The 2026 World Economic Forum Annual Meeting recently concluded. While geopolitical issues continued to dominate the core agenda, the voices of Chinese tech companies also garnered significant attention. During the conference, Li Yifan, Co-founder and CEO of Hesai Group, noted in a conversation that a significant change at this year's forum was the shift in AI discussions from exploring "possibilities" to focusing on practical "implementation." Furthermore, he deeply felt that in the face of a global environment where the "old rule-based order has been broken," an increasing number of countries are adopting a more proactive and pragmatic attitude towards seeking cooperation with Chinese industry.
Li Yifan observed a shift in the phase of AI technology during this Davos trip. He stated that discussions on artificial intelligence at the forum were more concentrated on specific scenarios, such as "which applications will experience explosive growth." "Last year's Davos coincided with the two weeks before DeepSeek's release, and everyone was still discussing the many possibilities of AI, whereas one of the most important topics this year is the implementation scenarios from a technical perspective," he remarked.
Geopolitics remains the "eternal theme" at Davos. This year, with speeches from figures like US President Trump and Canadian Prime Minister Carney, the idea that the "old order is being broken" became a hot topic of discussion. "In the absence of clear rule constraints, how can we still engage in more trade, exchanges, and mutually beneficial cooperation?" He believes this topic is "both profound and difficult to resolve."
Yet, within this complex international political environment, Li Yifan identified a positive change: China's role is being viewed more rationally, and countries that previously held relatively nuanced or hesitant attitudes towards China are now more willing to engage in dialogue. Behind this shift in attitude lies curiosity and recognition of China's development path. "Originally, many people did not particularly recognize China; now some have become very approving, while others have at least become very curious," he sensed, adding that many countries want to understand "What exactly did China do right? In which areas might they be at a disadvantage later if they don't cooperate with China?"
A key objective of Li Yifan's visit was precisely to objectively showcase the actual achievements already made by China's smart vehicle and AI industries to the world, and to encourage various parties to shift their thinking from "defensive regulation" to "how to cooperate and embrace the technology."
So, what exactly have China and Chinese companies done right? "People might think it's cost, but actually, it's not," Li Yifan pointed directly to the core issue. "Cost is the final step; it's not even a cause, but rather an end result. Because with enough customers and sufficiently large production volume, costs naturally become lower."
What truly impresses overseas clients is, first, the mass production capability and quality system validated by the Chinese market; second, the company's consistently leading R&D investment and financial health, which gives clients confidence that they are partnering with a company for the long term; and finally, the cost advantages derived from scale. "Clients will examine the company's overall operational status, including market share, gross margin, net profit margin, cash on hand, etc., feeling assured because, in the long run, this is how one can sustainably outperform the competition," he explained.
Behind this sense of assurance lies the deep barriers constructed by China's unique industrial ecosystem. Li Yifan believes that the reason Hesai Group possesses such strong competitiveness compared to overseas competitors is because it has ridden the "tailwind" of China's smart vehicle industry development. From an enterprise perspective, Li Yifan emphasized that China possesses a solid manufacturing foundation and efficiency advantages in its supply chain. "The US holds more advantages in services and software, while China can build a very robust industrial chain and supply chain system, which is not only large in scale but also deep. In this aspect, China has performed far better than the US." This ecosystem, built upon a vast market, a complete supply chain, and rapid iteration capabilities, is difficult for overseas competitors to relocate or imitate.
Using the automotive and LiDAR industries as examples, he noted that the cost efficiency and time efficiency at every layer of the industrial chain are exceptionally high. In regions like the Yangtze River Delta and Pearl River Delta, "people can communicate and understand things within one or two hours, with iteration speeds measured in days, or even half-days"—a pace "unheard of" in the West. "I sometimes joke that by the time we've developed the product, you're just replying to the email," Li Yifan stated frankly, "At such a speed, obviously we will progress faster."
Beyond hardware advantages, the simultaneous enhancement of China's software capabilities has fostered stronger integrated innovation. "Smart cars and robots are both results of the combination of hardware and software." He believes that the best hardware, combined with deep integration and iteration with AI software, means "China might have more opportunities than the US."
For internationally competitive Chinese companies like Hesai, how can they better expand overseas in the current environment? Li Yifan proposed the need for "more communication" + "ecosystem synergy." He admitted that many current constraints "indeed stem from a lack of understanding," such as concerns over privacy and data security. The solution lies in "more communication." This communication operates on two levels: first, at the technical level, proactively explaining data security and privacy protection issues to resolve misunderstandings; second, at the strategic level, showing the world that the rise of China's strategic industries has become an undeniable fact.
Ecosystem synergy makes overseas cooperation for Chinese companies simpler and smoother, as evidenced in Hesai's collaboration with NVIDIA. Previously, the typical collaboration model for overseas automakers involved independently selecting and integrating various suppliers, developing a system, and then providing it to NVIDIA. What NVIDIA received was the automaker's choice, which wasn't necessarily optimal for NVIDIA's system. Now, NVIDIA's new strategy with the "NVIDIA DRIVE AGX Hyperion 10 Platform" is to provide a pre-validated, mature, complete reference system with sufficient data, models, and completed adaptation.
Li Yifan explained, "For automakers, there is a strong incentive to accept the system recommended by NVIDIA. Because if an automaker doesn't accept it and introduces a new system, they would need to recollect data, train models, and change recommended configurations, which could take two to three years and incur more costs." While this model may reduce the degree of customization, it significantly enhances efficiency and reliability, "finding a more efficient way for NVIDIA and the entire automotive industry."
NVIDIA's new strategy is also positive news for Hesai Group's expansion in overseas markets. At the CES exhibition held in January, Hesai was selected by NVIDIA as the LiDAR partner for the "NVIDIA DRIVE AGX Hyperion 10 Platform," enabling Hesai to transition from independently dealing with various automakers to efficiently reaching global customers through the NVIDIA ecosystem.
As a safety-critical component that helps measure the speed, distance, and shape of objects, the performance reliability of LiDAR is a core industry concern. Li Yifan admitted that while "zero false positives" is mathematically impossible, getting infinitely close to zero is the company's "obsession." "A false positive probability of 0.0001, no matter who it affects, is a quite dangerous thing," he stated. Previously, Li Yifan has mentioned in multiple interviews that invisible metrics like LiDAR's non-false-positive rate, non-missed-detection rate, and high safety would become truly insurmountable, systemic chasms. Hesai's product released last year has pushed safety standards to "physically zero false positives." So, how does the LiDAR industry build a competitive "moat" around safety standards?
Li Yifan believes this requires a quality system that runs through the entire product lifecycle. "Quality is more important than manufacturing," he emphasized. Behind this lies a quality system honed by the stringent demands of China's smart vehicle industry, continuous optimization and improvement based on feedback from massive real-world scenario data, and vertical integration capabilities achieved through self-developed chips and self-built factories.
Mass production capability was one of the key factors in Hesai winning major orders from top European OEMs. "We shipped over 1 million units last year," he revealed, "so [clients] realized: rather than choosing a local partner that hasn't truly achieved mass production yet, it's better to choose a partner that, although a bit farther away, has been fully validated."
How can competitiveness be maintained in the long term? Li Yifan stated, "It's impossible to avoid the emergence of competitors; one must understand that new competitors will always continuously appear in the market." However, he indicated that sustained high R&D investment, respect for the market, and the persistence in perfecting everything from patents to R&D direction constitute Hesai's competitive barriers. According to the "Automotive LiDAR Patent Landscape Analysis - 2025" report published by the globally renowned intellectual property strategy consulting firm KnowMade, Hesai ranks first globally in the LiDAR industry in terms of the number of patents and first in comprehensive patent strength.
Regarding the future of technology, he prefers to view LiDAR as "the infrastructure for AI." "Everything will move in the future," he cited the view of NVIDIA CEO Jensen Huang, "'Anything that moves will be autonomous,' so it will most likely need a 3D sensor."
In the field of robotics, Li Yifan believes that in the next two to three years, robot applications in specific scenarios will continue to scale up, such as those for food delivery, cleaning, and logistics transport. Within the next five to ten years, general-purpose "humanoid" robots will become the main trend. "This type of robot might not necessarily look completely like a human, but its hardware is not limited to one industry; it can do all kinds of things and be applied across different industries. You just need to input new data and train its model, and it can perform different tasks." He compared this model of standardized hardware and differentiated software to how people use iPads. "The hardware form of the Pad is the same, but different software runs on it."
Comments