An anonymous reader quotes a report from CNBC: On Thursday, AMD said it's entered into an agreement to acquire Taalas, a Toronto-based startup that makes chips for inference. Taalas' accelerators are customized, or hard-wired for a single AI model, rather than being general purpose. In exchange for that loss of flexibility, Taalas' technology promises a less-expensive chip that it says can produce output for specific models thousands of times faster than a traditional GPU. An AMD representative declined to provide a purchase price for the transaction. Taalas has raised a total of $219 million in venture funding since its 2023 founding.

Taalas' current chip runs a small version of Meta's Llama 3.1 model, though the company is working on chips for bigger and more advanced models. It's manufactured using an older Taiwan Semiconductor Manufacturing Co. process, and uses speedy SRAM memory on the chip itself. Taalas CEO Ljubisa Bajic says on the startup's website that the company "developed a platform for transforming any AI model into custom silicon." "From the moment a previously unseen model is received, it can be realized in hardware in only two months," Bajic wrote.

The latest NVMe specifications add SSD-level virtualization that can simplify live VM migration by preserving storage identities across servers. The updates also introduce support for post-quantum cryptography, controller-based rate limiting, voltage monitoring, and factory-reset capabilities. The Register reports: Announced by the NVM Express consortium, all 11 of the suite of NVMe specs have been updated with new features and engineering change notices. These represent the next step in the evolution of the standard, it says, which was created as a protocol to support storage devices connected to a system's PCIe bus. Perhaps the most significant new capability is PCIe Exported NVM Subsystem Migration. This extends existing NVMe virtualization to locally-attached PCIe SSDs. It does this by abstracting the physical drives into host-defined virtualized NVM subsystems, to allow for virtual machine (VM) mobility without storage reconfiguration.

When a VM moves from one server to another, its storage also needs to move with it in a way that's non-disruptive to any applications running in that VM. "With NVM Subsystem Migration, NVMe SSDs can present exported NVM subsystems that hide the complexity of the underlying hardware," says Mike Allison, a senior director at SSD maker Samsung and NVM Express board member. "Instead of interacting with physical controllers and namespaces, the host only sees exported controllers and namespaces. This creates a clean separation between what the VM sees and what's happening under the hood," Allison explains on an NVM Express blog.

"One of the key innovations here is providing the host with control over exported identifiers. During migration, those identifiers can be carried over exactly from the source to the destination. That consistency is crucial: even if the underlying hardware uses different internal IDs, the VM sees no change and can pick up right where it left off with no storage reconfiguration required," he says. In effect, the NVMe layer now enables the virtual machine manager (VMM) to rely on virtualization built into the SSD. The flash drive itself exposes logical, virtualized storage constructs, offloading this complexity from the VMM. You can learn more about the updated NVMe specifications here.

An anonymous reader quotes a report from Ars Technica: Nowhere is the US data center boom bigger than in Texas. But less than a year after declaring Texas the "epicenter of AI development," Governor Greg Abbott has declared a moratorium on all new power grid connections for data centers -- at least until developers provide more information about their projects' potential impacts on the grid and communities. The Republican governor directed regulators in an August 3 announcement at the Public Utility Commission of Texas and the grid operators at the Electric Reliability Council of Texas (ERCOT) to perform a "comprehensive verification and audit of all data centers advancing through ERCOT's interconnection process." As an independent system operator, ERCOT oversees a power grid that operates separately from the rest of the United States and provides services to most of Texas.

Texas has aggressively courted data center development with its availability of cheap land and relatively abundant energy resources, along with offering state incentives, like tax breaks and fewer regulations. That puts the state on track to surpass Virginia in becoming the largest US data center market. But the recent AI boom and the accompanying frenzy of data center development threaten to overwhelm the Texas grid on paper, despite the state leading the country in adding new power generation. The ERCOT interconnection queue currently includes more than 1,800 projects representing over 474 gigawatts' worth of requests to connect to the Texas grid -- more than five times Texas' record peak electricity demand -- and about 90 percent of those power connection requests come from data centers.

"That unprecedented load growth could endanger the reliability and stability of the Texas electric grid," according to the statement from Abbott's office. Many of those data center projects may never materialize for various reasons. But ERCOT has still forecast that data center demand and other factors could drive statewide electricity demand to double the current demand record by 2032, according to The Texas Tribune. The review will examine each project's projected electricity consumption, reliance on the grid and state incentives, ownership, and water use. It will also assess measures intended to "reduce impacts on neighboring property owners and communities," including noise controls, lighting, traffic improvements, setbacks, and emergency planning.

Ars Technica notes that the directive does not address air pollution or greenhouse-gas emissions and does not apply to data centers generating their own power on-site.

BrianFagioli writes: Sandisk and SK hynix have published the first open technical specification for High Bandwidth Flash (HBF) through the Open Compute Project. HBF is designed to create a new memory tier between High Bandwidth Memory and traditional SSD storage, giving AI inference systems more near compute capacity without relying entirely on expensive HBM. The specification supports capacities up to 512GB using 8-high and 16-high NAND stacks, with bandwidth tiers ranging from about 0.4TB per second to 3.0TB per second. It also uses the UCIe chiplet interface, which could make HBF easier to integrate with CPUs, GPUs, and other accelerators. Google and Tenstorrent participated in the standardization effort, but commercial products and independent benchmarks are still missing.

Apple's stock "fell just shy of 10% on Friday," reports Yahoo Finance, "after CEO Tim Cook warned about the impact of the global memory shortage on the company's business." During Apple's third quarter earnings call, Cook said the company paid significantly more for memory in the quarter and expects that to further increase in the current period. And while Apple is able to offset some of that price jump, it won't be able to tackle it all... The CEO said that iPhone and Mac sales outpaced Apple's own expectations and that a lack of flexibility in the supply chain is making it more difficult to keep up with demand.
Yahoo Finance cited an investment analyst who predicts overall gross margins for Apple's iPhone could drop from 38% to 34.5%. But another analyst sees a scenario where Apple "raises iPhone prices, unit growth will slow, and as unit growth slows, so will user growth, which we think ultimately will slow Services growth." (Still, Yahoo Finance predicts Apple's new leasing program "could help address those concerns.")

Apple has another controversial option, according to the blog 9to5Mac: Bloomberg reports that US senators from both sides of the aisle are urging Apple CEO Tim Cook to commit by August 21 to not using memory chips from Chinese suppliers CXMT and YMTC....

Apple is not required to obtain U.S. government approval to purchase chips from the companies, but doing so without the administration's support could expose it to significant political ramifications. Which is why, in an interview with The Wall Street Journal ahead of Apple's recent price hikes, Tim Cook said that "everything needs to be on the table," adding that "we should look at all supply...." More recently, the Journal reported that Apple's use of Chinese memory chips could extend beyond China, with the company seeking the administration's blessing to use components from CXMT and YMTC in products sold elsewhere outside the US. The report also detailed Micron's efforts to persuade the administration to reject Apple's request, arguing that allowing Chinese suppliers into Apple's supply chain could undermine domestic memory production...

[Bloomberg's reported that U.S. lawmakers warned] other companies could follow Apple's lead, potentially undermining domestic memory production and planned investments in states such as Indiana, Idaho, New York, and Virginia... [T]he senators sought details about any information Apple has shared with CXMT during the component qualification process, noting that the transfer of controlled technical information to advanced chipmaking facilities in China may require a Commerce Department license.
The blog MacRumors notes that Apple has already increased prices for Macs and iPads in June because of surging memory prices. Apple CEO Tim Cook said Thursday "we did it because we're in what I would characterize as a 100-year flood on memory pricing with exponential increases in memory prices." Cook did not comment on whether Apple plans to raise iPhone prices when the iPhone 18 Pro models and first foldable launch this September, but multiple analysts believe prices will go up. Cook said Apple is expecting to pay higher memory costs in the September quarter, though Apple will be able to partially offset it with lower costs on some non-memory components and a stockpile of inventory.
Sky News points out that "It was Tim Cook's final earnings appearance before his retirement after 15 years at the helm of the company." Apple, which recently topped Nvidia as the most valuable listed company, has been largely spared the volatility in share price seen by chipmakers and big spenders on artificial intelligence. Apple continues to generate cash without the huge investment spending that its Wall St peers are dealing with "and that showed across most parts of the operation," said Thomas Monteiro, an analyst at Investing.com.

Google DeepMind's Gemini Robotics 2 combines vision, language, and action models to control multiple types of robots, including humanoids performing tasks such as organizing shelves, tying bags, and replacing lightbulbs. "It's another milestone in our path towards really getting towards what we call like physical AGI, which means we get a robot to do anything that a human can," Carolina Parada, head of robotics at Google DeepMind, tells WIRED. From the report: Gemini Robotics 2 combines several different AI models into a single system. Taken together, they allow a robot to make sense of its surroundings and how to act in it. A vision language model (VLM), which understands images and video, can communicate with humans and reason how to perform different tasks. Two vision language action (VLA) models, trained to understand how to move in physical space, control the robot's full-body movement as well as the movements of grippers or hands.

In video demonstrations shared ahead of the release, the company showed several different robots performing complex tasks autonomously using the amalgamated model. In one demo, Apptronik's Apollo 2 robot used hands from a company called Sharpa to tidy shelves. Google DeepMind trained the model to perform these tasks using a mix of human teleoperation, video examples, and simulations -- it's not yet possible for AI models to perform a wide range of complex tasks without specific training.

[...] Parada says Google takes a multi-layered approach to safety, with guardrails applied on each model layer. It's also introducing ASIMOV-Agentic, a new benchmark for measuring the safety of various AI systems collaborating to control a robot. The benchmark detects whether a command will result in harmful or uncertain outcome.

The FCC's ban on Chinese-made robots extends well beyond humanoids to quadrupeds, research platforms, and many robot vacuums from allied countries. Supporters call it a major boost for domestic robotics, but critics warn that cutting researchers and startups off from affordable foreign hardware could slow U.S. innovation instead. Ars Technica's Jeremy Hsu examines who stands to gain and who stands to lose from the prohibition: Such an import ban would apply to some of the most affordable robots primarily produced by Chinese companies, including Unitree's humanoid robots that are used by robotics labs and researchers for tasks such as experimental robot surgeries. US consumers would also likely lose access to the newest robot vacuum cleaners that are mainly manufactured by Chinese companies such as Roborock. But the ban also broadly applies to foreign-made robots produced by countries nominally allied to the United States, including Japan, South Korea, and Germany. [...]

The ban on foreign-made robots could theoretically encourage more US and foreign companies to set up manufacturing facilities in the United States. There are already multiple companies racing to scale up production of humanoid robots in US factories, including Agility Robotics, 1X Technologies, and Figure AI. Tesla has been attempting to shift production away from older electric vehicle models and toward its Optimus humanoid robot. Boston Dynamics has already been making its Atlas humanoid robot, along with its four-legged Spot robot and wheeled Stretch robot, at its main facility in Waltham, Massachusetts. The US robotics company is also planning to massively scale up manufacturing of the Atlas robot under South Korea's Hyundai Motor Company, which gained full ownership of Boston Dynamics in July 2026.

"This is one of the strongest technology-security actions in modern US history," wrote Evan Beard, CEO of Standard Bots, in a LinkedIn post. "The message is unambiguous: robotics is a technology America must lead and own -- and foreign-subsidized robots will not be allowed to unfairly dominate US robotics as they did solar." Similar praise came from Rush Doshi, director of the Initiative on China Strategy at the Council on Foreign Relations, who, in a social media post, described the FCC decision as "one of the most significant actions taken so far in support of the US robotics ecosystem."

However, several robotics researchers and analysts interviewed by The Robot Report expressed skepticism about any potential boost to US competitiveness in robotics. Some even warned that the ban could prove counterproductive for US robotics efforts to develop humanoid robots. "In the near term, the measure could slow US physical AI innovation by cutting startups and researchers off from future low-cost Chinese platforms before comparable Western alternatives exist," said Georg Stieler, a global robotics advisor and managing director for Asia at Stieler Technology & Market Advisory, in an interview with The Robot Report.

US domestic production of robots lags behind China in terms of mass manufacturing at lower cost, said Rueben Scriven, a senior analyst at Interact Analysis. "This announcement is more likely to inhibit the US humanoid robotics industry, as the presence of low-cost Chinese humanoid robots has been helping educate the US market through promotional and entertainment use cases -- an effect this policy risks undermining," Scriven told The Robot Report. The report notes that previous FCC bans have done little to help create competitive U.S. alternatives, with restrictions on Chinese drones instead prompting companies to sell barely disguised versions of DJI technology.

NextEra and Brookfield plan to invest more than $100 billion to transform a former uranium-enrichment site in Kentucky into a data center campus with a generating plant. "The privately funded effort will include construction of 2 gigawatts of natural gas-fired power at or near the site in Paducah in western Kentucky, and as much as 2.6 gigawatts of battery storage capacity," reports Bloomberg. " For context, a single gigawatt of capacity is roughly the output of a traditional nuclear power plant and can power about 750,000 homes at any given moment." From the report: Brookfield will develop and operate the 1.8 GW data center campus, which will occupy portions of the sprawling 3,556-acre Energy Department site once used to produce weapons-grade uranium and fuel for nuclear reactors. Operations using the now obsolete enrichment technology known as gaseous diffusion stopped in 2013 and the site is now subject to cleanup efforts. Construction is expected to be completed in 2031, the Energy Department said. "The U.S. government is leveraging its assets -- like our federal lands -- to add power generation, create jobs, and ensure the United States wins the AI race," Energy Secretary Chris Wright said in a statement. The project is expected to create 8,000 construction jobs and 600 permanent positions, the department said.

The Trump administration has banned newly authorized foreign-made humanoid and four-legged robots, along with power inverters, citing "unacceptable risks" to the country's national security. FCC chairman Brendan Carr said the agency was doing its part "to secure America's critical supply chains." The BBC reports: The FCC has added the items to its Covered List -- a register of goods and services that are deemed a risk to US national security. The ban applies to new foreign-produced advanced robotic devices and power inverters and does not prevent the sale or import of any existing models that had been previously authorized by the FCC.

The FCC cited concerns that the use of foreign-made inverters could allow overseas firms to turn them off, steal data, facilitate remote access and surveillance by "foreign government actors, or be otherwise exploited through a cyberattack." It added that the use of robots made outside the US could allow "malign actors to surveil Americans, enhance the capabilities of foreign intelligence services, or to remotely commandeer the robots."

The review embargo has lifted for the new Framework Laptop 13 Pro, and the consensus across the board is that it is a massive leap forward in terms of build quality and battery life. The main issue reviewers complained about is the sky-high price, with the higher-end model jumping dramatically from $2,100 up to $2,900 due to the memory shortage crisis. (Some note that the price "nearly doubled" overnight while they were in the middle of testing.)

In his video review, Marques Brownlee says, "This is their best build yet. They're finally doing what people have been asking for: a premium... modular... laptop." ZDNET agrees that this device "represents a new approach for Framework and a maturation of the brand's catalog," noting that the new construction is "sleek and airtight, with no gaps, spaces, or evidence that was even put together by your hands at all." Tom's Hardware echoes this enthusiasm, declaring that "The Framework Laptop 13 Pro's sturdy design, haptic trackpad, and bigger battery feel like they should have been there all along".

Battery life, which has historically been a weak point for Framework, is now a standout feature. Ars Technica points out that the combination of Intel's efficient Panther Lake chips and a new 74-watt-hour battery is "finally long enough to decisively eliminate 'mediocre-to-poor battery life' as the laptop's biggest downside." They also said it's "Framework's nicest-looking, nicest-feeling, most polished laptop design."

For Linux users, the machine appears to deliver on its promises. Phoronix says the device "is designed with great Linux compatibility in mind," offering a seamless out-of-the-box experience for distributions like Ubuntu and Fedora.

As mentioned above, the overarching complaint is the extreme cost. Ars Technica notes that Framework unfortunately "switched to an exotic new upgradeable LPDDR5X RAM format just in time for those modules to become astronomically expensive." Still, Marques Brownlee summarizes the long-term value proposition perfectly: while it might be painfully expensive on day one, "the longer you keep this laptop, the more it feels worth it" because you can easily repair and upgrade it over time. Compared to Apple's flagship, the Framework is "80% of the quality, way more modular," he says.

According to The Information (paywalled), a state-backed Shanghai company has begun mass-producing China's first homegrown immersion DUV lithography machines, with about five units expected this year for SMIC, Hua Hong, and CXMT. Roughly 20 more units are expected in 2027. Tom's Hardware reports: The Information didn't name the manufacturer, but its sources described the operation as having pulled DUV development teams from several Chinese companies, one of them the state-backed startup Shanghai Yuliangsheng Technology. SMIC has been testing a Yuliangsheng immersion tool since September 2025. Most components in the new systems are domestic, though some critical parts still come from Japan, and delays at local suppliers have held back output this year.

[...] China accounts for around 20% of ASML's net sales this year, down from 33% in 2025, driven mainly by mainstream logic demand... Independent analysis from the AI Futures Project in June put commercial-scale Chinese immersion DUV in the mid-2030s, with ASML holding 98.7% of the immersion market. Qualifying the new machines for production lines could take many months, and they trail ASML's tools on performance and build quality. China's domestic EUV effort, which Reuters first reported as a working prototype in December, remains years away.

From the site tech-news site XDA Developers: The consumer tech economy needs more DRAM manufacturers, and it needs them more urgently than perhaps at any point in its history. Anyone keeping track of the pricing trends will know that memory prices have climbed into the stratosphere on the back of AI infrastructure's relentless demand, and every consumer device that depends on RAM has traveled from inflated to unaffordable, quarter after quarter, with almost no relief in sight. Which is exactly why a story out of Shenzhen is worth paying attention to, even in its current rumor phase. According to the details circulated by various semiconductor analysts on X, the Chinese government, DRAM maker SwaySure and Chinese tech giant Huawei have partnered to set up a 12-inch memory production plant. If the reports hold up, it could change memory economics for years to come, but not necessarily in the way you'd hope...

SwaySure itself is a known player. The company was registered in Shenzhen in March 2022, is reportedly backed by municipal government, and appears alongside Huawei on the US Commerce Department's Entity List, suggesting Washington has realized the connection between the two entities. Although, that's also where the story becomes a little less straightforward. None of these claims have been independently confirmed, and many of the technical details circulating online trace back to a single X account. Huawei has previously denied operating the wider network of state-backed fabs that analysts have reported. Even the widely quoted figure of 140,000 wafers per month isn't new, and public records indicate SwaySure's production line is already operational. What gives the report weight, then, is the fact that Huawei has spent years building this kind of domestic semiconductor ecosystem through partners that appear to operate independently.

The rationale that analysts on X keep revisiting has little to do with consumer memory. Huawei's Ascend AI accelerators require high-bandwidth memory (HBM), and US export controls have all but cut Chinese entities from procuring HBM from Samsung or SK Hynix... Owning the supply outright solves a problem that purchasing power can't, which is why the DRAM push is read as a bid to secure captive HBM for Huawei's own silicon rather than to flood the market with cheaply manufactured DDR5... The catch, however, is the fact that most of this new and rejuvenated capacity is expected to be absorbed by China's own domestic demand, rather than to relieve the global shortage... This development does not seem to be an isolated incident, given the state of DRAM economics. CXMT, China's largest manufacturer of DRAM, is already running close to 265,000 wafer starts per month as of now as per analyst estimates, and is projected to hit 350,000 by the end of the year, bringing it within the touching distance of Micron's estimated 385,000....

[T]he oligopoly seems to be facing the first credible challenge to its capacity discipline in decades, and it seems to be coming from a country building fabs much faster than the competition could hope to match. Now, whether it tangibly impacts the consumer market economics in the time to come may be a story in its own right in a few years, but for now, for what it's worth, it's a new direction, and much of it feeds optimism.

A Canadian lab has run a chemical process on real spent nuclear fuel "and pulled out 90% of the long-lived danger in 24 hours, the part that forces a burial site to last 100,000 years," notes the blog Autonocio, "with the leftovers meant to fuel a reactor." The standard plan for spent nuclear fuel is to wait it out. You pull the used bundles from a reactor, sit them in a pool of water for seven to ten years while the heat and radiation come down, seal them in concrete casks, and look for somewhere deep and geologically dull to leave them for the next hundred thousand years. Canada has been hunting for that burial site since the 1980s and still doesn't have one in the ground.

A company in Saint John, New Brunswick thinks most of what makes that waste dangerous never needed to go in the ground at all. Moltex Energy Canada says a chemical process it calls WATSS can strip 90% of the long-lived material out of used Canada Deuterium Uranium [CANDU] fuel in 24 hours, and that the concentrated leftovers become fuel for a reactor it wants to build on the same site. The recovery step isn't a slide in a pitch deck anymore. In 2025, World Nuclear News reported that Canadian Nuclear Laboratories ran the process on real used fuel from a commercial Canadian reactor and confirmed the 90% figure.

None of it is generating power yet. What exists is a validated chemical step and a reactor design waiting in line at a regulator. The rest is a 2030s problem.
"The chemistry has a lab result behind it. The reactor does not exist..." the article points out. "Moltex is aiming to have its first WATSS and SSR-W units running at Point Lepreau by the early-to-mid 2030s... Not everyone buys the pitch. Critics have argued the reprocessing creates its own stream of byproducts, that the economics are unproven, and that a single site's stockpile is finite."

But Moltex "isn't alone in trying to burn nuclear waste instead of bury it," the article notes, with Switzerland and Denmark "chasing the same goal a different way." Switzerland's Transmutex drives a subcritical reactor with a particle accelerator, feeding it spent fuel alongside thorium... Denmark's Copenhagen Atomics is building a thorium molten-salt reactor that fits inside a shipping container and runs on the leftovers from conventional plants.
Thanks to long-time Slashdot reader kwelch007 for sharing the article.

Politico reports that in the Virginia area alone there's dozens of data center projects "that altogether need 70 gigawatts of power -- equivalent to 70 nuclear plants" -- currently seeking connection to their grid. But there's also concerns about a Virginia data center powered with natural gas and backup diesel generators: When Virginia's top environmental regulator received an analysis warning of data center pollution, it took him less than five minutes to forward it to seven people on his staff... Internal emails obtained by POLITICO through a public records request paint a portrait of an agency that moved quickly to defend the only data center in Virginia that is powering itself -- in what former officials and environmental health advocates described as an unusual effort to shape the debate around an industry whose global epicenter is in the state. The agency's pushback focused on the report's findings that the facility in Loudoun County could release harmful amounts of air pollution through eight bus-sized natural gas turbines -- as allowed by permits granted by the Department of Environmental Quality [DEQ] itself... The request comes as data centers are devising new ways to produce their own power as a way to temper growing public discord over rising electricity prices amid the AI construction boom.

Virginia has more than 600 data centers. The report [commissioned by Virginia's 54-year-old environmental nonprofit Piedmont Environmental Council] raised concerns that the data center's on-site power system, which also includes dozens of backup diesel generators, could cause tens of millions of dollars in health damages to people living around the facility, owned by Vantage Data Centers... [The analysis also argued the pollution could lead to 3.4 to 6.5 premature deaths annually.] People living near the Vantage facility in Sterling say its natural gas turbines produce constant noise and air pollution. The permits for Vantage issued by DEQ in 2023 allow cumulative emissions of seven different pollutants, including 95 tons per year of nitrogen oxides and more than 56 tons of soot. Both contribute to asthma and heart attacks.

Those figures are in line with other state permits for minor sources of pollution, but DEQ's permits for Vantage have sparked local concerns because of the facility's location in a residential area... [Virginia environmental quality officials] raised questions about whether the report wrongly described the facility's potential air pollution as dangerous, when soot levels in Loudoun County are deemed acceptable by America's Environmental Protection Agency (EPA). Health experts, including former EPA air quality official Michael Korber, told POLITICO that soot pollution can negatively affect human health even at EPA-approved levels. The World Health Organization's standard for ambient soot pollution is nearly half of what EPA suggests is safe.
Some current and former staffers on Virginia's Department of Environmental Quality believe it's inappropriate for the agency to issue aggressive statements on the healthfulness of the data centers, with one former department leader saying Virginia's DEQ "is not the health department."

Ars Technica reports: Hyundai Motor Company's plan to put humanoid robots to work by 2028 is not part of current negotiations with striking South Korean autoworkers, according to the company.

The automaker is disputing news reports that partial labor strikes by the Hyundai Motor union at the world's largest automotive plant in South Korea were spurred by concerns about the company's planned deployment of humanoid robots in the United States starting in 2028. [The planned robots are built by Boston Dynamics, now a wholly-owned subsidiary of Hyundai.] A Hyundai statement shared with Ars describes the union's demands as focusing on compensation-related issues such as wage increases, bonuses, and an extension of workers' retirement age. "Potential deployment of robots in Korean production facilities is not part of the current labor-management discussions," according to the Hyundai statement...

Hyundai emphasized that its current plan only covers the initial deployment of the Atlas humanoid robot at Metaplant America, an electric vehicle factory near Savannah, Georgia, starting in 2028. "Decisions about future Atlas deployment at other facilities will be made thoughtfully, in accordance with local operational needs, and in dialogue with the employees and workforce representatives at those sites," the company stated. However, The Wall Street Journal described the Hyundai Motor union as making "unprecedented demands seeking to enshrine job protections in the era of robots and AI," and characterized certain compensation demands as hedging against potential reductions in work hours caused by AI adoption and robotic automation. "Remember that without labor-management agreement, not a single robot using new technology will be allowed to enter the workplace," the union told Hyundai in an internal letter reported by Reuters.

Hyundai management and the labor union are currently reviewing a proposed wage reform that would "provide factory workers with greater income stability even after the carmaker's planned deployment of humanoid robots," The Korea Times reported. But experts cautioned that the wage overhaul, which would convert hourly pay for overtime and night-shift allowances into fixed wages, could come at the expense of company productivity.