The New Cold War: How China and the USA are competing for control of the AI market
A BCG study reveals the technological battle between the two superpowers, with huge gaps in investments and computing power alongside a Chinese advantage in costs and implementation, in an attempt to understand what the future of artificial intelligence will be.
The competition between the USA and China in the field of artificial intelligence has turned from a struggle over developing the most advanced model into a broad technological war. According to a new study by the strategic consulting firm BCG, both superpowers are currently building completely separate AI systems, ranging from chips and infrastructure to models and applications.
The USA leads primarily due to massive investments, computing power, and infrastructure. Since 2023, American startups in the field have raised about 380 billion dollars, compared to only 39 billion dollars in China. Additionally, in 2024, the 20 largest American technology companies invested about 318 billion dollars in research and development, compared to 68 billion dollars for the Chinese companies examined.
There is also a significant gap in the field of computing. According to the study, American computing power adapted for AI is eight times greater than that of China, and the output of data centers in the USA stood at more than 50 gigawatts at the end of 2025, compared to 31 gigawatts in China.
On the other hand, China is trying to close the gap through cheaper models and rapid implementation. For example, the cost of using the Chinese model Kimi K2.6 was 1.71 dollars per million tokens, compared to 11.25 dollars in the American model GPT 5.5. Furthermore, between 2020 and 2024, China was responsible for 78% of AI patent families worldwide, compared to only 7% in the USA.
The split between the superpowers also creates a new problem of technological incompatibility. Systems developed in the American environment, such as NVIDIA's CUDA, do not work directly with Chinese chips like those from Huawei, which makes it difficult for companies to switch between systems.
According to BCG, the decision of which AI system to use will become a geopolitical decision in the near future, not just a business one. International companies may be required to operate two different systems to operate in both Western markets and in China.
The study notes that despite the American advantage in investments and infrastructure, the high centralization of the AI industry in the USA also creates a risk, due to the interdependence between chip companies, cloud services, and artificial intelligence laboratories.
