Last week, Shanghai hosted China’s largest AI event: the World Artificial Intelligence Conference (WAIC), with 500 exhibitors, 1,500 products on display, more than 300,000 attendees, and even an appearance by Chinese Premier Li Qiang.
But despite its impressive scale, the conference left me disappointed. I was hoping to see the industry’s technological advances. Instead, WAIC confirmed my suspicions: There is a gap between what China’s AI can do and the cutting-edge innovations emerging from Silicon Valley.
WAIC exhibitors focus on robotics and large language modeling (LLM), with only a few generative AI companies among them. More than half of the companies at WAIC, including major tech companies and even some state-owned telecom companies, are showcasing their new models.
In Shanghai, Baidu founder Robin Li encourages attendees began developing practical AI applications instead of continuing to refine their LLMs. He emphasized that a powerful and widely used AI application would benefit society more than another model that could process huge amounts of data but had no practical applications.
The AI applications on display in Shanghai were mostly chatbots like ChatGPT, with the exception of Kuaishou. Text to Image Application Kling, a Sora-like product that I found really impressive.
As I walked around the showroom, I noticed that most of the chatbots asked for prompts in English, rather than Chinese, which made me suspect that many of China’s AI programs are actually running on models developed outside of China.
It’s clear that the models still need some tweaking. One consumer suggested a text-to-image app from Moore Threads featuring “a cute little boy with brown hair, sitting in a garden.” The result was a baby with rosy skin, eyes that didn’t line up with his face, and a disproportionately small body.
I leave the conference in agreement with Alibaba Chairman Joe Tsai’s Candid Admit China’s development of artificial intelligence was at least two years behind the United States at the beginning of this year. That means US and Chinese companies are not really on the same level, so it’s difficult to make direct comparisons.
The key issue is that Chinese LLMs are limited in their use of data within the Great Firewall. As an investment bank Goldman Sachs noted late last year“LLM performance improves with scale—more parameters, more good training data, more training runs, and more computations.” There is simply less information on the isolated Chinese internet than on the open internet with sources in many different languages.
AI companies outside of China have more data to use for training purposes. An AI developer in China would struggle to keep up.
The limitations of limited access to advanced GPUs are also evident. US policies Restrict Access to advanced chips and chip manufacturing technology will mean Chinese companies lag their non-Chinese friends.
Yet despite these limitations, Chinese AI developers are still looking for opportunities to innovate.
A lot of strong talent from the country’s mature consumer tech ecosystem is moving into AI. Most of the founding members of “Four Tigers”—Baichuan, Zhipu AI, Moonshot AI, and MiniMax—have all worked at major tech companies. Their strong consumer and product intuition is why they are now leading China’s AI application space. From a consumer perspective, their products are on par with many leading US applications.
There’s also progress on the hardware front. Huawei’s Ascend AI processor, in particular, looks like it could be miles ahead of competitors. The Chinese tech giant, which currently uses chips made by SMIC, claims its Ascend 910B AI chip can outperform Nvidia’s A100 chip in some tests, especially when using large AI model training.
China’s AI developers face several fundamental hurdles, such as a challenging environment, a lack of advanced chips, geopolitical isolation, and national security concerns that limit the mobility of talent and capital.
Together, these restrictions will create two parallel AI ecosystems: one inside China and one outside China. The United States will maintain its lead in developing this transformative technology.
But just because the United States has a technological edge doesn’t mean China’s AI developers will be left behind. Chinese companies have always started out a step behind their non-Chinese rivals, but fierce competition and a willingness to experiment have helped them catch up with — and in the case of consumer internet companies, even surpass — the rest of the world.
In the world of AI, the United States and China are both friends and rivals. We should hope that their geopolitical rivalry does not stifle innovation and collaboration.
Opinions expressed in commentary on Fortune.com are solely those of the author and do not necessarily reflect the views and beliefs of Fortune.
