Autonomous & Self-Driving Vehicle News: Tesla, Waymo, WeRide, Helm.ai, Torc & More

In autonomous and self-driving vehicle news are Tesla, Waymo, WeRide, Helm.ai, Reuters, FORT Robotics, Mapless AI, American Rheinmetall, Harbinger, DENSO, Carnegie Mellon University, Massimo Group, Torc Robotics, Mila, Voltera, Revel, Swift Ride, Geotab, STRADVISION, aiMotive and AEye.

Investigation Find Flaws & Problems in Tesla’s Self-Driving

A Reuters investigation indicates that Tesla’s Full Self-Driving safety statistics rely on flawed methodology, and data labelers responsible for training the artificial intelligence system report a lack of trust in the technology. The investigation reveals that Tesla inflated its safety claims by comparing internal airbag-deployment data against federal crash statistics that utilize a lower threshold involving tow-truck dispatches. When evaluated using comparable airbag-deployment metrics, the data showed the system was roughly three times farther between crashes rather than the ten times safer figure claimed by company executives.

Former data labelers tasked with reviewing video from vehicle external cameras reported frequent failures, including inadequate responses to emergency vehicles, motorcyclists, and construction zones. Dedicated analysis of near-miss footage also highlighted instances where the software failed to detect pedestrians and children in crosswalks. Additionally, workers noted instances of significant speeding following the introduction of more aggressive driving profiles.

The report details that Tesla engaged in extensive high-definition mapping of specific operating zones ahead of public robotaxi demonstrations in California and Texas, challenging public assertions that the camera-based system operates entirely without localized digital mapping. The labor-intensive annotation required for these limited operational design domains has reportedly restricted the scaling of the unsupervised fleet, which maintains a limited deployment footprint nearly a year after initial pilot launches.

 Waymo Deploys Sixth-Generation Autonomous Ojai Vehicles to San Francisco and Los Angeles Fleets

Waymo has commenced deployment of its sixth-generation driverless system on Ojai vehicles, initiating employee and guest testing across San Francisco and Los Angeles. Engineered on an electric vehicle base platform supplied by Chinese manufacturer Geely’s Zeekr subsidiary, the Ojai configuration serves as the primary asset for the Alphabet-owned company’s planned market expansions. The upgraded Waymo Driver architecture utilizes lowered-cost hardware and enhanced sensor capabilities designed to stabilize vehicle operations within adverse weather conditions.

The sixth-generation hardware suite leverages an integrated 17-megapixel imager alongside modernized lidar and radar components, reducing total camera count while expanding spatial perception. To address inclement weather performance, the perception system incorporates localized cleaning mechanisms to mitigate sensor occlusion from ice, rain, and road debris. While the proprietary autonomous compute, sensor integration, and data processing occur domestically, the reliance on Geely platforms has prompted federal legislative scrutiny regarding telematics data security. Waymo confirmed the autonomous driving software remains isolated from the OEM, and noted that the sixth-generation architecture will additionally deploy on Hyundai Ioniq 5 platforms alongside existing fifth-generation Jaguar I-PACE fleets.

WeRide Deploys Level 4 Robobus at Roland-Garros for Third Consecutive Year

WeRide, in partnership with Renault Group, has deployed its autonomous L4 Robobus at the Roland-Garros tennis tournament in Paris for the third consecutive year. Operating as the exclusive autonomous public shuttle service for the event, the deployment utilizes beti as the ground fleet operator to manage high-density transit demands during the Grand Slam tournament. The autonomous shuttle operates on a 2.8-kilometer fixed route along Avenue de la Porte d’Auteuil, connecting three key transit nodes within a 12-minute journey time. The 2026 deployment maintains extended night-time operations following a successful 2025 pilot, serving as a high-visibility validation of WeRide’s scaling European footprint, which includes commercial operations and public-road L4 pilot programs across France, Spain, Belgium, Switzerland, and Slovakia.

Helm.ai Standardizes Full HD Generative Simulation with GenSim-3 and VidGen-3 Systems

Autonomous driving software developer Helm.ai has introduced GenSim-3 and VidGen-3, its next-generation generative AI foundation models designed for high-end ADAS, autonomous vehicles, and robotics. The platform is the first to deliver native Full HD (1920×1080) resolution across a synchronized 6-camera, 360-degree surround-view configuration. Generating a total aggregate canvas of 12 megapixels per timestep, the models achieve a fivefold increase in pixel density compared to existing industry benchmarks for generative world models, which typically operate at sub-HD or VGA resolutions. The transition to native 2-megapixel output per camera addresses the “Sim-to-Real” domain gap that occurs when training advanced perception neural networks on low-resolution synthetic data. Because production Level 2 and Level 4 autonomous vehicle architectures employ high-resolution physical sensors, the simulated training data must match the hardware’s native pixel density to optimize perception accuracy. Helm.ai’s architecture acts as a hardware-accurate virtual sensor twin by deliberately replicating physical sensor anomalies, including lens flares, sensor banding, and dynamic exposure blinding, ensuring the simulation mirrors real-world edge-case conditions. The dual-model pipeline provides automakers with two primary functional capabilities: GenSim-3 executes high-fidelity scene transfer, allowing engineering teams to restylize real-world video logs by dynamically altering weather, illumination, and object textures across all six cameras simultaneously; VidGen-3 generates completely synthetic driving sequences, simulating human-like agent behaviors and traffic logic from scratch. Helm.ai developed this Full HD capability using a highly optimized compute architecture on a cluster of a few hundred advanced GPUs, offering a more efficient computational footprint compared to traditional end-to-end models requiring massive hardware scaling.

FORT Robotics Acquires Mapless AI to Integrate Teleoperation and Active Safety into Physical AI Platform

Functional safety and machine control developer FORT Robotics has acquired Mapless AI, a Boston- and Pittsburgh-based specialist in vehicle teleoperation and autonomy supervision. The transaction expands the capabilities of FORT’s proprietary Trust Platform by incorporating remote human-in-the-loop teleoperation and onboard active safety systems. The integration marks a shift in FORT’s product portfolio from safety-certified machine control hardware and software toward a comprehensive infrastructure for supervised autonomy across industrial environments. The acquisition introduces two primary architectural layers to FORT’s existing distributed control technology. The remote teleoperation system allows off-site specialists to monitor and control autonomous vehicles or machinery over long distances, addressing enterprise demand for human-in-the-loop fallback systems without exposing personnel to high-risk environments. Concurrently, the onboard active safety layer integrates environmental perception technology, transforming traditional reactive safety mechanisms into proactive architectures. This enables real-time threat perception, predictive path planning, and automated contingency maneuvers based on immediate environmental sensing. Mapless AI, founded by automotive and robotics veterans from Bosch, Apple, Uber, Aptiv, and nuTonomy, brings deep technical expertise in automotive functional safety standards and real-world robotics deployment. FORT plans to leverage this engineering asset to accelerate market penetration into complex sectors including logistics, construction, defense, and last-mile delivery. Moving forward, the unified platform aims to allow a single remote operator to supervise and intervene across multi-vehicle fleets, decoupling labor from hazardous zones while maintaining supervisory control.

American Rheinmetall and Harbinger Partner on Hybrid Drive-by-Wire Military UGVs

Defense contractor American Rheinmetall and industrial electrification manufacturer Harbinger have formed a strategic partnership to develop and deploy a family of robotic and uncrewed ground vehicles (UGVs) for the U.S. Department of War. The collaboration integrates American Rheinmetall’s combat vehicle systems engineering and modular mission architectures with Harbinger’s commercially derived, dual-use hybrid electric chassis. The joint effort aims to accelerate the deployment of scalable, attritable robotic systems for tactical wheeled logistics, contested resupply, and manned-unmanned teaming operations. The core technology platform centers on Harbinger’s fully drive-by-wire, autonomous-ready medium-duty chassis. This architecture features a scalable battery system paired with a gas-powered range extender, providing extended operational endurance alongside tactical advantages such as silent watch capabilities and reduced thermal and acoustic signatures. By leveraging a high-volume commercial foundation, the partnership intends to deliver cost-competitive, sovereign military robotics capable of being mass-produced to meet Department of War requirements for modern combat effectiveness and industrial readiness. Both entities maintain a domestic engineering and manufacturing footprint. American Rheinmetall is expanding its production and integration facilities across Michigan, Ohio, and Maine, while Harbinger handles in-house design and assembly of powertrains, battery packs, and chassis at its California headquarters. Following Harbinger’s recent acquisition of Phantom AI for advanced driver-assistance systems (ADAS) integration, the companies intend to launch joint vehicle demonstrations to pursue near-term prototyping opportunities via Commercial Solutions Openings and Other Transaction Authorities.

Waymo Initiates High-Definition Mapping Fleet Deployment in Northern Virginia

Alphabet-owned autonomous vehicle developer Waymo has commenced fleet operations in Northern Virginia to map the municipalities of Alexandria and Arlington. Because fully autonomous passenger transport remains prohibited under current Virginia vehicle codes, the mapping fleet is operating under manual human control. High-definition spatial mapping utilizes vehicle-mounted lidar, radar, and camera sensor arrays to construct precise 3D digital representations of road geometry, traffic control infrastructure, and environmental variables, serving as the technical prerequisite for eventual robotaxi service deployment. The operational expansion coincides with legislative efforts by Virginia’s Autonomous Driving Work Group and the state Senate to formulate regulatory frameworks for licensing driverless passenger and freight vehicles. However, regional legislative timelines suggest that commercial autonomous operations may not receive regulatory authorization until 2028. Waymo representatives indicated that while the high-definition mapping pipeline typically requires a 12-to-18-month capitalization and validation phase prior to commercialization, the company has no immediate timeline for launching its Waymo One ride-hailing service in the Commonwealth. Waymo’s presence in the Washington, D.C. metropolitan area represents a strategic expansion into the highly regulated mid-Atlantic corridor, contrasting with its established commercial operations in sunbelt markets. The company faces localized legislative headwinds, including a recently stalled autonomous vehicle authorization bill in Maryland and pending regulatory permitting hurdles within Washington, D.C. Beyond regional policy challenges, Waymo is addressing operational edge cases, having recently issued a voluntary software recall to update automated driving system responses to flooded, high-speed roadways following severe weather disruptions across its southern regional hubs.