Hesai Bets on Spatial Intelligence to Avoid Being Trapped in the Auto Cycle

Author | Chai Xuchen

Editor | Zhou Zhiyu

While domestic automakers are still locked in price wars over "intelligent driving democratization," LiDAR companies have already begun looking for their next ticket.

On April 17, Hesai released its 6D full-color LiDAR super-sensing chip and announced that its corporate strategy would upgrade from "spatial perception" to "spatial intelligence." This means enabling machines not only to see their surroundings but also to understand spatial relationships and act autonomously.

In line with this shift, Hesai also unveiled Kosmo, a spatial intelligence hardware product aimed at the consumer market, as well as robot power modules serving as the core foundation of the physical AI era.

This company, previously highly tied to the automotive supply chain, is now urgently seeking to transform its identity.

Over the past few years, Hesai has benefited from the rise of China's new energy vehicle (NEV) sector. Companies such as Li Auto, XPeng, NIO, Great Wall Motors, and Leapmotor have densely equipped their vehicles with LiDAR, helping Hesai quickly become one of the global leaders in shipment volume. This success also convinced capital markets that it is the "seller of shovels" in the intelligent car era.

However, the story of selling shovels has its ceiling.

The automotive market is ultimately a cyclical industry. Automaker price pressure, platform-based procurement, and diverging intelligent driving routes are all squeezing profit margins for suppliers. As LiDAR gradually shifts from a high-end option to a standard feature in mid-range models, the industry's competitive logic has switched from "technical scarcity" to "cost efficiency."

Against this backdrop, Hesai turning its gaze beyond the automotive sector comes as no surprise.

From Selling LiDAR to Selling Spatial Understanding

The LiDAR industry once enjoyed a golden age.

At that time, the market believed that autonomous driving absolutely required high-precision 3D perception, with LiDAR being the optimal solution. Consequently, numerous startups flooded into the track, driving valuations skyward. Companies like Hesai, Huawei, RoboSense, and Innoviz Technologies emerged one after another.

But several years later, industry realities have become more complex.

On one hand, Tesla's approach has proven that pure vision systems can achieve a certain level of advanced driver-assistance systems (ADAS). On the other hand, Chinese automakers have rapidly turned LiDAR into just another line item on a specification sheet. Products that once cost tens of thousands of yuan per unit are now priced in the low-thousand-yuan range.

A head of intelligent driving at a major OEM once stated plainly: "When purchasing LiDAR today, the first metric we look at is no longer the maximum detection distance, but the BOM cost."

This places immense pressure on all suppliers.

Hesai does possess moats. It boasts strong mass production capabilities, in-house chip development skills, and resources from top-tier customers. Yet, even the most powerful supplier finds it difficult to escape the reality of aggressive bargaining by vehicle manufacturers.

Thus, Hesai began telling a larger story: spatial intelligence. At the launch event, Hesai Technology CTO Xiang Shaoqing proposed, "Spatial perception is merely the first step; what truly matters is enabling machines to understand space, operate within it, and serve it."

In other words, the company is moving from selling hardware to selling system capabilities. If past LiDAR was responsible for "seeing the world," future spatial intelligence products will address "what to do after understanding the world."

Kosmo is the first tangible product. According to Xiang Shaoqing, Kosmo can comprehensively capture and record the 3D world. With continuous iteration, it holds the potential to become a "pocket camera" for a broader user base. Simultaneously, it can support training for physical AI by providing spatial data to robotics manufacturers for world model training.

An industry analyst told Wallstreetcn: "Even if the automotive LiDAR market is large, it essentially serves dozens of carmaker clients. Once consumer electronics succeed, however, it opens up a terminal market of hundreds of millions. These are two entirely different scales."

Hesai clearly sees this distinction.

Why Now?

Hesai chose to pivot at this moment because the growth dividend of the automotive market is beginning to taper off.

While NEVs continue to grow, supply chain profits are being continuously compressed. To fight for market share, automakers pass down price war pressures layer by layer to component suppliers. A source in the industry pointed out to Wallstreetcn that although LiDAR penetration rates are rising, average selling prices continue to fall, making profitability based on unit price increasingly difficult.

As LiDAR technology matures, its application scenarios have also expanded.

An industry investor told Wallstreetcn: "LiDAR was expensive in the past due to complex mechanical structures, high chip costs, and low mass-production yields. Today, with chip integration, solid-state design, and scaled manufacturing, cost reductions have far exceeded expectations."

A person close to Hesai told Wallstreetcn: "The company's core judgment over the past two years has been: LiDAR is not the endpoint; spatial data is the endpoint. 'If we only make sensors, we will eventually enter price competition. But if we master the entry point for spatial understanding and interaction, the imagination space is completely different.'"

The Picasso SPAD-SoC released by Hesai this time essentially integrates capabilities that were previously scattered across multiple modules into a highly compact form. After chipification, the product is smaller in size, consumes less power, offers better cost control, and is more suitable for consumer-grade scenarios.

A supply chain source noted: "Many technologies initially serve high-end industries before eventually trickling down to the mass market. LiDAR is currently at this stage."

Moreover, AI provides a new entry point. Without large language models, spatial perception remains merely a hardware upgrade. However, with generative AI and embodied intelligence, machines need to understand the 3D world in real-time, redefining the value of LiDAR. Robots need obstacle avoidance, grasping, and navigation; AR glasses require spatial modeling; home assistants need to understand room structures; and unmanned devices need to recognize human-object relationships.

These are tasks that traditional cameras alone cannot easily accomplish. Hesai management stated at the event: "In the future, every intelligent agent will require reliable spatial input."

The implication is clear: Hesai wants to transform LiDAR from an automotive component into foundational hardware for the AI era.

Consumer Electronics Is No Easy Battle

The problem lies in the fact that consumer electronics is never a market that suppliers can easily win.

Although the automotive industry involves long cycles and strict certifications, once a supplier enters the chain, orders remain stable and product lifecycles are long. Consumer electronics is entirely different: rapid iterations, fierce pricing, and high brand concentration mean that even a slight mistake can lead to replacement.

More critically, consumers will not pay simply because "this is LiDAR"; they only pay for experiences.

Therefore, if Hesai wants to enter the C-end market, it cannot simply shrink automotive-grade radars and sell them. It must prove that it can deliver new experiences that users are visibly willing to pay for.

For example, home robots navigating precisely in the dark; AR devices achieving centimeter-level spatial positioning; smart security systems accurately identifying abnormal behaviors rather than false alarms; and companion robots understanding human positions and movements.

If it fails to achieve this, LiDAR may remain nothing more than an expensive add-on.

A consumer electronics investor told Wallstreetcn: "Many supply chain companies mistakenly believe that strong technology alone allows them to succeed in the consumer market. However, the consumer market competes on product definition, channel capabilities, and brand mindset. These are not the same thing."

This is the true challenge facing Kosmo.

Hesai's past strength lay in B-end delivery capabilities—stability, reliability, mass production, and automotive certification. The consumer market, however, demands aesthetics, interaction, ecosystems, content, and operations.

Moving from B-end to C-end is often not an addition but a complete reconstruction of organizational capabilities.

Therefore, compared to directly attacking the mass consumer electronics market, robotics may be a more realistic landing spot for Hesai. Industry insiders view robotics as combining B-end procurement logic with C-end growth potential.

Robots require reliable hardware and prioritize performance parameters while operating in an early stage where the supply chain landscape has not yet solidified. Whether it's warehouse robots, delivery robots, home robots, or companion robots, all require spatial perception capabilities.

The robot power modules showcased by Hesai this time also indicate that it does not want to be just "eyes" but also "limbs."

Because making only sensors keeps a company as a component vendor; entering execution systems brings it closer to becoming a robotics platform company. In a future robot, the synergy between perception systems, motion systems, and decision systems creates whole-machine value far exceeding that of a single sensor.

A robotics entrepreneur told Wallstreetcn: "Whoever can simultaneously provide standardized modules for perception + motion will have the opportunity to become the Tier 1 of the robotics era."

Hesai clearly aims to compete for this position and revalue itself. Moreover, competition in the robotics industry is far less brutal than in the automotive supply chain. Top players have not yet formed stable barriers, and technical routes are changing rapidly, leaving a window of opportunity for latecomers.

Next, whether Hesai can emerge depends on two questions: First, can Kosmo find a killer use case instead of remaining a mere technology showcase? Second, can the company upgrade from a parts-thinking mindset to a platform and product-thinking mindset?

The new script has only just begun. Regardless of success or failure, Hesai has taken the first step to demonstrate its stance: it does not belong solely to the automotive industry. If it succeeds, Hesai's boundaries will be completely opened.