Major US technology companies spent more than $ 350 billion on AI-related infrastructure last year, with expectations that the volume of spending will exceed $ 400 billion by 2026, a level far exceeding the investments of any other country, led by China, which is estimated to spend only about $ 100 billion.
Supporters of this trend in the West argue that the large spending reflects the strength of American companies and the depth of capital markets, giving the United States a decisive advantage in what is described as an existential race over the future of artificial intelligence.
A comprehensive bet versus a more balanced strategy
In turn, this approach raises fundamental questions about whether an excessive focus on artificial intelligence may lead to winning this particular race, versus losing a broader race for long-term economic superiority.
While the United States has chosen to bet almost entirely on artificial intelligence, China is pursuing a more diversified strategy, spreading its investments across multiple areas, including electric cars, batteries, robotics, solar and wind energy, as well as advanced manufacturing. Although these sectors are less glamorous, their returns are less risky and more predictable.
Although Beijing describes artificial intelligence as a” national strategic priority, " its biggest spending is on clean energy technologies and manufacturing. In 2024 alone, China invested approximately $ 940 billion in renewable energy projects, electric grids, and energy storage, a figure that far exceeds its investments in artificial intelligence.
Silicon Valley and the general artificial intelligence mani
China treats artificial intelligence as a supporting tool for industry, not as an end in itself to create “super intelligence, a goal that all leading technology companies have in mind.
In contrast, Silicon Valley seems to be more fascinated by artificial intelligence, as business goals overlap with semi-philosophical perceptions about general artificial intelligence (AGI).
A Western trend believes that achieving advanced artificial intelligence is the highest goal of humanity, with a conviction of the continued exponential progress of technology, although technological history often indicates a slowdown in innovation after the first breakthroughs.
Observers warn that the monopolistic and concentrated nature of the US technology sector increases the dangers of “herd thinking”, with huge spending decisions concentrated in the hands of a limited number of companies, amid ongoing uncertainty about the real economic return of artificial intelligence. This trend coincides with the decline in American investment in clean energy and scientific research.
Less expensive models and a more pragmatic approach in China
In China, companies and the government rely to a greater extent on simpler, less expensive, and often open-source AI models, compared to advanced and closed American models. This approach is seen as more flexible and less costly, with a focus on practical uses rather than the pursuit of major theoretical breakthroughs.
The figures indicate that this approach is beginning to bear fruit: China produces about 70% of the world's electric cars, controls 80% to 85% of the global manufacture of solar panels, in addition to more than 75% of battery production.
Analysts believe that the rhetoric of the” artificial intelligence race " also serves the interests of technology companies in the United States, as a political and economic pressure tool to justify unprecedented spending levels, and warn that any regulation or government intervention could weaken the competitiveness of the United States.
In the end, the question remains open whether artificial intelligence is really the most important path to economic prosperity in the Twenty-First Century, or whether a unilateral bet on it may turn into a major risk. Of course, the answer is still unknown, and it may not become clear until years later, when it becomes clear whether this huge investment achieves a long-term vision or is just an expensive technical obsession.