Editorial | China’s low-cost AI is showing up America’s high-cost tech model
The real China threat is in exposing America’s AI myths by showing what can be achieved at much lower costs and with greater efficiency
Whenever an American breakthrough in artificial intelligence (AI) is announced, China is not far behind. That has long alarmed policymakers and industry insiders in the United States. Whether by handicapping China with dodgy trade sanctions and export restrictions with its allies, or handing out billions of taxpayer dollars to favoured industry players, the US just can’t shake off its Chinese competitors.
At a comparatively low cost and with less computing power, DeepSeek’s V3.2-Speciale matches the Gemini 3 Pro in reasoning capabilities and offers efficient results. It has scaled up less powerful chips and made them work together to rival the most advanced and expensive ones that power American models. Chinese AI developers have been able to do more with less, which amounts to a fundamental challenge to the development and investment models of AI in the US.
The American notion is that AI is capital-intensive, perhaps more so than any other US industry today. This has meant the need for trillion-dollar investments to develop the most advanced chips and build data centres that eat up water and power supplies. Some of those investments have led to accusations of circular or self-reinforcing deals. The fear is that the AI investment craze has become a bubble waiting to burst. Indeed, the US stock market has been held up by a handful of AI-led tech firms with trillion-dollar valuations. Whether they are really worth that much is anyone’s guess.
The real China threat here is that it has exposed America’s AI myths by showing what can be achieved at much lower costs and with greater efficiency, while offering the technology as open source. Perhaps China has the US chip war to thank for developing its own world-class AI industry.