The US is still ahead of China in artificial intelligence. However, perhaps the key to China’s success lies in its open-source model ecosystem, combined with aggressive development in semiconductor design and manufacturing. Our world is not static, and the world of artificial intelligence is where momentum matters.
The Power Curve
Something small today on a power curve of high compounded growth can quickly become an unstoppable force. China’s hypercompetitive business environment and rapid distribution of knowledge are the keys to developing extraordinary momentum.
The US’s AI Action Plan, released this week, champions open-source development. This is a good step for the United States, but AI isn’t a single monolithic technology. Different stages of growth exist globally. The United States has scaled AI infrastructure, enabling a significant advantage in artificial intelligence implementations. But, is it sustainable? Can the US afford the energy bill?
AI technology will continue to progress. But, there is no single finish line. It’s not a single, monolithic technology that needs to be developed and completed.
The AI Flywheel
Advantages in AI translate into economic growth and reinvestment – and the flywheel continues. One advantage cycles into another, and this AI flywheel forms the foundation of ongoing economic power, interdependence, and competition.
Specifically, regarding the open-source flywheel, China has an edge, as evidenced by DeepSeek, Alibaba, and other Chinese companies; many of their offerings surpass those of Google and Meta. A small advantage compounds into an insurmountable lead if China’s open-source development continues to outperform.
This is understandable, as US companies tend to take a more proprietary and secretive approach to their foundational models. They spend enormous amounts on capital and recruiting. While knowledge circulates, it is slow and expensive to disseminate.
In China, its open AI ecosystem is more competitive, driving down pricing, and intense competition is developing stronger, competitive entities with a quickly rising power curve. That keeps China’s AI flywheel spinning faster and is the US’s most significant vulnerability to global AI dominance.
The Software and Hardware Curves
The US’s software products are superior, and its reinvestment in software advantage is substantial, potentially sustainable, and the U.S.’s greatest single competitive advantage. However, while Chinese AI software products may still lag behind those of the US, the slope of the Chinese curve and rate of development are getting steeper, which may erode the US’s global competitive advantage.
The other component to the AI race, hardware, is uniquely competitive. At the same time, NVIDIA’s edge is not only advanced hardware design but also a software ecosystem, facilitated by its CUDA AI hardware development and operating system, which gives it a long-term sustainable advantage not only in performance but also in development.
Chinese companies cannot compete directly with Nvidia and other US hardware designers, but they may be able to leapfrog the technology with a different approach. Alternative architectures may prove effective, but now geopolitics, trade restrictions, and the drive for self-reliance, all combined to create an uncertain but intense competitive environment in hardware technology.
The Zero-sum Game
Rhetoric surrounding AI often implies that it is a zero-sum game. Other general-purpose technologies emerged, benefiting all nations, with no country’s advancement at the expense of others. It’s no different with AI.
AI can potentially be transformative, and although current geopolitical rhetoric does not allow for cooperation or collaboration, AI progress and innovation are ultimately a global collaborative effort. If done correctly, it benefits many more and it does not come at the expense of any one nation.
That should be the AI Action Plan.
