Autonomous Driving Is Becoming AI’s Next Frontier: QCraft’s $100M Bet Explains Why

As AI shifts from digital systems to real-world applications, autonomous driving is emerging as a key testing ground for next-generation models

QCraft, an autonomous driving technology company founded in Silicon Valley, has raised $100 million in a new Series D round. The funding was backed by a group of investors including Ningbo Ninghai Xingtaihe Fund, Wonderland Capital, and Liangxi Science and Innovation Industry Investment Fund, along with strategic participation from an automotive OEM and an electronics components supplier.

The company said the capital will be used to accelerate research in areas such as world models and reinforcement learning, while also expanding its global talent base. While funding rounds in the autonomous driving sector are no longer uncommon, the emphasis on “physical AI” marks a notable shift in positioning—aligning QCraft with a broader industry narrative around real-world AI systems.

QCraft is framing its next phase of growth around the idea that autonomous driving represents a direct pathway into physical-world intelligence.

Chairman and CEO Dr. James Yu described the transition as a move beyond current AI capabilities:

“We are moving from systems that mimic human-like intelligence toward ones that can surpass it. Over the next decade, much of that progress will happen in the physical world—and autonomous driving is one of the most direct ways to get there.”

At a technical level, this direction is reflected in the company’s focus on world models and reinforcement learning—two areas gaining traction across advanced AI research. World models aim to help AI systems build internal representations of how the real world works, while reinforcement learning enables them to make decisions through trial and feedback. Together, these approaches are seen as critical for systems that must operate in unpredictable, real-world environments.

In this context, autonomous vehicles are not just applications—they are continuous learning systems, generating large volumes of real-world data that can be used to train increasingly capable models.

Scaling Deployment in a Competitive Market

QCraft’s technology is already deployed at scale. Its QPilot system has been integrated into more than one million vehicles across nearly 30 production models, in partnership with close to 10 automotive manufacturers.

The company’s QPilot Pro product, which enables urban Navigate on Autopilot (NOA) on a single 128 TOPS chip, has also gained industry attention for its efficiency and performance.

Beyond passenger vehicles, QCraft has established a presence in autonomous logistics, with commercial deployments operating in cities such as Jinhua, Wuhu, and Ningbo. The company has also outlined plans to expand into robotaxi services, with pilot programs expected in 2026 and broader rollout targeted for 2027.

These milestones highlight a key distinction in the autonomous driving landscape: while some players remain focused on research and testing, others—including QCraft—are prioritizing real-world deployment and iteration.

Autonomous Driving as an AI Testbed

The broader significance of QCraft’s strategy lies in how autonomous driving is being repositioned within the AI ecosystem. Unlike generative AI applications, which operate primarily in digital environments, autonomous systems must navigate physical constraints—uncertainty, safety, and real-time decision-making. This makes them a more complex, but also more meaningful, benchmark for advanced AI capabilities.

In China, where large-scale testing and deployment are more feasible, companies have been able to accelerate this transition from research to application. This has created a competitive dynamic where progress is measured not just by model performance, but by real-world validation.

QCraft’s focus on physical AI places it within this emerging category of companies attempting to bridge the gap between simulation and reality. As AI continues to evolve, the center of gravity is gradually shifting from digital intelligence to systems that can operate in the physical world. QCraft’s latest funding round underscores this transition. While the concept of “physical AI” is still taking shape, autonomous driving is increasingly seen as one of its most immediate and practical applications.

The next phase of competition in AI may not be defined solely by model capabilities, but by how effectively those models can interact with—and learn from—the real world.

Quick Takeaways:

• QCraft raised $100M in Series D funding, signaling continued investor interest in autonomous driving and real-world AI systems.
• The company is positioning itself around “physical AI”, reflecting a broader shift from digital AI (chatbots, content) to real-world applications.
• Autonomous driving is emerging as a key testbed for next-gen AI, especially with technologies like world models and reinforcement learning.
• QCraft already has significant scale, with its technology deployed in over 1 million vehicles across multiple OEM partnerships.
• Focus is shifting from research to real-world deployment, particularly in areas like urban driving and autonomous logistics.
• China-based players are accelerating adoption, benefiting from faster real-world testing and commercialization environments.
• The “physical AI” narrative is still evolving, but it points toward AI systems that can perceive, decide, and act in complex environments.
• Autonomous logistics and robotaxis remain key growth areas, with pilot programs and expansions planned in the coming years.
• Big picture: The next phase of AI competition may be defined not just by model performance, but by how effectively systems operate in the physical world.