QCraft Secures $100M as Xpeng Enters Robotaxi Market
China's autonomous driving sector received a major boost on March 23, 2026, with two significant announcements signaling a shift toward commercialization. Autonomous driving company QCraft announced it completed a new $100 million Series D financing round. The investment, which included a major domestic automaker and several industry funds, will fund research into advanced world models and reinforcement learning, technologies central to the concept of "Physical AI." QCraft's CEO, Yu Qian, stated the firm is shifting its strategic focus toward Level 4 autonomy.
On the same day, electric vehicle maker Xpeng established a new first-level division dedicated to its Robotaxi business. The unit will manage product definition, development, and operations, with plans to begin demonstration services for passengers in the latter half of 2026. Xpeng intends to leverage its second-generation end-to-end model to launch three distinct Robotaxi vehicle models by 2026, marking a significant commitment to building a vertically integrated autonomous mobility service.
Cost and Policy Breakthroughs Lower Commercial Barriers
The recent industry activity is fueled by critical advancements that have dismantled long-standing barriers. Technologically, the industry is converging on end-to-end large models, which directly translate sensor input to driving commands, simplifying the software stack. This shift, combined with a move toward factory-installed systems over costly retrofits, has driven down expenses dramatically. For instance, rival Pony.ai has reduced the cost of its latest Robotaxi hardware suite by 70%, including a 68% reduction in lidar sensor costs.
Regulatory hurdles are also clearing. In December 2025, China's Ministry of Industry and Information Technology granted the first L3 autonomous driving permits. By mid-January 2026, vehicles from automakers like Deepal had already accumulated over 70,000 kilometers of L3 driving on public roads. Crucially, these pilot programs established that automakers hold primary responsibility during autonomous operation, resolving a key legal ambiguity that had previously slowed commercial deployment.
'Physical AI' Race Escalates with Expanding Robotaxi Fleets
Companies are now translating the "Physical AI" narrative into quantifiable commercial targets. The race to build and deploy large-scale Robotaxi fleets is accelerating. WeRide plans to deliver 2,000 mass-produced Robotaxis with partner Geely by 2026, while Pony.ai aims to expand its own fleet to 3,000 vehicles. These ambitious targets are underpinned by falling production costs, with some analysts forecasting that a Robotaxi's manufacturing cost could drop to $10,000 per unit at an annual production volume of 100,000.
This push for scale highlights different strategic approaches. Xpeng is pursuing a full-stack, self-developed model for its Robotaxis. In contrast, QCraft is establishing itself as a key technology supplier to automakers. The company's "QPilot" assisted driving system has already been deployed in over one million vehicles across nearly 10 OEM partners, with plans to integrate its advanced Navigate on Autopilot (NOA) capabilities into more than 50 new vehicle models in 2026. This parallel progress of both integrated automakers and specialized suppliers indicates a maturing and increasingly competitive market.
