THE AI wars have begun. For years, America sought to hold China back when it hindered access to its semiconductors, and more recently, its AI chips. The plan failed spectacularly this month when Chinese AI company DeepSeek took the tech world by storm with its own model that rivals the performance of the best that ChatGPT has to offer, developed at a fraction of the cost. Since its launch on Jan 10, the app has shot to the top spot on the Apple App Store. The real ripple it has created, however, is wiping nearly $1tr off the market capitalisation of major US tech firms as investors reassess the global AI playing field. The writing had been on the Great Firewall, for those who bothered to read it. While US officials fretted over Taiwan Semiconductor’s fabs, China’s tech giants were busy perfecting the art of algorithmic efficiency. DeepSeek’s success has shown us how constraints can breed innovation. Denied access to Nvidia’s most powerful AI chips under US sanctions, Chinese engineers optimised older, less powerful hardware to train DeepSeek. According to the Financial Times, DeepSeek developed its model in just two months with an investment of under $6m, while OpenAI reportedly spends over $5bn annually.

This technological leapfrog has profound implications. The notion that American sanctions could meaningfully delay Chinese AI development now appears naive. More concerning for Western policymakers is the possibility that China’s AI sector has evolved to be more resilient precisely because it was forced to innovate under constraints. The result is a new world order in artificial intelligence. For the West, the challenge remains in maintaining its technological edge, while engaging with a rival whose capabilities can no longer be dismissed. What might surface tomorrow is not known at this point. The only certainty is that America’s strategy of containment has failed. The dragon, it seems, is finally spreading its wings.

Published in Dawn, January 31st, 2025