WeRide, a global leader in autonomous driving technology, today unveiled WITT (World Intelligence Toward Truth), a Physical AI Cognitive Foundation Model designed to build AI cognition of the physical world through trusted facts extracted from real-world experience. The model introduces Atomic Physical Facts (APFs), the smallest verifiable units of information about the physical world, establishing a new fact-based cognitive framework for Physical AI. By transforming real-world operational data into trusted facts and trusted facts into learning signals, WITT aims to improve the reliability and efficiency of AI systems in complex environments.

WITT is built on four core capabilities: Fact Extraction, Fact Reasoning, Fact Verification and Fact Curation. These capabilities create a complete pipeline spanning scene understanding, event attribution, data validation and learning curation. The model identifies and extracts three categories of Atomic Physical Facts from real-world driving data: standard driving facts, multi-agent interaction facts and physically ambiguous conditions. It analyzes key events, behavioral relationships and evolving risks within a scene, while identifying underlying causes and potential trajectories.

To reduce hallucinations commonly associated with general-purpose AI models, WITT evaluates outputs across six dimensions: vulnerable road users, ego-vehicle behavior, surrounding vehicle behavior, scene understanding, comprehensive fact and traffic facilities. The model introduces factual confidence scoring and validates conclusions against external physical evidence. WITT achieves an average factual error rate approximately one-third that of leading general-purpose AI models in autonomous driving scenario understanding tasks.

In real-world operations, WITT automatically identifies high-value facts and routes them into the most effective learning workflows to maximize model improvement. Rare long-tail scenarios can be returned to WeRide GENESIS, the company’s proprietary general-purpose simulation model, for simulation training and scenario expansion. High-frequency everyday scenarios support reinforcement learning and workflow optimization, while abnormal or ambiguous data are directed into review processes to prevent valuable information from being discarded as noise.

Compared with general-purpose AI models that often rely on hundreds of billions of parameters, WITT delivers strong performance with a significantly more efficient architecture. The model reduces token costs by up to 98%, processes up to 10,000 minutes of vehicle-operation video per day on a single GPU and delivers up to 200 times greater data-processing efficiency in comparable workloads. In labeling workflows, a single request to WITT can generate more than 100 dynamic tags, enabling rapid retrieval and validation of real-world driving video.

Supported by this Physical AI flywheel, WeRide has become the world’s only company to achieve large-scale commercial deployment of both L4 autonomous driving and L2++ intelligent driving systems. In the L4 domain, WeRide has obtained autonomous driving permits across eight countries and markets, deployed autonomous driving products in more than 40 cities across 12 countries, and operates a fleet of more than 3,000 autonomous vehicles. Its Robotaxi services have achieved regular, large-scale fully driverless commercial operations in Guangzhou, Beijing, Abu Dhabi and Dubai.

WITT Performance Metrics

WeRide Inc. (NASDAQ: WRD) has accelerated the international expansion of its L2++ intelligent driving business by initiating on-road testing and localization validation in Germany, France, and Japan. This strategic move aims to address the challenge of delivering consistent driving experiences across diverse regulatory frameworks and road infrastructures. However, the company's stock has fallen 36% year-to-date, trading lower in Monday’s premarket session as investors assessed the expansion update against a bearish technical setup.

The solution, built on WeRide's proprietary one-stage end-to-end intelligent driving algorithm and integrated with Bosch's automotive-grade domain controller platform, is being validated for global adaptability. The unified AI architecture integrates perception, prediction, planning, and control within a single model, supporting both driving and parking scenarios. This design facilitates rapid adaptation to varying traffic regulations, road environments, driving cultures, and climate conditions, from Germany's unrestricted autobahns to right-hand-drive markets across Asia.

Commercial Performance and Adoption

While expanding its international footprint, WeRide continues to demonstrate strong performance in China's complex traffic conditions. Over the past six months, Chery Exeed ES and ET models equipped with WRD 3.0 have secured six consecutive victories in the China Intelligent Driving Challenge. The solution has also achieved design wins across more than 30 production vehicle programs with leading automakers, including Chery and GAC Group.

Technical Analysis and Market Position

Despite the commercial progress, WRD remains below its major moving averages, keeping the longer-term trend bearish. The stock trades 5.4% below its 20-day SMA of $5.83, 18.2% below its 50-day SMA, and 31.9% below its 200-day SMA. Momentum is showing early signs of stabilization, with the MACD remaining above its signal line suggesting selling pressure may be easing. Key resistance is at $6.00, while support lies at $5.00, near the stock's 52-week low of $5.18. The shares remain near the bottom of their 52-week range of $5.18 to $12.55.