Pony.ai and Moore Threads have formed a strategic partnership to combine China's domestically developed GPU computing power with advanced autonomous driving algorithms, accelerating L4 autonomous driving deployment and commercialization.

Strategic Partnership Details

Pony.ai and Moore Threads recently announced a strategic partnership that focuses on the deployment and large-scale application of L4 autonomous driving technology. The collaboration centers on training and optimizing Pony.ai's core technologies—its world model and Virtual Driver system—using safe, reliable AI computing power to empower autonomous driving iterations and commercial deployment.

The partnership represents the first time Pony.ai has applied domestically developed AI computing at scale in key training and simulation phases. This marks the formal entry of Moore Threads' domestic full-function GPUs into the core domain of autonomous driving, serving as an important example of coordinated innovation in China's artificial intelligence industry chain.

Technical Collaboration Framework

The two companies will jointly advance the training adaptation and validation of Pony.ai's world model and in-vehicle models based on Moore Threads' MTT S5000 training-and-inference integrated AI computing card and the KUAE intelligent computing cluster. By building an autonomous driving computing ecosystem together, they aim to achieve full-link collaboration across algorithms, data, computing power, and applications.

This collaboration is designed to accelerate the maturity and cost optimization of L4 autonomous driving technology, providing solid support for the high-quality development of smart mobility and logistics. The partnership represents a true fusion of "AI algorithms" and "AI computing power," with Moore Threads' leading domestic high-performance GPU capabilities providing strong and reliable computing support for training Pony.ai's world models and in-vehicle models.

Moore Threads' Computing Capabilities

As one of China's rare full-function GPU leaders, Moore Threads has achieved rapid iteration across four generations of GPU architecture under its "five chips in five years" plan. The company has built a comprehensive product matrix covering artificial intelligence, scientific computing, and graphics rendering, fully empowering cloud, edge, and end-to-end scenarios.

The KUAE intelligent computing cluster built on the MTT S5000 platform has reached internationally advanced levels in training very large models, providing a solid computing foundation for autonomous driving large-model training. Moore Threads' unique graphics rendering strengths also support autonomous driving simulation, scenario reconstruction, and high-fidelity visualization, enabling efficient full-link loops from R&D to testing and validation.

Pony.ai's Autonomous Driving Technology

Pony.ai has built industry-leading L4 autonomous driving capabilities with its entirely self-developed PonyWorld world model and Virtual Driver system. The company has become the first in China to provide fully driverless Robotaxi services across four major tier-one cities.

According to Pony.ai Founder and CEO Peng Jun, the PonyWorld model, based on reinforcement learning, can produce more than 10 billion kilometers of testing data per week and generate hundreds to thousands of variations of "hazardous scenarios." This allows the Virtual Driver to evolve through repeated training, enabling safety performance that far exceeds human driving.

Commercialization Progress and Future Plans

The strategic cooperation closely aligns with Pony.ai's clear L4 commercialization roadmap. As of December 31, Pony.ai's Robotaxi fleet had reached 1,159 vehicles, exceeding annual targets. The company's seventh-generation Robotaxi system achieved positive unit economics in Guangzhou, marking a key breakthrough in commercial closure.

Pony.ai aims to scale its Robotaxi deployments to more than 3,000 vehicles by the end of 2026. The partnership with Moore Threads is expected to improve L4 autonomous driving model training efficiency and overall system performance, accelerating progress toward that scaling goal.

Industry Impact and Significance

Moore Threads Founder, Chairman and CEO Zhang Jianzhong emphasized that the strategic alliance with Pony.ai represents the first deep collaboration between domestic full-function GPU computing power and leading autonomous driving algorithms. He noted that this partnership will fully showcase the excellent performance and reliability of Moore Threads' GPUs in high-intensity AI training and complex physical simulation scenarios.

The collaboration represents a key step in supporting national strategic emerging industries with independent computing power. By combining Pony.ai's advanced autonomous driving algorithms with Moore Threads' domestically developed GPU computing capabilities, the partnership aims to create a comprehensive autonomous driving computing ecosystem that can accelerate the commercialization of L4 autonomous vehicles in China and potentially beyond.

The partnership between these two Chinese technology leaders demonstrates the country's growing capabilities in both autonomous driving technology and domestic semiconductor development, potentially reducing reliance on foreign computing hardware while advancing the commercialization of autonomous vehicles.