QCraft Raises $100M to Power Physical AI Leap

QCraft just closed a $100 million round to turn AI into street-ready autonomy.

China’s AI hardware push just got a turbo boost—from a brainy promise to a road-proven reality. The Series D funding brings in a mix of strategic backers, including a leading domestic automaker, Ningbo Ninghai Xingtai He Fund, Liangxi Sci-Tech Innovation Industry Fund (managed by Bohua Capital), Huade Sci-Tech, and a major automotive electronics supplier, among others. Supply chain disclosures reveal the investment’s unmistakable signal: carmakers and AI developers are pooling capital to accelerate a move from simulation to pavement.

Dr. Yu Qian, QCraft’s co-founder, chairman, and CEO, framed 2026 as a hinge year: “the greatest opportunities over the next five to ten years lie in the physical world,” and autonomous driving is a gateway to that realm. The company is recalibrating its bets toward Level 4 autonomous driving while doubling down on world-model and reinforcement-learning technologies. In practical terms, that means QCraft intends to push beyond driver-assistance toward fully autonomous systems that can operate with limited human oversight on real-world streets.

On the product side, QCraft is touting a two-pronged path. Its Chengfeng ADAS, categorized as L2++ in industry parlance, has already surpassed 1 million units deployed, illustrating the company’s mass-production capability and scale. At the same time, the leadership says a world-model plus reinforcement-learning solution will be officially unveiled soon, signaling a major leap in how the company designs perception, planning, and control loops for autonomous platforms. The mix of investors—an automaker, automotive-electronics suppliers, and state-leaning tech funds—reads as a coordinated push from China’s auto and AI ecosystems to translate clever algorithms into road- and factory-floor realities.

The funding also underscores a broader policy-and-market alignment. Mandarin-language reporting indicates that domestic capital is increasingly channeling into “physical AI”—AI that interacts with the real world, notably in mobility and robotics—soaring beyond purely digital feats. This aligns with ongoing national programs to upgrade manufacturing through intelligent connected-vehicles and autonomous systems, with funding and partnerships that blur the line between software, sensors, and vehicle platforms.

From a manufacturing perspective, the scale implied by 1 million Chengfeng units deployed matters. It signals not merely a prototype triumph but a livable, ongoing production capability that can feed a growing ecosystem of suppliers, integrators, and fleet operators. It also raises practical questions for global manufacturers sourcing complex AI-enabled mobility tech: will domestic Chinese players push toward tighter vertical integration, making it harder for foreign AI stacks to displace local ones in China? And how quickly can the same level of autonomous capability be translated into the more stringent safety and regulatory regimes that govern international markets?

Two practitioner takeaways stand out. First, for automakers and Tier 1s, this round reinforces a trend toward “one-stack” AI platforms embedded in vehicle platforms—where perception, decision, and control are co-developed with the vehicle’s hardware and sensors—rather than stitched-together modules sourced from separate vendors. Second, for robotics and manufacturing managers, the emphasis on world models and reinforcement learning points to a future where simulation-to-real transfer becomes a keystone capability: the more data and compute a Chinese player can throw into its world-model loop, the faster it will adapt to new environments on the factory floor and on the road. In short, the QCraft signal is less about a single product and more about a domestic R&D and manufacturing treadmill designed to outpace the hype with repeatable, scalable deployments.

What this means for global sourcing and competition: expect greater domestic consolidation in China’s AI-autonomy stack, with a potentially faster cadence of updates and deployments for vehicles and robots. For rivals, the message is clear—borderless markets notwithstanding, the center of gravity for physical AI hardware and software is tightening around Chinese teams that can couple advanced algorithms with proven mass production.

Sources

QCraft Raises $100M to Accelerate Push Toward General Physical AI