QCraft unveils physical AI model at Beijing auto show Chinese autonomous driving company QCraft unveiled its physical AI model and QPilot MAX assisted-driving solution at Auto China 2026 in Beijing on Friday, as it seeks to apply physical AI technologies to assisted-driving systems and Level 4 applications such as robotaxis and autonomous logistics vehicles. The company said the physical AI model is built on a framework combining world models and reinforcement learning. The model includes a cloud-side component used offline to generate training data and simulate complex traffic scenarios, while the in-vehicle model handles driving tasks. QCraft said QPilot MAX is an assisted-driving solution supported by more than 500 TOPS of computing power and is designed to improve performance in complex urban traffic conditions. The company said its broader assisted-driving systems are available on 25 production models and are expected to be added to more than 50 models in 2026. The scale of deployment gives QCraft more real-world data and helps validate system reliability, it said. Yu Qian, co-founder and CEO of QCraft, said autonomous driving is one of the clearest early applications for physical AI. He said physical AI remains at an early stage and still needs further technological advances before wider adoption. Autonomous driving, however, already has large volumes of driving data and a more mature engineering base, making it a practical starting point, he added. Yu said QCraft is using world models and reinforcement learning to improve the training of autonomous-driving systems, as real-world testing is time-consuming and may not cover enough rare or complex traffic scenarios. The approach allows the company to generate more scenarios in simulation and use the resulting data to improve in-vehicle driving systems, he said. Yu said QCraft is focusing on improving the vehicle's AI decision-making capability rather than simply adding more
Apr 27, 2026 · via chinadaily.com.cn
About
Press
Copyright
Contact us
Creators
Advertise
Developers
Terms
Privacy
Policy & Safety
How YouTube works
Test new features
NFL Sunday Ticket
© 2026 Google LLC
Apr 27, 2026 · via youtube.com
Gasgoo Munich- Shenzhen Hangsheng Electronics Co., Ltd. unveiled three core technology foundations for AI Car smart vehicles on April 24 during the Auto China 2026. The lineup includes the Kongming 3.0 advanced intelligent driving solution. It also features the DSSAD autonomous driving data recording system. Additionally, it offers a full-stack hardware-and-software package. This package features the Mozi 3.0 central computing platform and the Moyuan AI OS. Image Credit: Hangsheng Electronics Kongming 3.0 centers on "software-hardware integration," leveraging an algorithm-defined chip architecture to achieve deep adaptation between software and hardware. This approach boosts performance while slashing system costs by over 30%. Built with an ASIL-D functional safety island, the solution supports both air and liquid cooling. This makes it viable for both internal combustion engine and electric vehicles. It represents a step toward parity in intelligence across powertrains. A leading automaker has already secured an order for the system. Mass production is slated for September 2026. The DSSAD autonomous driving data recording system complies with China's GB 44497-2024 national standard. It captures critical data to establish liability in accidents and refine algorithms. This data includes vehicle status, system commands, environmental perception, and driver inputs. Designed with ASIL-D functional safety, the system features secure hardware boot. It includes anti-tampering mechanisms and power-loss protection. The protection allows 300 milliseconds of independent operation after a total shutdown. It also offers fire and water resistance. Mozi 3.0 achieves "five-domain fusion," centrally managing the cockpit, intelligent driving, connectivity, vehicle control, and edge AI computing. By replacing traditional distributed controllers with one or two high-performance units, the platform triples resource utilization. It also slashes cross-domain data latency to under 3 milliseconds. It offers computing power ranging from 100+ to 1000+ TOPS. The Moyuan AI OS features an AI HMI. This includes a panoramic 3D cross-screen interface and
Apr 27, 2026 · via autonews.gasgoo.com
Government-Funded Study to Assess Autonomous Vehicle Integration at UK Airports - Safer Highways - 2 minutes ago - 2 min read A new government-funded feasibility study will investigate how autonomous vehicle technology could be introduced to improve transport operations within UK airport airside environments. Airports currently operate a range of separate vehicle fleets—such as cars, minibuses and buses—each with its own dispatch and scheduling systems. This fragmented approach can lead to inefficiencies and added operational complexity. The study will examine whether these systems could be brought together into a more coordinated, integrated model. The project is being led by autonomous systems developer Fusion Processing as part of the Connected and Automated Mobility (CAM) Pathfinder programme. Its focus will be on assessing the potential role of SAE Level 4 autonomous vehicles, which are capable of operating without a driver in controlled conditions, within airport settings. The CAM Pathfinder initiative, outlined in the Government’s Advanced Manufacturing Sector Plan, is intended to support growth in the UK’s connected and automated mobility sector. It is being delivered by the Department for Business and Trade in partnership with Innovate UK and Zenzic. Fusion Processing’s concept centres on managing a mixed fleet through a Remote Operations Control Centre (ROCC). This approach would allow operators to oversee multiple vehicle types from a single location, with real-time monitoring and the ability to intervene remotely if necessary. The system would also incorporate features such as demand-responsive booking, improved fleet utilisation and automated charging. As part of the study, a range of vehicle types reflecting current airport use will be considered, from smaller units for individual transfers to larger buses for higher-capacity movements. The work will also explore commercial aspects, including potential leasing and financing models alongside technical deployment considerations. Several UK airports are expected to participate, with additional operators
Apr 26, 2026 · via saferhighways.co.uk
In autonomous and self-driving vehicle news are Tesla, Pony.ai, NVIDIA, Moia America, Beep, Innoviz Technologies, SiMa.ai, Hesai Technology, Karsan, WeRide, Massimo Motor, PlusAI, Churchill Capital Corp IX, Arbe Robotics and Zoox. In this Article Tesla Robotaxi Rates Rise 41% But Maintain San Francisco Price Leadership Tesla’s Robotaxi service remains the most economical ride-hailing option in San Francisco despite a 41% average price increase since December 2025. Real-time pricing data from aggregator Obi shows that while Tesla is moving toward more sustainable market rates, its fares remain approximately half the cost of Waymo and significantly lower than Uber and Lyft on a per-mile basis. The data, collected between March and April 2026, indicates a broader market trend where Waymo, Lyft, and Uber also saw price increases of 18%, 12%, and 5%, respectively. Industry analysts suggest Tesla’s pricing shift reflects a transition from early-stage adoption tactics to sophisticated marketplace modeling. Although daily price variability has increased—ranging from $8 to $14 per ride—Tesla consistently offered the lowest price on every day measured by the Obi app. The service currently operates with safety drivers, yet it continues to set the pricing floor for autonomous mobility, forcing legacy providers and premium autonomous competitors to adjust their economic strategies as the sector scales. Tesla Faces Litigation Over FSD Capabilities And Hardware Limitations Tesla Inc. is confronting a consolidated legal challenge in the U.S. Court of Appeals for the Ninth Circuit following class certification for approximately 3,000 California owners in the LoSavio v. Tesla litigation. Lead plaintiff Tom LoSavio, a retired attorney, alleges the OEM engaged in misleading marketing regarding the autonomous capabilities of Hardware 2.0 and 2.1 suites. The U.S. Chamber of Commerce recently filed an amicus brief supporting Tesla’s effort to decertify the class, arguing that the plaintiffs failed to demonstrate uniform exposure to specific
Apr 26, 2026 · via autoconnectedcar.com
Paul Lee 특파원 hoondork1977@alphabiz.co.kr | 2026-04-27 06:29:01 [Alpha Biz= Paul Lee] Huawei announced plans to invest up to CNY 80 billion (approximately USD 17 billion) over the next five years to strengthen its autonomous driving capabilities. Speaking at the “Qiankun Technology Conference” in Beijing, Jin Yuzhi, CEO of Huawei’s smart car solutions unit, said autonomous driving requires sustained long-term investment and revealed that the company will allocate CNY 18 billion to R&D in smart driving technologies this year alone. Huawei aims to enhance the performance and reliability of its “Qiankun” autonomous driving system and solidify its position in the smart mobility market. The company is also accelerating data accumulation, a key factor in autonomous driving development. Huawei said its platform has surpassed 10 billion kilometers in cumulative driving data and is currently deployed across 50 vehicle models from 25 Chinese automakers, with about 1.7 million vehicles sold. Industry experts identify Tesla as Huawei’s main global competitor, particularly in terms of real-world driving data. While Chinese firms remain competitive in the domestic market, Tesla continues to lead globally, according to industry analysts. China has also supported market expansion by conditionally allowing Level 3 autonomous vehicles, where drivers must remain ready to intervene when prompted by the system. [ⓒ 알파경제. 무단전재-재배포 금지]
Apr 26, 2026 · via m.alphabiz.co.kr
Geely reveals EVA Cab autonomous robotaxi to rival Tesla, launching 2027 Watch out, Cybercab. Geely has unveiled its own autonomous taxi, with high-level self-driving capability. Geely has lifted the covers off a new autonomous electric car, dubbed the EVA Cab, at the 2026 Beijing motor show. The EVA Cab is a driverless car that can provide autonomous 'robotaxi' services through LiDAR sensors and satellite connectivity, and is already being tested in ‘pilot operations’ in cities such as Hangzhou and Suzhou. The Geely EVA Cab is a rival to the two-seat Tesla Cybercab, which was revealed by Elon Musk in October 2024, and has just commenced production in the US. While testing of the EVA Cab is already underway, Geely says the autonomous vehicle is scheduled to launch in 2027 through a subsidiary called CaoCao Mobility. Underpinning the operation of the driverless vehicle is what the brand calls the ‘Quantum Level AI Electronic and Electrical Architecture’, which is claimed to allow for a secure connection to a cloud to power the vehicle, assisted by AI tech. Autonomous driving comes through a new-generation LiDAR system, which uses 2160 laser lines that can detect up to 600 metres ahead, and can image up to 25.92 million data points per second. The EVA Cab is rated to Level 4 autonomous driving capability, which allows it to operate without human attention or intervention in a 'geofenced' area. It is a higher level of autonomy than the 'Level 2' offered by Tesla’s Full Self-Driving Supervised software currently optional for subscription in Australia, and forms the basis of the Level 4 tech bound for the Cybercab. The EVA Cab has electric sliding doors on each side, with four seats that face each other. The interior is more spacious for occupants, without the requirement of a driver’s seat
Apr 26, 2026 · via drive.com.au
Autoblog and Yahoo may earn commission from links in this article. The Next Phase Is Underway Tesla has officially begun production of the Cybercab at Giga Texas, marking a significant shift in its positioning as more than just an electric vehicle maker. However, there’s a major catch, as unsupervised driving without traditional controls has yet to be approved. The Cybercab units rolling off the production line appear to lack a steering wheel and side mirrors, and likely pedals as well. Without approval for unsupervised autonomous driving, the two-seater robotaxi cannot enter service, though Tesla previously said its Plan B would be to install steering wheels and pedals. If anything, the start of production suggests the company is confident the Cybercab’s original purpose can be delivered. Purpose-built for autonomy Cybercab in production now at Giga Texas pic.twitter.com/Y9qG3KyWBa— Tesla (@Tesla) April 23, 2026 Meeting the Requirements Interestingly, the National Highway Traffic Safety Administration imposes an annual 2,500-unit production limit on vehicles that do not meet the Federal Motor Vehicle Safety Standards (FMVSS), some of which concern equipment the Cybercab appears to lack. However, Tesla Vice President of Vehicle Engineering Lars Moravy replied “No” to a question on X about whether the production cap applies to the Cybercab, a claim seemingly backed by drone footage showing a production unit with an FMVSS sticker. While awaiting further developments, Tesla has expanded its unsupervised Model Y robotaxi operations to geofenced areas in Dallas and Houston. The company has also outlined plans to expand into more cities this year, including Phoenix, Miami, Orlando, Tampa, and Las Vegas. The Cost-Cutting Play The recent report that Tesla is building private Supercharger sites – likely intended exclusively for its robotaxi fleet – also suggests the company is moving forward with its plans. The decision to remove the steering wheel
Apr 26, 2026 · via autos.yahoo.com
Nexteer's Global First Steer-by-Wire Goes into Production PR Newswire BEIJING, April 26, 2026 BEIJING, April 26, 2026 /PRNewswire/ -- Nexteer Automotive helped a leading Chinese new energy vehicle (NEV) manufacturer bring the world's first production passenger vehicle with a full drive‑by‑wire chassis to market. The vehicle features Nexteer's steer‑by‑wire (SbW) system as a key enabler. The SbW featured in this vehicle marks Nexteer's first SbW system in mass production, representing a major step forward for the technology — moving from development and validation to full-scale production. Certified in late 2025, this system achieved the world's first ASIL D functional safety approval from DAkkS (German Accreditation Body) through close collaboration with the OEM. This certification reflects global top-tier performance in fault diagnosis, redundancy, and safety monitoring. Key features include: - Multi-layered redundancy design: Dual controllers, dual power supplies, multiple communication links, and dual actuation paths — achieving redundancy at system, hardware, and software levels. This ensures that in the event of a single fault, the backup path takes over within milliseconds with no loss of steering function. - Full‑scenario functional safety mechanism: Multi‑level monitoring and fault handling strategies covering sensors, controllers, actuators, and communication links. - Variable steering ratio: Automatically adjusts steering angle and effort based on vehicle speed and driving mode, balancing agility and comfort. - Intuitive road‑feel simulation technology: Software‑defined steering feedback delivers a more responsive and precise driving experience, adaptable to a wide range of driving scenarios. - Open interface for autonomous driving: As a key actuation layer for ADAS and autonomous driving systems, it provides real‑time, precise control capabilities, supporting the development of intelligent transportation systems. Steer-by-Wire: Electronic Signals Replace Mechanical Links, Flexible Configurations for Diverse Needs By decoupling the mechanical link between the hand wheel and the road wheels, steer-by-wire replaces conventional mechanical connections with electronic
Apr 26, 2026 · via mycarrollcountynews.com
An autonomous delivery ground drone struck a pedestrian in a crosswalk. Initial investigations point to a 3D segmentation failure in the perception system. To clarify the incident, the native LiDAR point cloud captured milliseconds before the impact was extracted, and the scene was recreated in Unreal Engine 5. The goal is to determine if the victim's reflective clothing caused the algorithm to mistakenly classify the pedestrian as a static object in the environment, such as a sign or a pole. Technical workflow: Open3D, Foxglove Studio, and Unreal Engine 5 ð ï¸ The forensic process begins with extracting the raw point cloud using Python and Open3D, filtering out environmental noise and isolating the critical frame prior to impact. This cloud is exported in PLY format for analysis. With Foxglove Studio, the LiDAR sensor data is visualized synchronized with the vehicle's telemetry, allowing the identification of the pedestrian's trajectory and the planning system's response. Subsequently, the scene is imported into Unreal Engine 5, where the urban geometry is recreated and the point cloud is positioned. A reflectivity filter is then applied to the points, simulating the behavior of the pedestrian's textile material. The results show that the points corresponding to the reflective jacket exhibit anomalous intensity, similar to that of road signs, which led the 3D segmentation model to group them within the static object class, ignoring their movement. Lessons for autonomous perception safety â ï¸ This case demonstrates that material reflectivity not only affects the sensor's range but can also induce fatal errors in semantic classification. The reconstruction in Unreal Engine 5 allows visualizing the algorithmic blind spot the vehicle had. For future systems, it is recommended to implement cross-validation between the point cloud and thermal or event camera data, as well as to train models with datasets that include pedestrians
Apr 26, 2026 · via foro3d.com
Gasgoo Munich- By 2026, the battleground for technology has shifted once again. Previously, the race was about range breaking 1,000 km, acceleration dipping under three seconds, and NOA coverage exceeding 100 cities. This year, the focus has clearly narrowed. Powertrains are no longer judged solely on energy density, but on discharge performance at minus 30 degrees Celsius. Assisted driving isn't just about the number of cities covered, but the synergy between algorithmic architecture and perception hardware. Cockpit interaction has moved beyond counting features to testing whether voice assistants understand nuances like "I'm a bit cold" — marking the arrival of AI agents. The industry's technology race is drilling down ever deeper. Ouyang Minggao, an academician at the Chinese Academy of Sciences and professor at Tsinghua University, predicts that during the 15th Five-Year Plan period, competition in the new-energy vehicle sector will shift from leading in single technologies to a battle of entire technological systems. The dividends from isolated breakthroughs are shrinking, replaced by the need for systemic integration across safety, efficiency, charging, and intelligence. The Auto China 2026 offers a clear window into this transition. Where Is the Tech Race Heading? New models serve as the primary stage for these technologies. At this year's show, alongside AI integration and smart driving, underlying technologies like chassis and battery systems are seeing major upgrades. These are no longer mere concept displays; they are being rolled out at scale in production models. Image Source: Li Auto Chassis upgrades are a clear signal. The Li Auto L9 Livis, for instance, features an 800V fully active suspension and a full steer-by-wire chassis with hardware pre-installed for Level 3 autonomous driving. XPENG's GX made its global mass-production debut with a steer-by-wire system and pre-installed Level 4 autonomous driving hardware. NIO's ES9 comes equipped with steer-by-wire, rear-wheel
Apr 26, 2026 · via autonews.gasgoo.com
Autoblog and Yahoo may earn commission from links in this article. Reshaping Tesla’s Roadmap Elon Musk used Tesla’s April 22 earnings call to outline a future that leans heavily into autonomy. The headline claim? The upcoming second-generation Tesla Roadster will be the only model in the lineup designed for full human control. Everything else is expected to transition toward self-driving capability through software and hardware evolution. He also framed the Tesla Cybercab as the brand’s de facto compact car. It is a two-seat vehicle that prioritizes space efficiency, based on Tesla’s claim that 90% of trips involve one or two occupants. Vehicles built before 2023 are limited by older hardware. Newer models equipped with Hardware 4 are expected to unlock full autonomy via updates to Full Self-Driving. For legacy owners, Tesla is offering retrofit pathways, trade-in credits, or camera upgrades. Musk also revived his most provocative idea yet. He suggested future Teslas could incorporate some form of flight capability. The Long-Delayed Roadster Tesla has spent years building anticipation around the new Roadster. First shown in 2017, the car has become a symbol of the company’s ambition rather than a product on the road. Musk has teased what he calls an “epic” reveal, claiming the car will redefine performance benchmarks for electric vehicles. Musk has openly stated that the Roadster is not being engineered with safety or autonomy as its primary focus. That runs counter to Tesla’s broader strategy. Instead, the Roadster is framed as a purist machine. It is meant to celebrate driving at a time when the rest of the lineup is moving away from it. That contrast is deliberate. It turns the Roadster into both a halo product and a philosophical outlier. Tesla Is Betting Everything on Autonomy Tesla’s confidence in its self-driving stack has grown more visible
Apr 26, 2026 · via autos.yahoo.com
P3 has embarked on a new development programme aiming to make autonomous driving safer, and help OEMs more rapidly develop and roll out new autonomous vehicles STUTTGART, GERMANY, February 24, 2026 /EINPresswire.com/ — P3 digital services has put together a team to create a middleware solution that resides between Android Automotive OS and a car’s ADAS (advanced driver assistance system). The new platform is being designed to translate high definition map information originating in the Android Automotive OS domain into a standardized protocol based on ADASIS which interfaces directly with the car’s core operational network and its ADAS. In the era of connected vehicles and ADAS, the employment of digital map data has been a game-changer. Traditionally accessible exclusively by navigation systems, when this data can be directly linked to other in-vehicle applications, the benefits to both OEMs and their customers can be transformational. With this approach, much more meaningful and actionable information is derived from the mapping system, and it’s sent straight to the car’s steering and traction systems, for example. This therefore enables greater responsiveness, earlier driving adjustments, and predictive functionality, delivering better and safer driving experiences, as well as greater energy efficiency. In the first phase, P3 translates information provided by Google VMS (Vehicle Map Service) into relevant live ADAS information. An overarching goal of this project is to create a solution that is vendor-agnostic, able to effectively interface with the car regardless of the supplier of its ADAS. This ‘universal language’ will enable transferrable functionality, free of the constraints of relying on proprietary technology. The ambitious initiative is being supported by a major international OEM. At the heart of P3’s endeavor is support for the ADASIS framework, ADASIS is an advanced set of interface specifications created to enhance advanced driver assistance systems. ADASIS has supported advancement
Apr 25, 2026 · via wisfarmer.com
Tesla is officially taking the training wheels off its autonomous future. After months of building modified test mules, the company has confirmed that the true, control-free Cybercab is now in mass production at Gigafactory Texas. In a promotional video shared on X, Tesla showcased a fleet of steering wheel-less vehicles rolling off the assembly line and autonomously navigating their way out of the factory and onto public streets. "Cybercab has started production," Elon Musk added, marking a major milestone for the program. This news comes just days after footage from Giga Texas revealed over a dozen newly manufactured units sitting in the factory’s outbound lot. Notably, these vehicles feature the minimalist interior Tesla promised: no driver controls, no pedals, and no side mirrors. The 2,500-Vehicle Question Not a Tesla App While seeing the hardware is exciting, a massive question has been looming over the program: How will Tesla bypass federal red tape? The U.S. National Highway Traffic Safety Administration (NHTSA) currently maintains a 2,500-vehicle-per-year, per-manufacturer limit on the deployment of autonomous vehicles (AVs) that lack traditional controls. However, Tesla’s Vice President of Vehicle Engineering, Lars Moravy, has finally put those fears to rest. When asked on X if the 2,500-vehicle cap applies to the Cybercab, Moravy simply responded: "No." This exemption is a game-changer. It appears that instead of asking for a special permission slip, Tesla is self-certifying the Cybercab under the Federal Motor Vehicle Safety Standards (FMVSS). An FMVSS sticker was previously spotted on a production unit, suggesting Tesla has found a way to meet safety standards without needing a traditional steering wheel, effectively allowing it to produce as many units as the factory can pump out. Ramping Up the Robotaxi Network This production surge arrives at the perfect time. Tesla recently expanded its Robotaxi service to Dallas and
Apr 25, 2026 · via notateslaapp.com
The Chinese car market has become incredibly competitive, and that battle is starting to be exported to other markets, including Europe. Automakers from China must find ways to differentiate themselves. In March 2026, XPENG launched its VLA 2.0, an autonomous driving system that isn’t just in public testing phase like Tesla’s FSD, but a full shipping product. It’s already reported to be a major factor boosting sales. Could this have a similar impact in Europe, where emotional brand attachment and heritage are much more important factors? VLA 2.0: Driving XPENG Sales In China Having experienced VLA 2.0 in action at the Beijing Auto Show, with myself in the driving seat of an XPENG P7 Ultra, I can confirm that it is very capable. It drove confidently around the urban streets of the city, navigated traffic, and coped with unpredictable motorbikes, narrow streets, and complicated junctions faultlessly. I’ve also experienced Tesla FSD as a passenger, which is impressive too, but VLA 2.0 had a more human-like operation, even pulling across lanes when a truck at a junction looked like it might try to enter the lane we were in. No human interventions were required at all across over 40 minutes of driving, apart from exiting and entering a barrier-controlled car park. “Our best feature is autonomous driving,” says He Xiaopeng, Chairman & CEO, XPENG. “VLA 2.0 has had very good results. The first generation was already the top autonomous driving system in China. With our latest generation, we are the best of the best. Our new generation can run everywhere, and even your mother is safe driving VLA 2.0. Our consumers use it almost 120% more than the previous generation. Next year, we also want to go global, including European and other countries. We are testing VLA in Europe. Maybe next
Apr 25, 2026 · via forbes.com
Thousands squeezed into the world's biggest car show Friday, snapping selfies beside Chinese electric vehicles as automakers showcased AI, humanoid robots and flying-car ambitions in a cut-throat market. Rows of influencers posed in front of gleaming models at the capital's cavernous international exhibition centre, darting to suitcases stuffed with outfit changes, while animated CEOs worked the crowds in front of massive LED screens. Legacy overseas brands such as Volkswagen, Toyota and BMW once dominated in China, but have lost market share in past years to domestic firms that beat them to the electric vehicle revolution and undercut them on price. Chinese manufacturers including BYD, Xiaomi and XPeng are now also at the forefront of integrating AI software and autonomous driving technology into their EVs. The Auto China exhibition, hosted at two side-by-side venues, spans 380,000 square metres (four million square feet), according to organisers -- more than 50 football pitches. More than 1,400 vehicles from hundreds of Chinese and foreign companies are on display from Friday, when the show opened to industry professionals and the media, and later to the public from April 28 until May 3. While traditional leading brands like Germany's BMW and Mercedes held sweeping areas of the vast halls, most of the event's mega stages were dominated by Chinese brands including BYD and battery giant CATL. - Robots, flying cars - At the sprawling expo, crowds cheered as XPeng chief executive He Xiaopeng unveiled the company's new GX, a six‑seat electric SUV. The imposing 5.2-metre vehicle incorporates AI technology and is aimed at breaking into the luxury market, He said. It would soon be followed by humanoid robots this year, He promised, and eventually by flying cars, which XPeng hopes to mass-produce. Foreign automakers are increasingly collaborating with Chinese firms to stay competitive. BMW has partnered
Apr 25, 2026 · via kten.com
He Xiaopeng told reporters at Auto China 2026 on April 25 that Xpeng’s Vision Language Action system already outperforms Tesla’s Full Self-Driving in complex Chinese driving scenarios , and staked a public wager on closing the remaining gap entirely by August 30. The bet is not metaphorical. He set it in December 2025 after personally testing Tesla’s FSD V14.2 in Silicon Valley and returning to China convinced that the technology had reached “near-Level 4” performance. The terms are public and specific: if Xpeng’s VLA 2.0 system achieves within China the same overall road experience that Tesla’s FSD V14.2 delivers in Silicon Valley by August 30, 2026, He will build a Chinese-style canteen in Silicon Valley modeled after the cafeteria at Xpeng headquarters. If VLA falls short, Liu Xianming, Xpeng’s head of intelligent driving, will run naked across the Golden Gate Bridge. He invited Elon Musk to come test the result personally. As wagers go, it communicates the timeline with unusual clarity. Xpeng launched VLA 2.0 in March 2026 and showcased it at Auto China 2026, the Beijing International Automotive Exhibition running through May 3. The Vision Language Action architecture replaces the fragmented pipeline approach , where separate AI modules handle perception, planning, and action in sequence , with a single end-to-end model that processes visual input and generates vehicle actions the way a human driver does: continuously, in context, without handoffs between subsystems. The practical result, in Xpeng’s own comparison data, is measurable: on a 20-kilometer complex urban test route, Tesla FSD V13.2.9 required five driver takeovers while Xpeng’s second-generation VLA required one. Morgan Stanley analysts who tested the system in early March 2026 described themselves as surprised by how quickly the gap with Tesla had closed. He Xiaopeng has publicly welcomed all competitors, including Musk, to experience the system
Apr 25, 2026 · via startupfortune.com
Tesla’s long-promised Tesla Roadster is back in the headlines again, and this time, Elon Musk is making a pretty bold claim. During Tesla’s latest earnings call, Musk said the upcoming Roadster could become the only future Tesla that people will actually drive themselves. That’s because the company appears to be betting heavily on autonomous vehicles moving forward. And if Musk follows through, traditional driving enthusiasts may not love where Tesla is headed. Musk Says Tesla’s Future Is “Almost Entirely Autonomous” During the company’s Q1 2026 financial presentation, an analyst asked Tesla about future vehicle plans. That included potential new family vehicles and smaller models. Instead, Musk shifted the conversation toward autonomy. “In fact, long term, the only manually driven car will be the new Roadster,” Musk said during the call. That suggests future Tesla products could prioritize self-driving capabilities over traditional driver engagement. The Roadster Still Hasn’t Arrived There’s just one problem. The second-generation Roadster still doesn’t exist as a production vehicle. Tesla first revealed the car back in 2017. Since then, it has faced repeated delays. Musk now claims Tesla could debut it “in a month or so,” though he also admitted it still requires significant testing and validation. Tesla fans have heard similar timelines before. Cybercab And Semi Are The Bigger Priority Tesla appears far more focused on autonomous vehicles like the upcoming Tesla Cybercab. Musk said production has already started in Austin. He also said the self-driving Tesla Semi will begin scaling production soon. According to Musk, both vehicles will ramp up slowly before accelerating production later this year and into 2027. That makes it pretty clear where Tesla’s priorities currently sit. The Roadster May Be Cool, But It Won’t Drive Revenue Musk also admitted the Roadster likely won’t have a major impact on Tesla’s finances. “I
Apr 25, 2026 · via autos.yahoo.com
AI Transforms the World: XPENG Showcases Its Full-Stack Physical AI Ecosystem at Auto China 2026 Rhea-AI Summary XPENG (NYSE: XPEV) presented a full-stack Physical AI ecosystem at Auto China 2026 featuring new models GX, MONA M03, next P7, X9, humanoid robot IRON, and a flying car prototype. The company released its first VLA 2.0 Intelligent Driving Report with user metrics and announced Campus and Underground Parking Safari features. Key metrics: Ultra-series orders +118% month-over-month; nearly 100,000 VLA 2.0 demo participants with 98% satisfaction; 98.52% daily activation in week-one; overseas deliveries of 45,000 units in 2025 (+96% YoY). GX equipped with up to four Turing chips (3,000 TOPS). Positive - Overseas deliveries reached 45,000 units in 2025 (+96% YoY) - Ultra-series orders increased 118% month-over-month - VLA 2.0 demo reach of nearly 100,000 participants with 98% satisfaction - First-week activation of intelligent driving at 98.52% among new owners - GX computing capability: up to four Turing chips delivering 3,000 TOPS Negative - None. Key Figures Market Reality Check Peers on Argus XPEV fell 6.31% with heavy volume, while key peers like RIVN (-3.2%), NIO (-2.48%), STLA (-2.36%) and F (-1.03%) were also down, and LI was flat, pointing to a mix of sector pressure and stock-specific weakness. Previous AI Reports | Date | Event | Sentiment | Move | Catalyst | |---|---|---|---|---| | Apr 17 | AI ecosystem preview | Positive | -0.8% | Previewed Physical AI ecosystem, VLA 2.0 and global expansion ahead of show. | | Jan 12 | AI strategy update | Positive | +8.4% | Announced P7+ flagship rollout, AI VLA2.0 platform and global delivery milestones. | | Sep 08 | AI tech showcase | Positive | +3.8% | Showcased AI mobility tech at IAA, including L4 autonomy and flying car plans. | | May 28 |
Apr 25, 2026 · via stocktitan.net
Bosch pushes Level 3 autonomous driving with AI and by-wire With AI-powered systems and redundant hardware, supplier Bosch is ushering in the next stage of mobility – from highways to urban rapid transit. The vision of the robot car, which largely takes care of itself and turns humans into passengers, is becoming more tangible. Assisted driving at Level 2 is already part of everyday automotive life for many. Bosch is now pushing the decisive transition to highly automated driving (SAE Level 3). At this stage, responsibility shifts from the human to the machine in specific application cases. The driver can take their hands off the steering wheel, take a nap, and look away from the road – a technological advancement that goes beyond comfort functions. At Bosch, the path to autonomy leads via a new approach. Earlier, rigid, rule-based programming reaches its limits in the complex reality of road traffic. The company's answer is: artificial intelligence in every software module. By using AI, the vehicle can react much more flexibly to unpredictable situations, it is said. Coupled with a redundant safety architecture that immediately takes over if a system fails, Bosch aims to create a basis of trust for Level 3 after many unsuccessful attempts by other automotive manufacturers and suppliers. The goal is a system that operates reliably at speeds of up to 120 km/h, even under difficult conditions such as poor visibility. According to Bosch, the driver will thus regain valuable time – whether on the highway or on multi-lane expressways in conurbations. The technology, according to the supplier, not only takes over lane keeping but also independently initiates lane changes and coordinates acceleration and braking. When milliseconds decide on safety An important building block of this “new freedom” is the vehicle's ability to swerve quickly and precisely
Apr 25, 2026 · via heise.de