MicroDisplays for consumer applications (like watches, gaming systems, etc) which require significant amounts of materials like the OLED deposition material and components like wire bonds and digital wafers, etc.

Several of these displays will use OLED deposition material, wirebonds, and even digital wafer manufacturing from Chinese sources so that they can compete on the low costs and satisfy the high volume demands of the consumer market. Therefore, even though defense firms can say they are buying from ‘non-Chinese’ sources, they actually are not.

John Koetsier: What does China still dominate?

Murray: Micro OLED and MicroLED technologies and volume production. Interestingly, China was spending hundreds of billions in OLED MicroDisplay capacity until recently and pivoted to MicroLEDs in the past several years. The answer as to why is both interesting and unfortunate for the U.S.

The Chinese Military tested OLED displays in programs like aviation helmets, night vision goggles, and thermal weapon sights and the performance was quite good over that of LCDs. However, the displays produced artifact images called burn-in and the brightness of the OLED displays degraded quickly when the displays were operated at high brightness levels when used in the daytime. In one specific day-time application the OLED display would burn out in less than one week of operation due to the amount of current required to operate the display degrading the cheaper OLED material so quickly. Furthermore, the ghosting or burn in of symbology images became an issue for weapon sights and pilot helmets alike.

So, China changed direction quickly and focused on developing MicroLEDs which are far brighter than OLEDs and, instead of organic material being used, a static LED is used which offers significant peak brightness of millions of nits, at lower power consumption and much lower heat dissipation.

With this new technology, China’s military can see us before we see them using these displays in head mounted displays, night vision and weapon sights. Furthermore, their long range targeting is also improved due to the smaller pixel structures of MicroLED vs. OLED displays. The smaller the pixel, the higher the magnification can be used on the image capture side of a weapon sight, etc.

The advantages of OLED MicroDisplays are they are far cheaper to manufacture than MicroLEDs, and many night time applications remain perfect for OLED displays along with medical, industrial, and consumer applications.

With respect to FPV drone headsets, as an example, Sony has an established OLED-material relationship with Idemitsu, and Idemitsu manufactures OLED materials in China. These systems are fully or partially occluded, meaning the display is fully enclosed in the headset (virtual reality) which is not see through and very close to the user’s eye, so extreme brightness and longevity of operation is not an issue in this application, making OLEDs the right choice. If these headsets move to ‘see through’ or augmented reality, the solution will likely require a MicroLED to be bright enough to compete with the daytime sun and surroundings of a drone pilot on the battlefield.

John Koetsier: Can U.S. manufacturing compete on cost?

Murray: Of course. But, we need to innovate and utilize our creativity and automation. Chinese companies do not need to worry about yields, profits, or paying healthcare insurance but, we can use AI and automation to level the playing field. Kopin has been investing in AI assisted automation for several years now and our new OLED deposition line is almost fully automated.

John Koetsier: How fast can production realistically move onshore?

Murray: We will be ready to produce OLED and MicroLEDs in 2027 and we began the process over 18 months ago.

John Koetsier: Will tariffs raise drone prices significantly?

Murray: Yes. They will raise prices to the point where U.S. manufacturers can compete now. Which is the point.

John Koetsier: How independent can the U.S. drone supply chain actually become?

Murray: In the next 12-18 months we can become fairly independent and fully independent in 24 months.

John Koetsier: Which manufacturers will benefit most?

Murray: Firms like Unusual Machines who make several critical components along with ONDAS and the Drone Dominance award winners like Neros will benefit greatly. Kopin will benefit from providing companies like these with our displays and optical solutions.

Jan Liphardt, founder and CEO, OpenMind

OpenMind builds an open, hardware-agnostic software layer for robots — think Android for robotics. Liphardt is also a professor of bioengineering at Stanford. These comments are from our podcast conversation about OpenMind’s Physical AI Readiness Index, which ranks China first at 83 out of 100, Japan second at 69.6, and the US third at 69.

On the cost problem:

“It’s not very useful for the US to have robots that cost 3X more than anywhere else. That immediately puts us at a disadvantage. And so somehow as a country we’re going to have to make a decision: do we want to have a credible supply chain for physical AI?”

On what’s missing:

“There’s price, but then there’s also the simple question of do we actually have everything here we need in the US? Rare earth metals come to mind, magnets come to mind, and all the gears and actuators and chips and so forth. So it’s not just a question of how much is it going to cost, but how much is it going to take for us to rebuild a lot of that here in the US?”

On supplier density, and why Shenzhen wins:

“One of the things that China has done very well is spatially concentrate everything that you might need in these incredibly focused areas — Shenzhen comes to mind — where within about a 10-kilometer radial distance you have everything you could possibly need, from carbon fiber to CNC to bearings to sensors to PCBs. Because it’s so concentrated, you have these incredibly rapid iteration cycles. If you’re building something, you can go across the street, you can talk to someone who builds gears. We certainly have a lot of the pieces potentially here in the US, but they’re typically in very different places. And those pieces that we do have are typically very small and there aren’t a lot of choices.”

On what the US should repurpose:

“We’ve been building cars here in the US since cars were invented. Ideally we could imagine repurposing some of that industrial base or some of those capabilities and really focusing them on maybe what will come after cars, which are advanced robots.”

On why the best AI models won’t be enough:

“Many people think that physical AI is all about AI. There’s this notion of if you have the best models, then you’re also going to win physical AI. But based on what we can tell, that’s not true. Part of the reason that’s not true is that countries, especially China, are open sourcing advanced models also for physical AI. That, of course, re-levels the playing field. Ultimately, it all boils down to who has the hardware supply chain for materials, ball bearings, sensors, magnets, PCBs, and who’s able to integrate all of that. From that vantage point, Japan is stronger than the US, which will come as a surprise to many people.”

On China’s real advantage:

“Because they’re manufacturing a significant fraction of what the world is buying, that puts them in a position to collect training data for manufacturing and automation on scales that are unimaginable in the US. So even though we have a razor thin margin in AI in general, the special case of models for physical AI — we may be at a disadvantage there simply because of our smaller manufacturing base.”

On the lesson from Ukraine:

“Ukraine is building millions of drones in very simple conditions, and they are able to iterate the design of these drones on time scales of days or weeks. I heard one example where experiences from the front lines are being transmitted to people building the drones that same day. What that is showing is that even in situations where you don’t have the entire supply chain, there are things you can do to greatly accelerate innovation. Mapped onto the US, you can imagine different states making it much easier to field test physical AI in hospitals, or to build roads or bridges.”

On whether the gap is closable:

“The gap is definitely surmountable. And so we shouldn’t be afraid. This is, if anything, information that helps us prioritize.”

On vertical integration at Tesla, Figure and 1X:

“In hindsight, with the new FCC rules, it makes complete sense… Of course, then you somehow need to work against this 3X cost differential. So does it make sense to vertically integrate? Absolutely. Is it difficult to do so? Yes. Is it difficult to do so and generate a cost-effective robot? Yes. But strategically, I think it makes a lot of sense.”

IDC: Ryan Reith, Group VP, and Navkendar Singh, Associate VP

From IDC’s analysis of the restriction, modeling a scenario in which the rule holds through 2030 across the robot categories IDC tracks in depth — household cleaning, professional and commercial, and humanoid.

On the scale of the hit:

“Using IDC’s blended average selling price assumptions, that growth gap between the restriction scenario and IDC’s published baseline compounds into more than $6 billion in foregone US robotics revenue between 2026 and 2030, with the single worst year, 2030, alone accounting for above $2 billion of it.”

On which category actually suffers:

“Humanoid robots, the category this restriction was ostensibly designed to protect, see the deepest cut of all: a 41% gap in 2027 widening to 58% by 2030. Look at the dollars and it is starker still: nearly four out of every five dollars of the category’s projected 2030 US revenue evaporates under the worst case.”

On why the badge doesn’t matter:

“It is about how dependent even vertically integrating US builders — Figure, Apptronik, Boston Dynamics — still are on China for the motors, actuators, batteries, and rare earth materials that go into every unit… IDC estimates China accounts for 82% of global humanoid shipments today. Until US builders reach that scale, whatever badge sits on the chassis, the bill of materials behind it still reads China.”

On who actually benefits:

“IDC’s model nudges professional and commercial and humanoid forecasts higher across EMEA and other parts of the world, particularly the UK, France, Germany, and Eastern Europe, as Chinese vendors redirect resources toward markets open to them. China’s own domestic forecast barely moves. It is a story about Europe, already the largest buyer of these tracked robot categories, getting a little larger still.”

On what vendors will actually do:

“Reshoring consumer robotics specifically is unlikely given thin margins; this accelerates a supply chain diversification trend already in motion rather than starting a new one. Tesla sourcing its own AI chips for Optimus across TSMC and Samsung’s Texas fab is one model other vendors may follow: hedge the component that carries the most geopolitical risk, chips today, magnets and actuators next, rather than moving final assembly to the US.”

The bottom line:

“A single trade rule, applied evenly on paper, lands completely unevenly in practice, and not in the direction most people assumed… It means the market is quietly sorting into two tracks: mass market categories that keep growing around the US, and higher value categories where the appearance of protection outpaces the reality of supply chain independence.”