XPeng Advances AI Tech with VLA 2.0 Upgrade & Robotaxi Progress

XPENG advances artificial intelligence and autonomous driving technology with major updates to its intelligent driving system and humanoid robot development. The Shenzhen-based electric vehicle manufacturer revealed significant progress in its second-quarter financial results and the rollout of VLA 2.0 Version 6.3.0, an over-the-air update improving autonomous driving capabilities across its vehicle lineup.

The company reported second-quarter revenue of $2.91 billion, an 8% increase from the same period last year and a 51.5% rise from the first quarter. Gross margin reached 20.7%, outperforming Tesla’s 16.8% margin. While vehicle margins declined to 12.1% due to costs associated with new model launches, services revenue—particularly technical research and development for Volkswagen Group—boosted overall profitability. Despite these gains, XPENG posted a net loss of $200 million, primarily driven by a 35% increase in research and development spending as the company expands globally and prepares multiple new vehicle launches.

A key highlight is the launch of VLA 2.0 Version 6.3.0, an enhanced autonomous driving system set for deployment in the coming weeks. The update introduces 4D perception, enabling the system to interpret past, present, and future driving scenarios more effectively. It uses historical data to predict up to six seconds ahead, a feature called X-Foresight, and supports real-time decision-making through updated streaming inference. XPENG claims a 300% improvement in response speed and a 20-fold increase in safety performance, positioning its system as “better than human.”

The company also confirmed its Level 4 robotaxi operations, having completed over 2,000 test orders on public roads. While regulatory approval remains stricter in China than in the U.S., XPENG expects broader commercialization soon. The VLA 2.0 update will be available on multiple vehicle models, including Ultra, Ultra SE, and Max trims, with some using a single Turing chip. XPENG emphasizes maintaining performance consistency across hardware configurations rather than limiting advanced features to premium models.

Voice interaction has also been significantly enhanced through the Master Agent system, enabling natural, context-aware commands such as navigating to a restaurant based on cuisine or pulling over at a safe location. The system processes voice commands locally using a dedicated Turing chip, reducing latency and improving real-time responsiveness.

Backed by a growing dataset of 110 million video clips and increased simulation capacity through X-World, XPENG continues to refine its models to handle complex scenarios like construction zones and ferry crossings. The company’s commitment to improving customer experience without charging for software unlocks reflects a strategic focus on long-term adoption and market leadership.

XPENG’s rapid advancements in AI-powered mobility signal a shift from traditional automaker to integrated technology enterprise, with physical AI at the core of its future growth.

#XPeng #AutonomousDriving #EVTechnology #Robotaxi #AIInAutomotive #ShenzhenTech #IntelligentVehicles #VLA2