China’s Solid-State LiDAR Makers Target Robot Market in New Phase of Competition
China’s leading LiDAR manufacturers are pivoting from autonomous driving to mobile robots, betting that low-cost solid-state sensors will fuel the next wave of intelligent automation.
Hesai Technology, Huawei’s HiSilicon unit and RoboSense are racing to integrate VCSEL and SPAD chips into compact, robust LiDAR systems aimed at robotaxis, logistics robots and consumer devices such as lawnmowers and pool cleaners.
The shift comes as demand broadens beyond advanced L2+ driver-assistance systems to robotaxis and industrial mobile robots, where simpler environments and limited compute power favor highly integrated, cost-efficient LiDAR designs.
In 2025, Chinese suppliers are clustering around this segment, seeking scale through solid-state architectures that can be produced using conventional semiconductor processes and priced aggressively for mass deployment.
For investors, the transition marks a potential new revenue stream for LiDAR makers after years of heavy investment in automotive programs with long qualification cycles. Success in mobile robotics could diversify end markets, reduce dependence on passenger-car platforms and accelerate time to volume production.
Solid-State LiDAR: VCSEL and SPAD at the Core
The new wave of solid-state LiDARs is built around two key components: VCSEL emitters and SPAD detectors.
VCSEL (vertical-cavity surface-emitting laser) devices emit light perpendicular to the chip surface and can integrate hundreds or even thousands of lasers on a single die. This simplifies manufacturing and sharply lowers cost; chips carrying thousands of VCSELs can be produced for only a few US dollars.
SPAD (single-photon avalanche diode) detectors offer extremely high sensitivity, capable of detecting single photons and operating reliably in near-total darkness. Their single-photon detection capability enables true all-weather perception. SPADs can be fabricated with standard semiconductor processes and integrated at chip level with VCSEL arrays, supporting scalable, compact solid-state LiDAR systems.
These characteristics are critical for mobile robots in factories, warehouses and consumer settings, where price, power consumption and robustness against lighting conditions can be more important than the long-range performance demanded by highway-driving LiDAR.
Hesai Pushes Deep Chip-Level Integration
Hesai Technology is pursuing full-stack in-house development, extending from system design down to core LiDAR chips. The company recently unveiled its self-developed RISC-V main control chip, Fermi C500, and a “photon isolation” safety architecture.
Fermi C500 integrates MCU, FPGA and ADC in a single RISC-V-based SoC, with a built-in point-cloud intelligence engine (IPE) and 256-core waveform processing array. The chip is designed to filter environmental noise and enhance point-cloud quality while consolidating control and signal-processing functions.
Hesai says it has completed four generations of chip platforms, now self-developing its lasers, detectors, laser drivers, transimpedance amplifiers, ADCs, digital signal processors and controllers. That level of vertical integration is intended to reduce cost and secure critical supply.
Its 800-line SPAD digital LiDAR ETX uses a self-developed SPAD-SoC from the company’s long-range automotive LiDAR line. Mounted behind the windshield, ETX can achieve a claimed 400-meter detection range at 10% reflectivity, with angular resolution up to 0.05° x 0.025°. The second-generation, fully solid-state, automotive-grade FTX model uses pure electronic scanning, with a horizontal field of view up to 180° and vertical FOV of 140°.
Tackling “Ghosting” and Cross-Talk Risks
As more LiDAR units are deployed in close proximity, interference between sensors has become a safety concern. Traditional SPAD-based architectures use a small number of lasers and dozens of reception channels that are not isolated, leading to cross-talk and artifacts.
Hesai highlights several resulting failure modes: “ghosting,” where the system detects objects that do not exist; “high-reflection expansion,” in which point clouds around highly reflective objects are distorted and inflated; and “mis-filtering,” where algorithms overcompensate and remove critical data.
The company’s “photon isolation” technology aims to mitigate these risks by mapping hundreds of lasers to hundreds of independent detection channels, ensuring that photons received by one channel do not interfere with others. Each laser can emit independently and supports dynamic exposure control, which is designed to suppress ghosting and cross-talk while maintaining fine-grained control over signal strength.
For mobile robots operating in cluttered indoor environments or multi-robot scenarios, such interference management could be essential to maintain safety and reliability as deployment densities rise.
Huawei HiSilicon Bets on System-Level Coordination
Huawei’s semiconductor arm HiSilicon has emerged as another major player in solid-state LiDAR, focusing on coordination between VCSEL, LDD (laser diode driver) and SPAD chips at the system level.
Its S1970 high-precision solid-state LiDAR solution is designed as a complete chain from VCSEL emission through LDD driving to SPAD reception. HiSilicon positions this as a system-level co-design approach to boost performance and simplify hardware integration.
The S1970 platform uses a back-illuminated SPAD detector with photon detection efficiency of up to 30% at 940 nm, enabling stable sensing in low light and under strong ambient illumination. A high-resolution time-to-digital converter with 250-picosecond time precision underpins more accurate distance measurements.
On the emission side, the VCSEL chip offers 50% power conversion efficiency and peak optical power of 100 W, allowing for energy-efficient yet powerful pulses. The accompanying LDD driver supports low jitter, high current and 12-channel time-division driving, maintaining tight pulse control even under high repetition rates.
HiSilicon says S1970 improves system integration and simplifies board-level design. Stacking automated guided vehicles (AGVs) using this solution are already deployed in Warehouse No. 7 of the Anbo Logistics Park, according to the company. The platform is also targeted at lawnmowing robots, surveying and mapping robots, municipal service robots and industrial drones.
RoboSense Recalibrates Strategy Around RISC-V
RoboSense has adjusted its technical roadmap and is now also developing solid-state LiDAR architectures that integrate application processors and microcontrollers and adopt RISC-V for both APU and MCU.
The company’s EM4 product reaches up to 2,160 lines of resolution. In the Geely Auto Haohan-S architecture and the Zeekr 9X Haohan H9 variant, RoboSense supplies a customized 520-line LiDAR, underscoring its continuing exposure to automotive platforms even as it develops new solid-state designs.
RoboSense says it has addressed several key challenges associated with SPAD-based LiDAR, including high-reflection expansion, ambient light interference, contamination blocking the optics, and noise from rain and snow. The company claims its solutions are ready for large-scale, on-vehicle mass production.
For investors, RoboSense’s pivot suggests it is seeking to stay competitive against rivals that have moved aggressively into chip-level integration, while leveraging its existing automotive relationships and product portfolio.
Mobile Robots as the Next Volume Driver
Mobile robots, including logistics carriers, service robots and specialized devices such as lawnmowers and pool-cleaning robots, operate in relatively constrained environments compared with highway-driving cars. Their controllers often rely on microcontroller-class compute rather than high-end automotive SoCs.
In such settings, low-cost, highly integrated solid-state LiDAR can play a pivotal role, compensating for limited onboard compute and enhancing navigation, localization and obstacle avoidance. Yet these robots still face complex dynamic environments, with challenges such as localization drift, navigation errors and lighting interference that can affect efficiency and safety.
The clustering of Chinese LiDAR makers into this segment in 2025 reflects a broader search for scalable, near-term revenue. While the global robotaxi market remains a major long-term target, the faster deployment cycles and diversified use cases of mobile robots offer an attractive bridge to sustained commercialization of solid-state LiDAR technology.