The AI race – the quest for performance or practicality

I had lunch today with a friend who’s really into AI. He invests in this sector, follows it closely, and uses it a lot in his private office. Naturally, we ended up talking about the ongoing “AI race” between the US and China—something that’s been popping up a lot in the media lately.

He agreed with a view I’ve come across many times: the US seems focused on making AI as big and powerful as possible, while China is more interested in getting it into people’s hands early and making sure it’s widely used.

The US has a long history of major tech achievements. Think NASA’s missions to the moon, the space shuttle, or building a world‑leading military. It’s no surprise Hollywood has so many American stories behind films like Apollo 13 and Oppenheimer. Big, bold projects seem to be very much part of the culture.

And the same mindset shows up in American tech companies. Many of the world’s largest, most influential tech firms are American—and they often aim straight for scale and speed. Chips? Make them as fast as possible. Models? Push for maximum performance. Data centres? Build them with enormous storage—and maybe even put them in space if you’re Elon Musk.

China, on the other hand, seems to be taking a different path. Sure, some of that is shaped by US export restrictions on high‑end chips and tech. But I think there’s also a cultural element. Chinese innovation often leans toward practicality. A good theory is great, but a good theory that helps more people make money? Even better.

This is probably why so many Chinese AI models are open source. Part is due to government policy. But it also gives everyone—from developers to everyday users—the chance to start building with AI early. That’s why AI is already showing up in phones, tablets, and even cars across China. The priority is clear: improve user experience at low cost. Lower cost means more users. Better user experience means more business.

Interestingly, Anthropic’s recent move in releasing tools like Cowork—also open sourced—feels like it’s nudging in the same direction. Cowork lets people automate tasks and workflows without needing to code. That’s practical. While many AI companies focus on attracting end‑users directly, the more pragmatic approaches—whether in China or at firms like Anthropic—might be what drives real productivity gains.

That’s something worth thinking about when looking at AI investments in 2026.

James Chu, CFA, Head of Investment Solutions