The laser-navigation modules inside robot vacuums are becoming a high-volume component business of their own. Shenzhen Camsense Technologies’ Hong Kong listing documents show that shipments of its triangulation LiDAR sensors rose from approximately 4.76 million units in 2023 to 9.31 million in 2025, while the average selling price fell from RMB 68.6 to RMB 43.3.
That is close to a doubling in unit volume and a price decline of about 37% in two years. The figures do not prove that finished robot vacuums became 37% cheaper, but they help explain why laser mapping, lift-up turrets and additional line sensors are appearing across more price tiers.
Camsense is largely invisible to consumers, yet its listing materials describe it as an important spatial-sensing supplier to all five of the world’s largest robot-vacuum companies in 2025. Recent industry reporting connects its products with Roborock and Xiaomi supply chains, without identifying the exact models involved.
LiDAR is moving from a premium feature to industrial scale
LiDAR allows a robot vacuum to measure distance, build a map and plan a repeatable cleaning route. Early laser-guided robots made the raised turret a visible premium feature. Today the sensor itself is increasingly one component in a broader package that may also include line lasers, cameras, structured light, cliff sensors and floor-detection hardware.
Camsense’s history illustrates the shift. The company says it entered consumer-grade LiDAR in 2017, launched its first sensor in 2018 and reached 500,000 LiDAR shipments in 2020. By 2024, annual LiDAR shipments had reached approximately eight million units.
The listing data then records 9.31 million triangulation LiDAR units in 2025 alone. Scale, manufacturing improvements and competition reduced the average component price even as volume increased.
Falling sensor prices do not guarantee a cheaper robot
A lower LiDAR price can reduce the cost of one part of the navigation stack, but a robot vacuum’s retail price also reflects motors, batteries, brushes, mopping hardware, pumps, docks, cameras, processors, software development, support and distribution.
Manufacturers may use savings in one component to lower the final price. They may instead spend the difference on more sensors, a retractable turret, improved obstacle detection or a more capable dock. The Camsense figures therefore show cost pressure at the supplier level—not a direct formula for showroom prices.
They also help explain why comparing robots by the presence of “LiDAR” alone has become less useful. Two machines can use related sensing hardware and still behave very differently because of sensor placement, mapping algorithms, route planning and firmware.
dTOF and line-laser shipments are growing even faster
Traditional triangulation LiDAR is only part of the story. Camsense commercialised direct time-of-flight, or dTOF, sensors in the second half of 2024. Its reported dTOF shipments rose from approximately 52,300 units in 2024 to 1.156 million in 2025. In the first quarter of 2026 alone, shipments reached approximately 1.257 million.
The company’s line-laser business expanded rapidly as well. Listing data records about 752,000 units in 2024 and 5.658 million in 2025. Line lasers are commonly used for nearby geometry and obstacle sensing rather than whole-room mapping, although the exact role depends on the device’s complete sensor architecture.
This diversification matches current product design. A modern robot may use one system for mapping, another for detecting low objects, and separate sensors for edges, carpets and drop protection. More sensors do not automatically produce better navigation, but they give the software more information to work with.
What the Roborock and Xiaomi connection does—and does not—mean
Vacuum Wars’ November 9 report, drawing on recent industry research and the listing information, says Camsense products have entered Roborock and Xiaomi supply chains. Camsense’s own public profile confirms that its spatial-sensing products are deployed at scale in the robot-vacuum sector, but does not provide a model-by-model customer list.
It would therefore be inaccurate to assume that every Roborock or Xiaomi robot uses a Camsense module, or that two models with sensors from the same supplier will navigate identically. Manufacturers can source multiple sensor generations, customise mechanical integration and apply different software to the data.
For buyers, the finished robot remains the correct unit of comparison. Mapping speed, route efficiency, obstacle avoidance, low-furniture access, privacy controls and long-term firmware support matter more than the name printed on an internal sensor.
Retractable LiDAR is one visible result of sensor miniaturisation
Camsense says it achieved mass production of its P-series lift-up 360-degree panoramic LiDAR in 2024. Retractable turrets are now a prominent feature in premium robots because they combine the broad view of a raised laser scanner with a lower body when the robot needs to pass beneath furniture.
The mechanism introduces its own engineering trade-offs. A moving turret needs reliable sealing, accurate position control and software that knows when the sensor can retract without losing navigation confidence. Smaller, cheaper sensors make the idea more practical, but do not eliminate those requirements.
Why this matters for parts and repairs
As spatial sensing becomes more specialised, visual similarity is a poor guide to replacement-part compatibility. A LiDAR module may differ in connector, voltage, communication protocol, mounting points, firmware calibration or optical window even when its housing looks familiar.
Goods Scene carries Roborock robotic-vacuum parts for supported models. Sensor, motherboard and navigation components should always be matched by the complete robot model and part number; supply-chain relationships are not evidence that modules are interchangeable between Roborock, Xiaomi or any other brand.
The larger industry signal
Camsense completed its Hong Kong listing on September 30, 2026, after raising funds intended primarily for research, development and manufacturing capacity. The more important signal for the cleaning industry is the scale behind the transaction: robot-vacuum navigation is no longer a niche optical subsystem produced in modest quantities.
Rapidly rising shipments and falling unit prices suggest that capable spatial sensing will continue to spread. The competitive question is shifting from whether a robot has LiDAR to how effectively it combines mapping, near-field perception and software—and whether those systems remain serviceable after years of use.
Community
Discussion