Soaring Autonomous Driving R&D Costs Push Automakers to Outsource

Facing high R&D costs and technological uncertainties, an increasing number of automakers are opting to outsource their intelligent driving systems. Wang Xianbin, Vice President of Gasgoo Automotive Research Institute, noted that aside from a few companies like Tesla, XPeng, and Li Auto that continue to pursue in-house development, suppliers such as Huawei Qiankun, Momenta, and Horizon Robotics are becoming the "new soul" of intelligent driving. In-house development demands massive investment: top-tier algorithm teams alone cost over RMB 100 million annually, while data collection, computing infrastructure, and safety testing represent bottomless expenses. Moreover, user trust in advanced intelligent driving remains low—only 20% fully trust these systems, with safety concerns cited by 60% of users. Against this backdrop, Huawei is projected to equip 3 million vehicles with its intelligent driving solutions by 2025, while Momenta has secured定点 (design wins) for more than 130 vehicle models. Automakers like Chery and BAIC are adopting multi-supplier strategies to enable flexible configuration. The diversification of technical approaches—including end-to-end architectures, Vision-Language-Action (VLA) models, and world models—further amplifies the risks of in-house development, making data and computing power critical barriers to entry. Going forward, automakers will shift their roles toward “product definers” and “experience integrators,” focusing on differentiated competitive areas such as intelligent cockpits, while leaving intelligent driving to specialized partners—ushering the industry into a new era of efficient collaboration.