QCraft raises $100M to bet on physical AI and robotaxis
What's the deal? QCraft, a Chinese-American autonomous driving startup, has closed a $100M Series D round. The funding was backed by Wonderland Capital, Ningbo Ninghai Xingtaihe Fund, Liangxi Science and Innovation Industry Investment Fund Partnership, and unnamed strategic investors from a major carmaker and an automotive electronics supplier.
The company plans to channel the capital into physical AI research — specifically world models and reinforcement learning — and into building out its global team.
Why now? QCraft's chief executive, Dr James YuDealroom has a profile for this one. Try Dealroom →, called 2026 "a critical inflection point" in AI, marking a shift from human-level to superhuman intelligence. The company believes autonomous driving is the most direct gateway into physical-world AI — systems that perceive and act in real environments, not just digital ones.
The timing tracks with a broader frenzy. Physical AI startups in China raised over $1.9B across 43 deals in the first two months of 2026 alone. QCraft itself hit a milestone: its QPilot driver-assistance system now runs in more than one million vehicles across nearly 30 models with close to 10 major carmaker partners.
The startup is also reportedly in the confidential filing queue for a Hong Kong IPO.
What could go wrong? QCraft's ambitions are steep. It plans to pilot robotaxis in 2026 and deploy them at scale in 2027 — a timeline the industry has repeatedly failed to meet. The jump from assisted driving to full Level 4 autonomy remains one of the hardest problems in tech, and regulatory hurdles vary widely across markets.
Competition is fierce too. QCraft operates in a crowded field alongside well-funded Chinese rivals and global players like Waymo. Scaling from one million assisted-driving vehicles to fully driverless fleets is a different game entirely.
The signal: This round reflects a clear bet: that autonomous driving will be the launchpad for general physical AI. QCraft's strategy of running Level 2 and Level 4 programmes in parallel — using mass-production driving data to train its self-driving systems and vice versa — is designed to create a flywheel that pure-play robotaxi companies lack.
The broader trend is unmistakable. Capital is flooding into companies that bridge AI and the physical world, from autonomous vehicles to logistics robots. Founded in Silicon Valley in 2019, QCraft has raised from investors including IDG Capital, CICC Capital, and Meituan's DragonBall Capital. With this round, it is positioning itself at the intersection of China's manufacturing scale and frontier AI research.
Sources:
Tech In Asia
36Kr
DealStreetAsia
Gasgoo
BusinessWire
B.S.
