Pony.ai Level 4 Autonomous Driving Partnership With Moore Threads Strengthens AI Computing Ecosystem

Pony.ai Level 4 autonomous driving development took a significant step forward on February 6, when the autonomous driving technology company signed a strategic cooperation agreement with Moore Threads, a Chinese developer of full-featured GPUs. The partnership highlights a growing focus on strengthening domestic computing capabilities to support next-generation autonomous mobility at scale.

Strategic collaboration to advance Level 4 autonomous driving

The cooperation centers on accelerating the large-scale commercial application of Level 4 autonomous driving technologies. Pony.ai and Moore Threads will work closely on training and optimizing Pony.ai’s PonyWorld model and its Virtual Driver system, both of which play a critical role in enabling high-level autonomous decision-making and vehicle control. By aligning algorithm development with high-performance computing hardware, the two companies aim to improve model efficiency, stability, and deployment readiness. This approach is expected to shorten development cycles while maintaining the safety and reliability standards required for advanced autonomous driving systems.

Integrated computing infrastructure for model training and validation

Leveraging domestic GPU and computing clusters

The collaboration will utilize Moore Threads’ MTT S5000 training-inference integrated intelligent computing card along with the MTT KUAE intelligent computing cluster. These platforms will support the joint training, adaptation, and verification of the PonyWorld model as well as in-vehicle autonomous driving models. Through this integrated computing infrastructure, Pony.ai and Moore Threads plan to enhance simulation accuracy, improve real-world model transfer, and optimize computing costs. The initiative is designed to create a unified autonomous driving computing ecosystem that connects algorithms, data pipelines, computing power, and real-world applications, ultimately accelerating the evolution and cost optimization of Level 4 autonomous driving solutions.