EV Battle Shifts to Batteries and Rapid Charging

Bernstein estimates Elon Musk’s TeraFab ambitions to supply 1 TW/year of AI compute would require up to 358 modern fabs and cost roughly $5 trillion—about 70% of the current U.S. budget. The calculation converts rack-level power targets into required chip and wafer volumes (Rubin/Rubin Ultra GPUs, Vera CPUs, and HBM4E memory), then into wafer starts and fab counts. Bernstein’s back-of-envelope assumes large die sizes, certain yields, and fab cost/capacity figures; Tom’s Hardware notes some assumptions likely over- or understate real-world fab throughput and costs, but the headline multitrillion-dollar scale remains plausible. The analysis highlights the massive industrial, capital, and supply-chain hurdles for any megascale AI chip strategy.

Elon Musk said SpaceX AI and Tesla will continue buying Nvidia chips at scale, signaling sustained demand from two of the largest AI and autonomous-vehicle customers. Musk made the comments amid industry chip shortages and escalating competition in AI accelerators; his endorsement underscores Nvidia's dominant position for neural-network training and inference. The remark matters because continued large-volume purchases by Tesla and SpaceX support Nvidia's data-center and edge chip business, influence supply chains, and shape competitors’ strategies in custom silicon and AI hardware. It also highlights ongoing collaboration between leading AI product companies and chip vendors as AI workloads expand across automotive, robotics, and aerospace applications.

Elon Musk announced a $25 billion 'Terafab' semiconductor factory aimed at producing high-performance AI chips and supporting SpaceX and other ventures. The plan envisions large-scale chip fabrication capacity to reduce reliance on external foundries, accelerate AI model training and inference, and secure supply for space and electric-vehicle systems. Musk framed Terafab as strategic for performance, cost control and national competitiveness in semiconductors. If realized, the facility would reshape supply chains, intensify competition with established chipmakers and foundries, and influence AI hardware availability for startups and hyperscalers. Key players are Elon Musk, his companies (notably SpaceX and possibly Tesla), and incumbent semiconductor firms and foundries that may be affected.

Elon Musk announced a plan to create Terafab — a chip fabrication effort across Tesla, SpaceX and xAI — aiming to produce a terawatt of compute per year, roughly 50x current global annual output. He claims an Austin advanced fab that can make “any kind of chip” plus lithography masks, a recursive rapid-production process, and “new physics” to push performance. Musk outlined two chips: one for inference in humanoid robots (he projects up to billions of units) and one for satellites, with ambitions to orbit terawatts of compute via megawatt-scale compute satellites. The plan depends on massive launch cadence and materials availability (helium shortages noted) and lacks concrete timelines or supply-chain feasibility. It matters because it signals audacious vertical integration across semiconductors, space launch, and AI hardware, but raises practical and logistical doubts.

Tesla and SpaceX announce $25B 'Terafab' chip factory – it reeks of desperation

Tesla and SpaceX announce $25B 'Terafab' chip factory – it reeks of desperation

Elon Musk unveils chip manufacturing plans for SpaceX and Tesla

Bloomberg : Elon Musk announces Terafab, an Austin-based project run by Tesla and SpaceX to manufacture robotics, AI, and space data center chips for Tesla, xAI, and SpaceX — Elon Musk said his Terafab project — a grand plan to eventually manufacture his own chips for robotics, artificial intelligence …

A Guardian investigation details at least five Cybertruck fires since its 2024 launch, including four fatalities and multiple lawsuits alleging design defects. Police and court records describe crashes where batteries reignited, fires exceeded cremation temperatures, and victims were trapped because electric door mechanisms and the truck’s lack of external handles hindered escape and rescue. Families and safety experts argue the Cybertruck’s heavy laminated glass, stainless-steel construction and novel door system exacerbate post-crash fire hazards. Tesla faces scrutiny as plaintiffs contend crashes were survivable but for the vehicle’s design choices; investigators and first responders report difficulty extinguishing battery fires and extricating occupants. The cases highlight questions about EV safety, emergency access and manufacturer liability.

WIRED’s Uncanny Valley podcast episode (Mar. 19, 2026) focuses on three tech storylines: Nvidia’s annual developer conference in San Jose, backlash around Tesla, and Meta’s shifting VR strategy. Hosts Brian Barrett and Zoë Schiffer describe Nvidia’s event as a major AI industry gathering and highlight CEO Jensen Huang’s comments about the company’s outlook, including a claim that Nvidia’s AI-chip revenue opportunity could reach at least $1 trillion through 2027. The episode also discusses Tesla “influencers” distancing themselves from the brand after recent controversies, signaling reputational risk among core supporters. Finally, it covers Meta’s decision to shut down Horizon Worlds on Quest—then reverse course—keeping it on limited support, a move framed as weakening the company’s metaverse ambitions.

The Guardian reports on a series of severe Tesla Cybertruck crashes in which occupants allegedly became trapped and vehicles burned intensely, leaving remains difficult to identify. One detailed case occurred in August 2024 near Baytown, Texas, where sheriff’s deputies encountered a Cybertruck engulfed in flames more than 10 feet high. Police reports say a fire extinguisher was ineffective and firefighters needed a hydrant, but the blaze persisted as the EV’s batteries repeatedly reignited. After the fire was extinguished, officers found human remains inside; paperwork nearby identified the driver as Michael Patrick Sheehan, a 47-year-old nurse practitioner who had owned the truck for three months. Victims’ families cited in the article blame what they describe as faulty design, raising broader EV crash and fire-safety concerns.

A Tesla Model X on Full Self-Driving (FSD) mode crashed into a concrete wall on a familiar Bay Area residential route; the driver—formerly head of Uber’s self-driving division—was injured but her children were unhurt. The incident underscores limits of current Level 2 driver-assist systems: the car logged flawless miles before the failure, then apparently behaved unpredictably during a turn, prompting a last-second human takeover that proved insufficient. The piece highlights the “moral crumple zone” where humans absorb blame for complex automated failures, and details disputes over access to vehicle logs—Tesla controls extensive driver and sensor data, drivers often receive fragments, and courts have been involved to retrieve evidence. The crash raises safety, transparency, and regulatory concerns for consumer autonomous driving tech.

A former head of Uber’s self-driving division recounts a terrifying crash in a Tesla Model X while using Full Self-Driving (FSD), noting the system drove flawlessly for years until a sudden failure left the car hitting a wall. The author and children survived with injuries, but the vehicle’s logs—capturing driver inputs and reaction times—shift culpability to the human in what researcher Madeleine Clare Elish calls the “moral crumple zone.” The piece highlights struggles drivers face obtaining complete telemetry from Tesla after crashes, cites past cases where data access was contested, and stresses that despite the “Full Self-Driving” label, Tesla’s system remains an SAE Level 2 driver-assist requiring human responsibility. It raises accountability and transparency issues for automated driving tech.

Tesla’s online booster community is fraying after a dispute over Full Self-Driving (FSD) transfers tied to Cybertruck purchases exposed tensions between loyal influencers and disaffected customers. Tesla initially said FSD could transfer to new vehicles bought by March 31, but later added a delivery deadline that left many buyers unable to claim the feature—prompting anger on X and public feuds among prominent Tesla accounts. The episode highlights how Elon Musk’s politics, changing product promises, and aggressive hype around FSD strain the fan ecosystem and can push long-time advocates to speak out. The split matters because influencer networks shape consumer sentiment, investor perceptions, and Tesla’s brand narrative in a market where promises about autonomy and pricing directly affect sales and stock value.

Elon Musk has ordered another round of job cuts and management shake-ups at xAI after its coding product and Grok chatbot failed to compete with offerings from OpenAI and Anthropic. Following a $1.25 billion merger with SpaceX, Musk brought in managers and engineers from SpaceX and Tesla to audit xAI, reassigned leaders, and pushed out multiple cofounders including Zihang Dai and Guodong Zhang. The reviews cited poor training data and implementation as reasons the coding tool lagged rivals. Musk has also redeployed Tesla AI staff to revive a “Macrohard” digital agent project and plans to leverage a growing GPU data center and X’s data to boost xAI. Staff say constant upheaval and burnout are harming morale and retention.

xAI and SpaceX have reportedly hired away two leaders from Cursor, according to the headline “xAI/SpaceX Poach Two Cursor Leaders.” No further details are available about the individuals’ names, their roles at Cursor, which organization (xAI or SpaceX) each joined, or the timing and terms of the moves. If confirmed, the hires would matter because leadership departures can affect a fast-growing developer-tools company like Cursor, while strengthening xAI and/or SpaceX with experienced product or engineering management. With only the title provided, it is not possible to verify the scope of the changes, whether the hires were coordinated, or how Cursor plans to backfill the positions.

The Information : xAI hires senior Cursor leaders Andrew Milich and Jason Ginsberg; Elon Musk said he expects xAI to catch up with rivals in coding by “the middle of this year” — Elon Musk's xAI is hiring two senior Cursor leaders as part of an effort to catch up with AI rivals in coding …

Elon Musk revived his “Macrohard” gag while pitching a joint xAI–Tesla project that he says will field fleets of “Digital Optimus” agents capable of “emulating the function of entire companies.” Musk described xAI’s Grok as the orchestration layer that would direct Optimus by processing five seconds of screen video and input events in real time, running on low-cost Tesla AI4 hardware alongside xAI’s Nvidia kit. He framed the effort as unique and real-time, despite a history of ambitious promises from Musk and recent quality and safety issues at Microsoft and Tesla (notably FSD setbacks). The announcement ties into Musk’s wider push for AI-driven software and Tesla’s strategic investments with xAI.

Group14 opens factory to produce battery materials for flash charging EVs

Grace Kay / Business Insider : Sources: xAI's AI agent project Macrohard has stalled as Tesla ramps up its own AI agent project Digital Optimus; Musk says they are a joint xAI-Tesla project — Follow Grace Kay … - XAI's Macrohard project has stalled amid leadership changes and a data project pause.

Germany’s rapid expansion of solar power is easing electricity costs, according to the headline, even as natural gas prices rise. The implied dynamic is that higher solar generation is helping offset more expensive gas-fired power, reducing wholesale power price pressure and potentially lowering costs for consumers and industry. Key elements mentioned are Germany, solar energy growth, power costs, and a jump in gas prices, but no figures, dates, or specific market data are provided in the available information. With only the title to go on, it is not possible to confirm the scale of the solar boom, the magnitude of the gas price increase, or how much power costs have changed, nor which institutions reported the data.

Bloomberg : How “dark factories”, powered by AI and robotics and requiring essentially no workers, are set to upend China's labor market, already stressed by tariffs — Workers who powered a tariff-defying boom tell a grim story of falling wages and vanishing jobs.

Tesla opens its first Megacharger station to Semi customers in California

U.S. automakers are increasingly adopting lidar for advanced driver-assistance and autonomous features after a surge of Chinese lidar suppliers drove down costs and improved performance. Key players include Chinese firms (which scaled production and pushed prices down) and U.S. automakers and Tier 1 suppliers now integrating lidar into vehicle architectures. The shift matters because cheaper, higher-performance lidar can accelerate deployment of safety-critical sensing, change supplier dynamics, and influence regulatory and competitive landscapes. It also raises supply-chain considerations, including dependency on Chinese component makers, potential for reshoring or diversification, and strategic partnerships to secure supply and expertise for production vehicles.

A blog post by a podcaster recounts notes prepared for an interview with Elon Musk and John Collison, recorded before SpaceX’s reported acquisition of xAI, focusing on the idea of orbital “space GPUs” and data centers. The author argues the main motivation for space computing is energy: satellites in sun-synchronous orbit could achieve near-100% solar capacity factor versus roughly 25% on Earth, avoiding batteries and weather. Launch economics are central, citing Falcon 9 at about $2,500/kg and a fully reusable Starship potentially below $100/kg. The post also highlights obstacles: energy is estimated at ~15% of data center total cost while chips are ~70%, and hardware failures are hard to service in orbit. It cites Meta’s Llama 3 paper (one failure every ~3 hours across 16,000 H100s) and Google’s Suncatcher TPU radiation findings, alongside Musk’s view that post-screened GPUs can be reliable.

Elon Musk told Dwarkesh Patel and John Collison that within roughly 36 months the cheapest place to run large-scale AI may be in orbit, arguing terrestrial power limits and permitting make space-based data centers and solar-fed compute economically attractive. In a wide-ranging discussion he also covered Grok and xAI’s alignment and business plans, challenges of scaling GPU-heavy datacenters, serviceability tradeoffs for orbital hardware, high-volume humanoid manufacturing for Optimus, geopolitical competition with China, SpaceX operational lessons, and even DOGE. The interview matters because it sketches a provocative infrastructure strategy for AI at hyperscale, linking energy constraints, regulatory friction, launch economics, and future hardware design—issues central to cloud, AI, and space-tech industries.

Xiaomi founder and CEO Lei Jun said during China’s 2026 Two Sessions that true driverless operation could be achieved within five years in limited, controlled environments, though full open-road, all-scenario autonomy for consumer cars will take longer. Lei announced a planned RMB 200 billion investment targeting chips, AI, and operating systems to build core capabilities and intelligent manufacturing for Xiaomi’s car efforts. He also previewed a new-generation SU7 model and said production ramped after February deliveries of the YU7. Lei urged stronger automotive intelligentization standards, coordinated governance for traffic safety in the smart-car era, and public education on safe assisted driving, noting current systems are not yet at L3/L4 levels.

Li Auto is reportedly preparing to unveil its first two-wheeled robot later this year after nearly a year of secret development under an internal Nexus team led by former Jiuguang Robotics hardware partner He Junpei. The company plans two humanoid lines — a two-wheeled model aimed at factory/manufacturing scenarios and a bipedal robot — and is said to be following a Tesla-like strategy: first iterate in closed, controlled environments before moving toward consumer markets. The project was advanced within Li Auto after internal debate about timing and scope, emphasizing building the robotic body/hardware first rather than early investments in large-scale open-source brain projects.

NIO announced that its 2,000,000th electric drive unit rolled off the line at the company’s Hefei second electric-drive factory; the unit is the rear drive system for the new ES8. The company emphasizes that electric drives are a core part of its 12 full-stack smart EV technologies and are self-developed and manufactured, with its drive systems having been named among the world’s top ten for four consecutive cycles. The ES8’s 900V high-performance flagship motor uses a 925V continuous winding technology and has won domestic awards. The milestone underscores NIO’s vertical integration in powertrain manufacturing amid rising deliveries and expanding charging/swapping infrastructure.

Bloomberg reports that Chinese autonomous driving startup Momenta has secretly filed for an IPO in Hong Kong, working with China International Capital Corporation and Deutsche Bank and aiming to raise at least $1 billion. The plan is preliminary and may change, with additional banks possibly joining; Momenta previously considered a U.S. listing and filed in 2024 but let that application lapse. Founded in 2016 by former Microsoft Asia AI engineers, Momenta is valued at over $5 billion after its latest financing and counts investors including GM (which invested $300 million in 2021), Mercedes‑Benz, SAIC, Toyota, Temasek, and Yunfeng Capital. The company also partners with Uber and Grab on robotaxi efforts. Momenta says it remains focused on operations amid listing rumors.

Cadillac officially revealed the Chinese name “凯威德” for its full-size electric road flagship SUV VISTIQ ahead of its China launch, and is offering a promotional 1,000,000-point blind-order incentive for early reservations. The VISTIQ shares lineage with the Escalade, measures 5222/2026/1799 mm with a 3094 mm wheelbase, and emphasizes off-road capability with a 19° approach angle and 193 mm ground clearance. It features an industry-first hidden in-cabin lidar and high-level driver assist from Momenta. The 2026 VISTIQ starts at $78,790 (~¥544k), packs a 102 kWh battery for up to 300 miles (≈483 km), delivers 615 hp and 650 lb·ft torque, and claims sub-4s 0–100 km/h and fast charging adding ~79 miles in 10 minutes.

NIO (Shanghai NIO Automobile Co.) has signed an agreement with Anting town government to build a new battery R&D base in Jiading, Shanghai, focused on next-generation power batteries. NIO, which runs global business and production R&D headquarters in Shanghai and already operates multiple test centers and a battery test facility in Jiading, will expand local battery development capacity with this project. The move aligns with recent strategic ties between NIO and Bosch announced during German Chancellor Merz’s China visit, covering collaborations on battery management, drive and chassis-by-wire systems, and vehicle electronics — signaling deeper industry cooperation on core EV technologies and supply-chain innovation. This investment underscores China’s ongoing push in EV battery tech and industrial clustering.

xAI’s first Grok model was developed and launched at extreme speed: the core team moved from an under-tuned base model and unfinished UI to a public release within about 36 hours after Elon Musk declared an all-hands “extreme hardcore” push. Former xAI member Toby Boren recounts the intense sprint—no stimulants, just adrenaline—and credits a small, skilled, highly motivated team and clear objectives for outpacing larger rivals. He cautions against performative overwork, urging focus on purposeful “hardcore engineering culture” rather than sacrificing wellbeing for show. The piece underscores xAI’s high-intensity work norms, internal culture debates, and how rapid execution can drive sudden visibility for AI startups.

Elon Musk said on X that Tesla could be among the first companies to develop a humanoid artificial general intelligence (AGI), potentially producing Optimus robots that run AGI. The remark aligns with Tesla’s pivot to convert the Fremont Model S/Model X lines into Optimus pilot production, with an initial target of one million robots per year. Musk has linked Optimus’ long-term potential to von Neumann probe–style self-replication, which would require true AGI. The piece notes xAI’s Grok project and recent xAI–SpaceX merger activity, suggesting deeper integration across Musk’s companies if Tesla deploys AGI-powered humanoid robots. Other AI leaders, like DeepMind’s Demis Hassabis, also emphasize AGI’s transformative impact.

NIO says its low-cost electric model, the firefly, has topped 100,000-yuan-class sales and reached over 40,000 deliveries within a year. Founders Li Bin and Qin Lihong announced during a livestream that firefly will roll out more special-edition trims in 2026 after last year’s limited “游牧美拉德”, “夜行生物” and “天马狂想曲” variants. The model is already sold in nine countries and NIO plans expansion into nearly 40 markets across Europe, North America and Southeast Asia; a Thailand launch is imminent and right‑hand‑drive production follows Singapore sales. Current incentives run through March 31, 2026, and the car’s specs include a 42 kWh battery, 420 km CLTC range and a 105 kW rear motor.

@ylecun: RT @ylecun: @elonmusk Tesla will be the first company to make Artificial Grokon Intelligence.

Tesla’s Berlin Gigafactory head Andre Thierig publicly refuted a German newspaper’s claim that the plant produced only 149,000 Model Y units in 2025, saying the figure is “completely wrong” and asserting the factory exceeded 200,000 vehicles despite a first-quarter line change and ramp-up. Thierig also disputed a reported 0.74% profit margin and highlighted Tesla’s >€5 billion investment in Grünheide since 2020, creation of ~11,000 well-paid permanent jobs, and a near-€100 million battery cell investment that will add hundreds of roles. The newspaper’s investigation team countered that its numbers derive from Tesla’s own annual financial statements and third-party data provider Inovev, and noted a drop in eligible factory-voter headcount. This dispute matters for perceptions of Tesla’s European production scale, financial health, and local employment impact.

Most newsworthy: Xiaomi Auto’s new-generation SU7 is now arriving at stores nationwide, with the company saying by mid-March the model will be displayed in 492 outlets across 143 cities and launch in April 2026 with a starting pre-price of ¥229,900. Xiaomi showcased specs: multi-tier safety kit (2200MPa rollover cage, four door beams, nine airbags including rear-side), reinforced chassis materials, triple-redundant door opening, and lidar plus 4D millimeter-wave radar for driver assistance. Powertrains use V6s Plus motors (320 hp standard, 690 hp Max) and a silicon-carbide high-voltage platform (752V/897V), with CLTC ranges up to 902 km. This rollout and hardware positioning signal Xiaomi’s aggressive push in EV manufacturing and smart-car competitiveness.

BYD has overtaken Tesla in annual EV sales in 22 countries and regions, driven by rapid global expansion and vertical integration. Nikkei/ S&P Global Mobility data show BYD’s exports and local production grew from 2021–2025 into 113 markets, with 2025 marking BYD as the world’s top EV seller and first-year outselling Tesla in the UK. Key factors include in-house battery and component manufacturing enabling lower prices, targeted early bus and taxi deployments (e.g., Uruguay), expanded logistics via China-backed ports to South America, and new plants in Thailand (2024) and Brazil (2025) with a Hungary plant due in 2026. Trade barriers and geopolitical pressure in North America and regulatory moves in the EU have pushed BYD toward localized production to avoid tariffs and access markets.

In February 2026, electric vehicles (EVs) accounted for over 98% of new car registrations in Norway, with Tesla regaining its position as the top-selling brand. The surge in EV sales followed a significant drop in January due to changes in value-added tax policies, which prompted many consumers to advance their purchases. Data from the Norwegian Road Traffic Information Council revealed that 7,127 new EVs were registered in February, while traditional fuel and hybrid vehicles made up only 2% of the market. Tesla's Model Y led the sales with 1,073 registrations, reflecting a recovery in demand. This trend is noteworthy as it indicates a return to normal market activity following previous tax adjustments.

The 2027 BMW iX4 electric coupe SUV has been spotted undergoing cold weather testing, with an expected launch in 2027. The vehicle features a sleek design with a sloping roofline and a front end similar to the iX3. It will be equipped with a 17.9-inch infotainment system and a panoramic iDrive display. The iX4 will offer two versions in the U.S.: the 40 xDrive and the 50 xDrive, with the latter boasting a 108.7 kWh battery, dual-motor all-wheel drive, 463 horsepower, and a range of up to 805 km on the WLTP cycle. This model highlights BMW's commitment to electric vehicles and advanced technology in the automotive industry.

BMW has unveiled a testing video of its first all-electric M3, codenamed ZA0, showcasing its performance on the Nürburgring racetrack. Set to launch in 2026, this vehicle will feature a groundbreaking four-motor drivetrain, exceeding 800 horsepower, making it the most powerful production model in BMW's M lineup. Unlike the standard Neue Klasse models, which utilize single or dual motor configurations, the electric M3 will also offer a rear-wheel drive mode for a traditional driving experience. Additionally, it will be equipped with a dedicated battery pack exceeding 100 kWh, emphasizing BMW's commitment to high-performance electric vehicles.

A recent report by Rhodium Group reveals that Chinese electric vehicle (EV) manufacturers, such as BYD, are significantly underpricing competitors like Tesla due to structural advantages. While state subsidies are often cited as the primary reason, they only account for 5% of the cost gap. Key factors include deeper vertical integration, greater scale, and lower overhead costs, including cheaper R&D and labor. For instance, BYD's vehicle costs are reduced by extended supplier payment terms, which provide a cost advantage of $214 per vehicle. The report suggests that Western automakers face challenges in competing on price without conflicting with their governments' industrial policies.

A report by the European Transport and Environment (T&E) organization reveals that the EU can significantly reduce local battery prices through its 'European Manufacturing' initiative. Currently, the cost difference between EU-produced and Chinese batteries stands at 90%, but this could shrink to around 30% by expanding local production. The EU is set to propose an 'Industrial Acceleration Law' prioritizing local manufacturing for public funding, covering key sectors like batteries and renewable energy. T&E emphasizes the need for local battery production to safeguard supply chains, with potential cost reductions projected by 2030. This initiative is crucial for enhancing the competitiveness of electric vehicles in Europe.

Tesla has raised the price of its new all-wheel-drive Cybertruck by 17%, from $59,990 to $69,990, just ten days after its launch. This price adjustment, which took effect on March 2, follows a brief promotional period that created urgency among potential buyers. CEO Elon Musk indicated that the price would depend on demand, suggesting a strategy to boost orders through perceived scarcity. The new model features significant upgrades over its predecessor, including dual motors and enhanced functionality, making it more competitive in the electric truck market. However, the price increase raises concerns about affordability and market positioning.

Elon Musk, CEO of Tesla, announced on social media that China is rapidly transitioning towards a solar and electric future, significantly reducing its demand for oil and gas. This statement aligns with a report from China's National Bureau of Statistics, which highlights the country's impressive progress in green and low-carbon transformation. By 2025, China's clean energy generation, including hydropower, nuclear, wind, and solar, is projected to reach 42,481 billion kilowatt-hours, marking a 14.4% increase from the previous year. Additionally, new energy storage installations are expected to exceed 100 million kilowatts, accounting for over 40% of global capacity, solidifying China's position as a leader in renewable energy.

Europe's efforts to localize electric vehicle (EV) battery production are facing significant challenges due to heavy reliance on Chinese supply chains. Recent reports indicate that several key battery projects in Europe have been canceled, highlighting the difficulty of decoupling from China's dominant manufacturing capabilities. Notable companies like Porsche and Northvolt have scaled back or halted their battery production plans due to economic viability issues. Despite investments in local production, European manufacturers still depend on Chinese imports for essential components. Policymakers are attempting to address this dependency through initiatives like the 'Battery Booster' program, which offers loans to local battery firms. The situation underscores the complexities of achieving technological sovereignty in the EV sector.

Xiaomi has successfully deployed its humanoid robots in an automotive factory, achieving a continuous autonomous operation for three hours at a self-tapping nut assembly station. The robots demonstrated a 90.2% success rate in simultaneously installing nuts while meeting a production cycle requirement of 76 seconds. This task involved precise handling of self-tapping nuts, which posed challenges due to alignment and grip variations. This deployment marks a significant step for Xiaomi in scaling robotic applications within the automotive manufacturing sector, with further developments expected in various operational stations.

Xiaomi has unveiled the safety features of its new SU7 electric vehicle, emphasizing enhanced safety measures across various domains. Key upgrades include a robust 2200MPa steel structure for better crash protection, nine airbags with added rear side protection, and a triple-redundancy door handle system that allows mechanical access in emergencies. The vehicle also incorporates advanced safety technologies such as laser radar and 4D millimeter-wave radar for improved risk detection. Xiaomi's battery safety standards exceed upcoming national regulations, ensuring higher resilience against impacts and thermal events. This focus on safety aligns with the growing demand for secure electric vehicles in the automotive market.

In February 2026, 鸿蒙智行 (Harmony Smart Mobility) reported the delivery of 28,212 vehicles, marking a 31% year-on-year increase. This brings the total cumulative deliveries to over 1.28 million vehicles. The company also highlighted its commitment to customer safety, providing vehicle safety checks to 69,396 owners during the Spring Festival and maintaining 736 service centers open during the holiday. Additionally, the AITO brand under 鸿蒙智行 achieved over 58,000 deliveries in January and February combined, with a notable 83% increase in January alone. This growth reflects the rising demand for electric vehicles in China and the company's competitive positioning in the market.