In July and August, Washington tightened restrictions on foreign-made advanced robotic systems and imposed steep tariffs on imported drones and their components, both moves citing national-security concerns. The drone tariffs take effect in September, with additional component tariffs following in 2027.

These moves are part of a broader U.S. effort to restrict foreign technology in strategically important industries. The FCC’s Covered List, established in 2021, initially targeted telecommunications and surveillance equipment from companies including Huawei, ZTE and Hikvision before expanding to foreign-made drones and, most recently, to advanced robotic devices.

The latest move comes as Chinese manufacturers have built commanding positions in both drones and humanoid robots, often competing at prices U.S. and European rivals struggle to match.

Taken together, the restrictions are raising a bigger question for the global robotics industry: If Chinese drones and humanoids are increasingly shut out of the U.S., where does the competition move next?

The restrictions may protect parts of the American market, but they don’t directly address China’s global manufacturing scale and cost advantages.

Industry analysts and executives who spoke with TechCrunch said the result may be less a clean U.S.-China split than a more fragmented global market, with Chinese companies expanding elsewhere while U.S. and allied manufacturers compete in markets where security requirements matter more.

The Scale Gap

The U.S. and Chinese robotics industries remain deeply connected, but the two countries enter the competition with very different advantages. Unlike semiconductors, robotics does not hinge on a single technology that one country can easily control, said Ankur Saxena, an investment director at TDK Ventures.

China dominates global humanoid robot manufacturing, with global shipments hitting 22,000 units in the first half of this year — the vast majority from Chinese manufacturers — according to a report by Counterpoint. U.S. companies, by contrast, are operating at a far smaller scale, said Soumen Mandal, a principal analyst at Counterpoint Research.

The world’s five largest humanoid robot makers by shipments — AgiBot, Unitree, Galbot, UBTECH and Leju Robotics — were all Chinese and together accounted for 86% of global shipments in the first half of 2026, according to Counterpoint.

That advantage could compound. Lower prices allow Chinese manufacturers to put more robots into use, generating real-world data that can improve their technology. Higher production volumes, in turn, can drive costs down further, Saxena said.

Mandal said Chinese humanoid makers are also pushing costs down by bringing more of the technology stack in-house and drawing on China’s existing manufacturing base. Unitree, for example, is developing more components internally, while automakers such as XPeng can draw on their experience in chips and vehicle manufacturing as they move into robotics.

“The United States leads in frontier AI, software and semiconductor innovation,” Saxena told TechCrunch. “China leads in manufacturing scale, supply-chain depth and cost.”

That manufacturing edge has let Chinese companies cut humanoid prices faster than most U.S. competitors can match.

“You cannot sanction your way around a cost curve. You can only out-build it, and America has yet to begin making the decade-long investment that will require,” Saxena said.

Where Does China Go Next?

The answer may increasingly be outside the U.S. Even if Chinese robotics companies lose access to the American market, they still have a large domestic market and room to expand elsewhere, particularly in regions where demand for affordable automation is growing, Saxena said.

Chinese robotics companies are already targeting price-sensitive markets with severe labor shortages across Europe, Southeast Asia, Latin America and the Middle East, said Mandal.

Mandal expects humanoid makers to follow a path similar to Chinese electric-vehicle companies: build scale at home, expand into overseas markets, and eventually establish local production. Countries facing labor shortages and demographic decline could become early markets for humanoids, particularly in manufacturing, where robots can take on repetitive work.

The drone market offers an early glimpse of what that more fragmented robotics landscape could look like. The industry is increasingly splitting into two ecosystems: a U.S.-led market built around American-made, NDAA-compliant systems, and a China-led market focused on low-cost, high-volume production, said Bentzion Levinson, founder and CEO of Virginia-based drone maker Heven AeroTech.

Levinson said Western manufacturers are unlikely to beat Chinese companies in the low-end consumer drone market, where cost remains a major advantage. Instead, U.S. and allied companies could increasingly compete in long-range autonomous systems for defense and critical infrastructure, where security requirements carry more weight.

Levinson sees the next competitive frontier shifting from the drones themselves to the technology that powers them and the equipment they carry. “The next battleground is over who owns the next-gen energy and payload architecture,” he said, pointing to battery constraints in particular. As drones become more capable, he added, battery limitations could make power systems an increasingly important point of competition.

Agility Robotics welcomed the FCC’s decision in July, saying it could address security concerns around foreign-made advanced robots before they become deeply embedded in the U.S. market, as has happened in the drone industry. The company pointed to its Digit humanoid, which is designed and assembled in the U.S., while also calling for continued access to the tools and technologies needed to advance robotics research.

A More Regional Robotics Market

“The alternative to China isn’t a purely domestic U.S. supply chain; it’s a diversified allied one,” Saxena said.

That could create opportunities elsewhere in Asia. Japan has decades of experience in industrial robotics and precision manufacturing, South Korea brings strengths in electronics, batteries and automobiles, and Taiwan is a major player in semiconductors. But none can simply replace China, Saxena said, given how deeply Chinese components remain embedded across the global robotics industry.

Asian manufacturers could emerge as a middle ground between lower-cost Chinese robots and more expensive U.S. offerings, Mandal said. South Korea’s Hyundai, which owns Boston Dynamics, and Japan’s Toyota are among the automakers investing in robotics, drawing on their expertise in vehicles, manufacturing and autonomous systems as they move into humanoid robots.

Yang Fang of Beagle Technology, a California-based agtech startup that uses AI and robotics software to turn conventional farm equipment into autonomous machines, told TechCrunch that robotics is likely to become more regional as companies design machines for the labor needs, working conditions and customers in their home markets. Chinese robotics companies, for example, may focus on products suited to China and nearby markets, while U.S. companies are more likely to build for industries across North America, he said.

The result may not be two neatly separated U.S.- and China-led robotics industries. Instead, the restrictions could accelerate the emergence of regional markets: Chinese companies competing on cost and scale across much of the world, U.S. and allied manufacturers gaining ground where security requirements matter most, and manufacturers in Japan, Taiwan and South Korea trying to carve out space between the two.

An appeals court in California has resurrected a 2022 lawsuit over claims Apple convinced them not to use a third-party repair service to fix cracked iPhone screens.

In the last few years, Apple has been more open to enabling repairs of its hardware performed outside of its support ecosystem. While its position has relaxed a lot, it still has to deal with its policies from prior years.

A filing for the Court of Appeal for the State of California in August reveals a panel has reversed an earlier court’s ruling to quash a lawsuit over Apple’s repair policies. The three-judge appeals Sixth Appellate District panel has brought the lawsuit back to life.

In 2022, plaintiffs Jesse Granato and Janice Zarad said they were told by an Apple employee in 2021 that using a third-party repair service would void their warranty coverage. The Courthouse News Service reports the plaintiffs claim this encouraged them to pay for a screen repair through Apple itself.

The lawsuit accused Apple of inflating the cost of independent repairs in various ways, including withholding parts, manuals, and software. It also allegedly conditioned warranties on using authorized repair shops.

A Santa Clara Superior Court judge agreed with Apple’s response that the lawsuit lacked standing under the state’s unfair competition law. Something the Court of Appeal overturned in 2026.

Repairing the lawsuit

In the unanimous decision, the judges explain that, as pleaded, Apple’s practices at the time “substantially injure consumers by allowing Apple to extract inflated payments for repairs or to increase demand for new products by incentivizing replacement over repair.”

The plaintiffs also couldn’t identify any benefit to Apple’s repair policies that outweighed the harm they caused, wrote acting presiding Associate Justice Cynthia Lie.

The panel decided that the plaintiffs had adequately pleaded under the unfair competition claim.

The plaintiffs also attempted to claim that Apple had violated the California 2024 Right to Repair Act, which Apple actually backed. However, the panel pointed out that the law was enacted after the initial suit was filed.

There was also an alleged violation of the 1975 Magnuson-Moss Warranty Act, by Apple allegedly refusing and instructing authorized repair shops to refuse in-warranty services, if repairs were performed elsewhere outside of any warranty.

However, while Lie explains the plaintiffs hadn’t adequately managed to claim the violation of the law, it was ruled that the claims violated the spirit of the law.

The case, revived, has been remanded to a lower court.

The introduction of optical media in the form of CDs meant a revolution in the world of gaming consoles, escaping the restrictions of a few dozen MB of storage and instead offering a theoretically infinite amount of storage through the miracle and tedium of swapping discs. After the SNES narrowly escaped getting a CD add-on and the N64 doubled down on cartridges, we can now at least get an impression of what it’d be like if the Nintendo Entertainment System had been gifted a CD add-on back in the 1980s, courtesy of a project by [Throaty Mumbo].

There’s also an accompanying video containing typical hijinks and a demonstration of this system. Initially [Throaty Mumbo] was going to make a SNES CD add-on to match that console’s initial prototype, but doing it for the NES seemed more fun. Obviously, since the NES is quite limited hardware-wise it was always going to be a struggle, even if the original front-loading NES conveniently has a mostly unused expansion slot.

On the custom PCB there is an RP2350B microcontroller that mediates between an Everdrive N8 Pro cartridge and an IDE CD drive, along with a PCM5102 audio DAC. The expansion port is hereby used to receive the audio samples on its audio mix input pin, along with power. After boot the game ROM is read off the CD by the MCU and streamed over USB to the Everdrive.

Combined with the previous work [Throaty Mumbo] did to revive the long-defunct Japanese online service for the NES, it’s an exciting time for fans of the iconic console.

In the very first moments of the universe, matter didn’t exist as we know today. A millionth of a second or so after the big bang, the universe was a dense, hot soup scientists call quark-gluon plasma (QGP). For several years, particle colliders—which smash molecules together at nearly the speed of light—have been able to replicate this state, but often using heavy elements like lead.

Now, a recent experiment by the European Organization for Nuclear Research (also known as CERN from its French acronym) has demonstrated this plasma can be produced by much smaller collisions. Since there’s no longer an accessible natural source of this primordial sludge, these micro big bangs can help reveal what happened in the first few minutes of our universe.

First, a little context. Quarks are what make up protons and neutrons, which, in turn, are the building blocks of atoms and thus all matter. Meanwhile, gluons—as their name suggests—stick quarks together.

During the first microseconds of the universe, quarks and gluons were not yet confined within protons and neutrons but instead formed an extremely hot plasma. As the universe expanded, the matter cooled, and the quarks condensed into larger particles.

After decades of studying QGP in large nuclear collisions, physicists are now trying to understand the limits of this strange state of matter. In particular, they are exploring just how much they can scale down a collision and still observe a collection of particles that behaves like a drop of fluid.

According to a recent article in Physical Review Letters, CERN and an international team of collaborators were able to generate the substance using oxygen-16 and neon-20. Both are less than a tenth of the weight of a lead atom, which was previously considered one of the lightest elements capable of generating QGP.

“We have pushed the boundary for how small the atomic nuclei can be while still re-creating this primordial matter—what you could call a ‘little big bang.’ We now know more about the fundamental conditions required for matter to transition into this extreme state,” You Zhou, a researcher at the Niels Bohr Institute in the Netherlands and a coauthor of the study, explained in a press release.

The scientists found that, despite the small size of the oxygen and neon nuclei, the collisions produced signals consistent with the behavior they expected to find in QGP. For an instant, the generated matter appeared to expand collectively like a fluid before cooling and reverting to particles.

“Hopefully, this will help us better understand how the plasma behaved during the first moments of the universe—and how it later evolved into the forms of matter that everything around us is made of,” Zhou added.

This story originally appeared on WIRED en Español and has been translated from Spanish.

The big news today is, of course, the successful launch and deployment of NASA’s Nancy Grace Roman Space Telescope earlier this morning. The space agency’s latest observatory lifted off at 7:26 AM Eastern from Launch Complex 39A at Kennedy Space Center aboard a SpaceX Falcon Heavy, and by 8:00 AM it was separated from the rocket’s upper stage and flying on its own.

While the sound and fury of launch is exciting, it’s just the beginning of the journey for Roman. It will take several months for the spacecraft to complete its roughly 1.5 million-kilometer trek out to Earth’s second Lagrange point (L2), where it will set up shop near — in cosmic terms, anyway — the James Webb Space Telescope (JWST). Along the way, it will switch on and test various systems and components, with its primary 300 megapixel infrared camera scheduled to power up in three weeks or so.

There’s a lot to cover about the Roman Space Telescope. Built from spy satellite spare parts donated by the National Reconnaissance Office and featuring a field of view 100 times greater than that of Hubble, its launch is widely considered to be one of the most important scientific milestones of the decade. We’ll be bringing you more about the past, present, and future of this flagship mission as it progresses.

From real space missions to virtual ones, this week the developers of EVE Online announced that 2.4 million lines of code that keep the massively multiplayer online role-playing game running would finally be making the switch to Python 3. Given the immense complexity of the codebase, it’s been stuck at Python 2.7 since their last overhaul back in 2010, a situation which has become increasingly difficult to manage as time goes on.

The announcement goes into a surprising amount of depth about the state of Python in EVE. We imagine most players couldn’t care less, but naturally the developers have strong feelings about the situation and perhaps thought it would benefit others in a similar situation to get their thoughts out there.

While there’s certainly an argument to be made that the only justification they really need for making the migration is that 2.7 hit end-of-life back in 2020, the developers explain that the more immediate problem for them was that various tools and libraries they wanted to use were no longer compatible with the Python 2.x series. They also point out hopes that speed improvements made in the latest version of Python will eventually translate into better game performance down the road.

In more terrestrial news, this week the necessary regulatory amendments were passed to make plug-in solar systems legal in the United Kingdom. Assuming the wiring meets the necessary requirements, consumers can pick up the hardware and install it themselves without involving an electrician, although they may still need to contend with landlords and local ordinances that may limit their ability to physically mount the panels. The rules as they stand now allow each residence to have four panels with a total combined output rating of no more than 2,000 watts, although critically, the system is only allowed to generate a maximum of 800 watts at the inverter. As the government and consumers get more comfortable with plug-in solar systems, these numbers will likely increase over time.

Solar isn’t the only area where DIY approaches are moving into the mainstream. This week, Citrix pitched a “different approach to endpoint resiliency”: an isolated Linux-based operating system called UniconOS that users can boot into should the computer’s primary Windows installation become compromised or otherwise inoperable. The idea is that an independent, read-only backup OS kept on its own partition will reduce downtime, since the computer can still be used while IT figures out what the hell happened.

This solution will sound suspiciously familiar to anyone who’s booted a live Linux system from CD/DVD/USB in the last few decades. Try not to keep yourself up all night wondering why you never pitched the idea to some hungry venture capitalists in exchange for a yacht in the Bahamas.

Finally, on the theme of new technology embracing the old ways, we bring you Defrag98, a web reincarnation of Microsoft’s dial-up era Disk Defragmenter tool. While it won’t actually improve the performance of your modern solid-state drive, you may find your own mood boosted by the wave of nostalgia when you see — and hear — the classic tool go to work.

That’s right, not only do the blocks dutifully flip from red to blue just like you remember, but all the while you’ll be treated to the unmistakable whirs and clicks of a spinning hard drive circa the turn of the millennium. Never forget what they took from us.

See something interesting that you think would be a good fit for our weekly Links column? Drop us a line; we’d love to hear about it.

If you have been paying close attention to artificial intelligence, you may have noticed your worries about job security growing instead of shrinking. According to the AI Monitor 2026 from TU Darmstadt, the more you understand AI, the more you worry it’ll take your job, not less. Based on a representative survey of over 2,000 people in Germany, the study breaks the long-held belief that fear comes from ignorance.

The more you know, the more you worry

Almost half of people with very strong AI knowledge, around 43%, expect AI could soon take over their work. That number drops sharply among people who understand AI less. Professor Peter Buxmann, who led the study, called this both remarkable and worrying, since most debates assume concerns stem from a lack of information.

His data says otherwise. People who understand AI best also see its disruptive potential most clearly. What matters even more than industry is whether your actual output can be digitized. If your job produces text, analysis, reports, code, or presentations, you’re far more exposed than someone whose work depends on physical presence. However, daily AI use at most workplaces still remains fairly limited for now.

A knowledge gap making things worse

Despite how disruptive AI already feels to many, most people haven’t been trained to handle it. Just 15% of respondents have participated in any kind of AI training so far, revealing a real skills gap in AI usage. The concerns people actually have are also surprisingly grounded.

AI hallucinations and data privacy top the list of worries, while only four in ten respondents consider a future superintelligence even likely. Younger workers reported more job anxiety than older ones, and overall concern about job security climbed across every occupational group compared to last year.

The fix isn’t telling people to calm down

Buxmann pushed back against simple reassurance as a fix. People who lose jobs lose more than income, they lose social connection and a sense of purpose too. Rather than dismissing these concerns, he argues they should be addressed through better training and honest conversations about how work is changing.

While generative AI will create new kinds of jobs, he says, it’s unlikely to replace disappearing roles on a one-to-one basis. So the real takeaway for policymakers and employers is that preparation matters more than reassurance ever will.

Welcome back to TechCrunch Mobility, your hub for the future of transportation and now, more than ever, the role AI is playing in it. To get this in your inbox, sign up here for free — just click TechCrunch Mobility!

Proponents of autonomous vehicle technology have long argued that robotaxis and other self-driving vehicles will reduce crash incidents and make roads safer. And there is some evidence of that. But that doesn’t mean there hasn’t been a cost to those who are working (or have worked) for the companies developing that tech. 

Sean O’Kane, senior reporter, special projects, dug into data submitted to the Occupational Safety and Health Administration and found that test drivers for Waymo and Zoox sustained more than two dozen injuries in 2024 and 2025 from hard braking or other sudden movements made by the autonomous vehicles. In some cases, these workers were sidelined for months after sustaining injuries like whiplash when the AVs stopped abruptly, and hard.

Check out his full story, which includes interviews with former and current workers. 

There is one important item to consider. You might ask yourself, well what about the other AV developers? Surely, this isn’t a problem isolated to Waymo and Zoox? And you’re probably right. It’s likely that test drivers for other AV developers are also getting injured, but the companies they work for may be exempt from OSHA’s reporting requirements.

Got a tip for us on this story or others? Email Kirsten Korosec at kirsten.korosec@techcrunch.com or my Signal at kkorosec.07, Sean O’Kane at sean.okane@techcrunch.com.

Deals!

I’ve been writing about Gatik, the autonomous vehicle startup known for its self-driving box trucks, since 2019 when it came out of “stealth.” At the time, I wasn’t sure if the startup would survive. The hype cycle was already chewing up and spitting out AV startups and no one seemed close to commercializing their tech. 

Gatik has not only survived, but it has also crossed over from R&D lab to commercial operator, albeit at a small-scale compared with traditional human-driven delivery companies. (The company’s box trucks are driverless and deliver freight from distribution centers to retails stores like Walmart.) And now, with $200 million in fresh funding, it’s pushing to scale. The round was led by Qatar Investment Authority and Koch Disruptive Technologies, with participation from Millennium Management, ARK Invest, and Intact Private Capital. 

This is Gatik’s largest funding to date. But I might argue that its multiyear commercial agreement with PepsiCo, which was signed in June and is part of $600 million in contracted revenue, is the bigger deal here. 

Other deals that got my attention …

Airbound, an Indian startup building autonomous drones, raised $37 million in a Series A round led by Greenoaks with participation from DoorDash, Lachy Groom, Lightspeed, and Humba Ventures.

Mubadala Capital, the alternative asset management arm of Mubadala Investment Company, agreed to acquire a majority equity interest in Arrive Logistics, a truckload brokerage based in Austin.  

Regent Craft, a Rhode Island electric startup developing and manufacturing electric seagliders, raised $120 million in a Series B round co-led by Mare Liberum and AE Ventures. (The seaglider is a class of vehicle called a wing-in-ground effect vehicle, or WIG.) Regent also secured about $120 million debt capital provided by Erebor Bank, which was launched by Anduril Industries founder Palmer Luckey. One note about Regent, a company I have followed for a while: The startup, which recently completed its 255,000-square-foot seaglider factory, is clearly pushing deeper into the defense sector, probably because there is money and partnerships to be had.  

Vista Global Holding, a private aviation group based in UAE, is considering a European IPO that could raise more than a $1 billion, Bloomberg reported.

Notable reads and other tidbits

According to a recent YouGov survey, more Americans oppose police license plate cameras than support them. Do you? Shoot me an email and share your opinion. 

Any, an electric two-wheeler startup out of Belgium, is placing a bet on cargo space. 

General Motors is facing increased scrutiny from U.S. safety regulators after hundreds of incidents, more than 20 crashes or fires, and at least six injuries involving brake problems in its EVs.

Ford has hired Dave Carroll as president of the company’s energy business. Carroll, who previously worked at ENGIE North America, will succeed long-time executive Lisa Drake.

Rivian CFO Claire McDonough is resigning from her position and will leave at the end of October. Her tenure at Rivian came during a tough, and exciting (ahem IPO), period for the company. And her departure comes at another critical moment for Rivian as it takes on some of its biggest projects to date, including the robotaxi deal with Uber and scaling production and sales of its R2 SUV. 

Uber is launching a new live video streaming feature that will allow parents to keep track of their children during rides. 

Waymo shared 10 lessons it has learned after its vehicles had driven more than 200 million autonomous miles. The first lesson — that multimodal sensors are indispensable — received the most attention because it is in direct opposition to Tesla’s camera-only approach. But there were other lessons that got my attention, including Waymo’s promotion of vision language models (an area that is starting to get a lot of attention). 

Meanwhile, Waymo is making more inroads overseas. The company announced that it plans to launch in Munich, Germany. 

One more thing …

One important clarification on what I wrote about last week. You might recall that Waymo shared information about its custom silicon chip — specifically a 5 nm ASIC chip that is designed to handle the massive influx of raw data before it reaches the core “brain” of the self-driving system. Waymo stated in its blog post:

“While these ASICs alone deliver over 1,000 TOPS of ML performance dedicated to front-end processing and ML models, we optimize across the full stack to maximize achieved performance, especially in the low-batch regimes we often operate.”

That line led me, and others, to believe that its one chip can deliver 1,000 TOPs (trillions of operations per second) of computing performance. I compared it to Nvidia’s Drive AGX Thor automotive processor, noting it was about the same performance. 

But alas, one eagle-eyed reader reached out with some questions, which prompted me to turn to Waymo for official answers. A Waymo spokesperson told me, “It’s for the system, not a single chip.”

That’s an important distinction.

The new Epson LifeStudio Grand UST projector may not ring any bells. But its rebranding hides the fact that it’s little more than a minor update to the Epson EpiqVision Ultra LS650. That was a decent projector when it came out, but not an amazing one.

With the LifeStudio Grand, Epson has made some improvements. The Google TV operating system works well, and there’s an extra HDMI port where its predecessor offered just two. It also doesn’t have whiny fans. But the image is still a little soft and not as dazzlingly colorful as RGB laser projection. The projector’s motion interpolation also needs a lot of work. And Epson’s switch from Yamaha to Bose for sound tuning hasn’t made the audio better.

With competitors improving considerably over the past year, the LifeStudio Grand is in troubled waters. If you can’t see the DLP rainbow effect and don’t mind drawing the curtains from time to time, you can get a much better projector for your money by shopping elsewhere.

Epson LifeStudio Grand review: Price & release date