In autonomous and self-driving vehicle news are Zoox, Waymo, Uber, XPENG, Tesla, Bosch, Cariad, Valeo and NHTSA.
In this Article
Zoox Details Production Robotaxi Design
Autonomous mobility developer Zoox has unveiled the production-intent iteration of its purpose-built, bidirectional robotaxi, introducing targeted cabin and exterior modifications derived from early fleet deployment testing and feedback from 500,000 riders. The structural and aesthetic revisions establish the finalized baseline for the company’s upcoming commercial assembly pipeline. Zoox plans to initiate volume manufacturing at its dedicated serial production facility in Hayward, California, with the capability to ramp output up to 100 vehicles per week to support regional ridehailing fleet expansions later this year, subject to regulatory clearance.
Interior modifications focus on cabin comfort, materials durability, and usability within the vehicle’s symmetrical, carriage-style seating configuration. The color, material, and finish palette has been transitioned to a lighter, monochrome design featuring aloe green seating alongside stone-grey flooring and structural trim to minimize visual distraction. Ergonomic adjustments include increased cushioning and re-contoured geometry across all passenger seats and headrests. Functional cabin interfaces were similarly updated, incorporating high-contrast component surfaces to prevent passengers from misplacing personal items, stabilized fluting on wireless device charging pads, expanded cup holder dimensions, and higher-luminance touchscreens.
Exterior modifications optimize physical visibility and acoustic communication pathways for public road interactions. The vehicle’s proprietary bidirectional color-rotating reflectors have been repositioned to enhance external classification of the robotaxi’s operational orientation. The primary entry interface integrates an upgraded microphone and speaker hardware array to expand two-way audio throughput. These acoustic enhancements are engineered to streamline audio communication between passengers and remote operations, while providing clearer external messaging channels for general road users, support personnel, and first responders.
Waymo Ends Phoenix Robotaxi Deal Uber
Waymo terminated its long-running autonomous vehicle partnership with Uber in Phoenix, Arizona, completely withdrawing its fleet from the ride-hailing platform after nearly three years of joint operations. The companies confirmed the system wind-down following user reports of vehicle unavailability within the Uber interface, marking the conclusion of passenger operations in May 2026. The Phoenix market served as the sole metropolitan area where Waymo deployed vehicles simultaneously through its proprietary Waymo One application and the third-party Uber network, following the initial commercial launch of the pilot partnership in 2023.
The dissolution in Phoenix highlights an escalating strategic divergence between the two entities as they scale autonomous mobility operations. Waymo continues to expand its independent direct-to-consumer footprint across multiple urban centers, deploying a fleet of approximately 4,000 vehicles generating over 500,000 weekly trips. To cement direct user relationships and capture full fare revenues, Waymo introduced a 29.99 USD monthly subscription membership model in San Francisco, Los Angeles, and Phoenix. Conversely, Uber has adapted its strategy to function as a platform-agnostic robotaxi aggregator, establishing integration agreements with over a dozen alternative automated driving developers including Zoox, WeRide, Avride, and Baidu to hedge against single-supplier dependency.
While the regional termination concludes the Phoenix pilot, the overarching commercial relationship remains active in secondary markets. Waymo fleets continue to operate via Uber in Austin and Atlanta under exclusive regional distribution agreements initiated in 2025. However, competitive friction continues to manifest globally as both firms prepare for distinct London market entries later this year, with Waymo planning an independent app deployment and Uber partnering with British autonomous mobility developer Wayve.
XPENG Unveils X-Mind Framework
XPENG introduced X-Mind, a predictive world model framework designed to enable autonomous vehicles to simulate future traffic scenarios prior to executing driving decisions. Unveiled at the CVPR 2026 Workshop on Foundation Model Deployment for Embodied Intelligence, the system utilizes a Visual Chain-of-Thought approach to transition autonomous driving from reactive perception-to-action systems to proactive reasoning. The framework integrates alongside X-World and X-Foresight to complete the company roadmap for physical AI foundational models, aiming for long-horizon forecasting and human-like driving performance.
The X-Mind architecture relies on three primary technologies to balance advanced reasoning with real-time automotive deployment. Thought Sketch constructs a cognitive representation that combines Bird’s-Eye-View layouts with driving priors, preserving road geometry, obstacles, and traffic controls while minimizing computational complexity. Recurrent Block Diffusion generates high-quality future scenes within a single forward pass, bypassing the latency constraints typical of conventional diffusion models. Finally, Visual Chain-of-Thought visualization exposes internal model predictions regarding obstacle trajectories and lane connectivity to increase system validation transparency.
The predictive framework was trained on hundreds of millions of real-world driving data frames to improve trajectory prediction accuracy during complex, long-tail edge cases. According to the developer, the resulting model achieves low inference latency compatible with current automotive-grade system-on-chip architectures. The software stack allows the vehicle to project how surrounding environmental elements will evolve in response to specific ego-vehicle maneuvers.
Tesla Driver Faces Manslaughter — Overriding FSD
Texas prosecutors have charged 44-year-old Tesla driver Michael David Butler with second-degree felony manslaughter following a June 19, 2026, crash in Katy, Texas, that resulted in the death of 76-year-old Martha Avila. Forensic telemetry recovered from the vehicle black box confirms the incident was precipitated by direct driver override rather than a failure of Tesla Full Self-Driving software. Investigators established that the system was operational during a residential delivery route when Butler manually depressed the accelerator pedal, bypassing FSD speed parameters.
Data logs show Butler depressed the accelerator to 100% capacity for approximately six seconds during a left-turn maneuver, accelerating the vehicle to 73 mph in a restricted zone before breaching a curb, becoming airborne, and colliding with a residence. Diagnostic analysis confirmed zero braking input in the final 60 seconds of operation and ruled out mechanical malfunctions, stuck pedals, or floor-mat interference. Forensic medical screening similarly cleared the driver of acute medical episodes or substance impairment at the time of impact.
The criminal complaint refutes Butler’s claim of a medical blackout, citing contradictions with vehicle log data, dashcam footage, and historical mobile browser queries regarding FSD performance. Tesla executive leadership reaffirmed that manual pedal inputs override autonomous speed controls and reiterated that drivers maintain legal operational responsibility. The National Highway Traffic Safety Administration has initiated a special investigation into the incident, which additionally faces a civil wrongful death lawsuit exceeding 1 million dollars filed by the victim’s estate against both the driver and the OEM.
Bosch and Cariad End Automated Driving Alliance
Bosch and Volkswagen Group software subsidiary Cariad have formally concluded their Automated Driving Alliance, completing the joint development of an artificial intelligence-based Level 2 advanced driver assistance system software stack for high-volume series production. The dissolution of the partnership shifts the relationship into a commercial phase, with both entities retaining independent rights to the shared intellectual property and underlying data pool. Bosch has confirmed global original equipment manufacturer orders for the completed software stack, while Cariad retains the capability to independently iterate the technology for individual Volkswagen Group brands.
The newly finalized Level 2 ADAS software architecture provides hands-free operation across diverse driving environments and is compatible across internal combustion and battery electric powertrain platforms ranging from entry-level to premium segments. Fleet validation operations spanned markets in Europe, Japan, and North America, utilizing a global validation fleet of up to 1,500 vehicles that included dedicated engineering test assets alongside active logistics, leasing, and service fleets. The software platform is officially slated for high-volume series deployment on the upcoming Volkswagen ID.EVERY1, scheduled to enter production in 2027.
The deployment of the jointly developed stack within a mainstream, mass-market model rather than a low-volume premium flagship is intended to validate system robustness, edge-case durability, and consumer acceptance at commercial scale. Furthermore, Bosch’s emphasis on deployment velocity highlights acute competitive pressure from domestic Chinese suppliers who have rapidly closed technological margins in advanced driver assistance systems, positioning performance parity within the Chinese market as the critical benchmark for global validation of the software stack.
Valeo North American Radar Contract
Automotive supplier Valeo has been awarded a contract to supply its advanced corner radar suite to a major North American electric vehicle manufacturer. The sensor platform is engineered to support Level 2+ and Level 3 autonomous driving architectures, providing 50 times the resolution of Valeo’s previous-generation technology. The hardware and embedded software package is designed to satisfy original equipment manufacturer requirements for low power consumption, package integration flexibility, and five-star NCAP and General Safety Regulation safety standards compliance.
The contract covers a modular, hardware-agnostic architecture capable of scaling across full vehicle product lineups, from entry-level advanced driver assistance systems to high-tier autonomous applications. The integrated sensor array enables active safety functions including automatic emergency braking, lane keeping assist, safe exit protection, and complex automated parking sequences. The system incorporates over-the-air software update capabilities, allowing vehicle manufacturers to deploy advanced “Eyes off” automated functionality to existing fleet assets post-production. The platform’s forward-facing radar components deliver long-range target detection thresholds extending up to 250 meters.
While the specific North American automaker was not named, the modular architecture aligns with Valeo’s Elevate 2028 regional growth strategy to expand state-of-the-art vehicle perception technology. Industry analysts note that the scalability of a single radar configuration reduces engineering integration costs for OEMs across multiple trim levels. Furthermore, the hardware selection highlights contrasting sensor approaches within the pure-play electric vehicle sector, where certain market participants emphasize vision-only perception systems while others increasingly integrate high-resolution radar topologies to achieve robust validation for upper-level autonomous operations.
NHTSA Starts Rulemaking to Remove Brake Pedals in AVs
The National Highway Traffic Safety Administration has initiated a formal rulemaking process to eliminate manual brake pedal requirements for vehicles designed exclusively to operate under Automated Driving Systems. The regulatory revision represents the fifth update to the Federal Motor Vehicle Safety Standards under the Department of Transportation’s current automated vehicle framework led by Secretary Sean Duffy. Under the proposed guidelines, historical vehicle stopping distance requirements will be strictly preserved, with the component exemptions applying solely to vehicle architectures that completely preclude manual driver control.
The proposed amendment follows a series of related federal safety standard revisions addressing transmission shifting mechanics, windscreen clearing systems, and tire identification placards. NHTSA Administrator Jonathan Morrison confirmed the regulatory shift is designed to dismantle architectural barriers for unique vehicle profiles while maintaining stringent performance oversight. The agency is concurrently developing empirical, real-world safety performance metrics specifically for ADS platforms and will retain full structural defect enforcement and recall authority over all participating autonomous vehicle developers.
The regulatory transition from prescriptive component checklists to performance-based capability testing serves as a critical commercial accelerator for purpose-built robotaxis, such as the Tesla Cybercab and the Zoox robotaxi, which are engineered from inception without traditional manual controls. To standardize these oversight updates, the Department of Transportation has commenced development of national AV performance standards and established a National AV Safety Forum to align automotive and technology stakeholders. The rapid sequence of safety standard modernizations indicates a deliberate agency strategy to establish a normalized federal framework ahead of scale production and commercial deployment of non-traditional vehicle designs.