DeepSeek and the Danger of Information Asymmetry in the Global AI Innovation Race

A few days ago, global tech news was suddenly taken over by the emergence of DeepSeek AI, a previously unknown Chinese startup that released an AI chatbot that claimed to have the capabilities of the best that the likes of OpenAI can offer, but developed with a fraction of the cost. The firms providing the tools to the global AI race, most notably semiconductor designer Nvidia, saw their share values tank in response, driven by investors fearing that future improvements in AI can be done without the constant increasing number and sophistication of chips and equipment that their producers previously claimed.

While plenty of third-party observers expressed skepticism about DeepSeek's capabilities and claims about efficiency, the reality is that the stockmarkets had little preparation for its emergence. Investors and tech insiders outside China had little knowledge that Chinese rivals have been working hard, perhaps for years behind the scene, to catch up with Silicon Valley on AI. They assumed American dominance is largely unassailable, at least in the short term, given the bans on exporting high-performance semiconductors to China as part of the ongoing trade war. 

Yet, for months and years, Chinese AI labs, companies, and researchers have meticulously published their work in open forums, with open-source codes freely available for all to use and papers detailing how each iteration improved upon the previous. These publications often had the real names and affiliations of the researchers, making it clear who is doing the work and where it is being done. Had tech insiders beyond China followed the incremental development, they could have priced in the possibility of frugal innovation upending the global AI race.

If there is any lesson that can be learned from the sheer scale of the surprise that greeted DeepSeek's AI model, and the resulting stockmarket rout, it is that in the ongoing race to lead the next stage of AI development, China has consistently gained an advantage in information asymmetry. Whereas Chinese researchers work in Silicon Valley and closely monitor new developments there, Silicon Valley insiders see little need to look at what China is doing, because they assume that nothing coming out of China could possibly be more advanced and thus worth learning about.

It is a mistake that America in particular can ill-afford to continue making. Scanning global tech headlines over the past years, one can easily come to the conclusion that China has become the only whole-the-system challenger to the US in disruptive innovation. While every country has their set of innovative firms, only US and China have been able to come up with entire innovative ecosystems that can entirely upend the lifestyles of people around the world. DeepSeek is just the latest example of China being the only peer competitor of the US in the most cutting edge technologies defining the future.

There will certainly be many more of those that surprise Americans if Americans choose to ignore the Chinese tech scene. While the American public has slowly comes to the realization of how widespread adoption of EVs and mobile payments have revolutionized urban living in China, few yet fathom the scale of Chinese investments in other technological frontiers, from quantum communication, to mass-produced flying cars, to supersonic travel, to nuclear fusion. Yet many of these R&D efforts largely unreported in Western media outlets, when breakthroughs come, they will come again as a massive surprise.

To correct this information asymmetry, Americans have to rely on their own eyes and ears, rather than the little that trickle through mainstream outlets. With anti-Chinese sentiment becoming a social norm among Americans of all colors and stripes, there is little incentive for profit-seeking media to report positively on China's technological prowess without risking backlash from subscribers and advertisers. For Americans to see China, they ought to instead head to China by themselves, as travelers, students, and businessmen working with Chinese collaborators, suppliers, and clients.

Yet, currently there is little sign that the asymmetry will disappear. With as few as 800 Americans studying in China at one time, the US is not prepared to train a new generation of China experts who are able to keep up with the latest development in its main competitor. In contrast, despite the tarnished reputation of China, plenty of Chinese students still aim to study in America, learning not just about American culture, but also gain firsthand experience in how technology is taught and created in the educational and corporate setting. This new generation of America experts in China will certainly give Silicon Valley more surprises in the future.