AI Trends

Robin Li: AI penetration at the application layer is lower than expected; if unaddressed, it could become a bubble

WallstreetcnUpdated Nov 17 20:19

AI summary

Speaking at the 2026 APEC CEO Summit, Robin Li said that AI penetration at the application layer is lower than expected, and that many enterprises have seen no clear improvement in revenue or profit after adopting AI. If this problem is not solved, AI could become a bubble, and the most urgent task is to find genuinely suitable application scenarios. Citing autonomous driving as an example, he said its safety is about 10 times higher than that of human drivers, that Apollo Go has expanded its open-road coverage in Dubai, and that it is testing in London in partnership with Uber and Lyft. He also mentioned that Baidu AI Cloud grew at close to 100% in the most recent quarter.

Body · AI translation

On November 17, the first day of the 2026 APEC CEO Summit was held at the Qianhai International Convention Center in Shenzhen. Robin Li, founder, chairman and CEO of Baidu Group, took part in two dialogues that day, one in Chinese and one in English. Around how AI can truly be implemented across the Asia-Pacific region and how the regional economy can build a new landscape, Li systematically shared his own judgments.

Li said the value of AI ultimately comes from the application layer. He noted that in recent years AI technology has made tremendous achievements at the levels of chips, cloud computing and large models, but in his view the power of AI must ultimately be unleashed through the application layer—only by truly applying AI to the real scenarios of each country, each company and each type of organization can technological capability be converted into practical value.

Li believes that autonomous driving is a typical example at the application layer. More than 1 million people worldwide die in road traffic accidents each year, and autonomous driving technology is about 10 times safer than human drivers, making it one of the applications that best demonstrates the value of technology. Apollo Go has continued to expand the range of open roads in Dubai, is conducting tests in London in cooperation with platforms such as Uber and Lyft, and testing in Hong Kong, China has also been underway for several years. "We hope more countries and economies will open up their roads so that this technology can truly be enjoyed by more people."

Speaking about cross-border cooperation, Li proposed a simple principle: focus. Companies must not spread their ideas, strength and efforts too thinly, and the same applies to cooperation. Cooperation is not about relying entirely on partners, but about each side leveraging its own strengths to accomplish something together.

He also mentioned that there is currently an obvious gap among different economies in AI development. The existence of risks is not a reason to stop moving forward; instead, policy guidance should be used to minimize risks and enable all parties to benefit. On computing power and infrastructure, he pointed out that many countries already have a foundation in computing power, but hardware and supporting facilities have not kept pace. "The global circulation of tokens" remains an unresolved issue, and all players in the industry need to optimize technology so that computing resources can be matched more efficiently worldwide.

In the dialogue, Li used "most urgent" as his keyword for 2026. He explained that over the past three years or so AI technology has iterated extremely fast, investment in all aspects has grown exponentially, and prices for computing power and cloud resources and token consumption are all rising rapidly, yet AI's penetration at the application layer is "lower than expected." Many companies have seen no obvious increase in revenue or profit after adopting AI. "If this problem is not solved, AI could become a bubble, so the most urgent task is to find application scenarios that truly suit AI."

Asked how small and medium-sized enterprises can "cross the threshold," Li pointed to organizational capability. He drew an analogy with the Industrial Revolution: once electric motors had been invented and installed in factories, but before the assembly line was invented, electricity's contribution to GDP was actually quite limited; today the capabilities of large models are already strong, but if a company's organizational approach cannot keep up, it will likewise struggle to truly benefit. Drawing on Baidu's practice, Li pointed out three principles for an intelligent organization. First, break down job boundaries; second, accumulate tacit knowledge; third, form a closed loop of workflows.

Discussing the industry landscape in the AI era, he pointed to a change that is taking place: AI is pushing companies like Baidu from asset-light operations toward asset-heavy ones. Computing power is tight everywhere in the world, and almost every company needs intelligent cloud services. In the most recent quarter, Baidu's intelligent cloud growth was close to 100%.

Looking ahead to regional coordinated development under the APEC theme of "Building a Sustainable Future for All," Li said that in the AI era the opportunities for cooperation among countries and companies have not decreased but clearly increased. He said that in almost every scenario, combining AI, agents and data can immediately deliver improvements in efficiency and effectiveness, and there are a large number of cooperation opportunities between companies and countries; all parties should use policy guidance to narrow the AI divide so that all economies can achieve progress in a balanced way.

In Li's view, the key to bridging the divide is not to set limits on one another, but for each side to bring out what it does best and accomplish things together. At the end of the roundtable discussion, Li answered with one sentence: "Humanity's pursuit of intelligence is infinite; let us all adopt the most advanced artificial intelligence as early as possible."

Source: The Paper Risk Disclosure and Disclaimer

Markets carry risk, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation or needs of individual users. Users should consider whether any opinions, views or conclusions in this article are suitable for their particular circumstances. Investment based on this article is at the user's own risk.

Source

Wallstreetcn