RoboSense (速腾聚创) posts sharp Q4 turnaround as robot lidar surges; questions remain over sustainable margins

Big swing in results — profit in Q4 after three losing quarters

RoboSense (速腾聚创) turned a year of losses into a dramatic fourth‑quarter reversal, reporting a full‑year net loss of RMB 145 million (1.45亿元) for 2025 but an adjusted loss of just RMB 54 million (0.54亿元) after stripping out non‑cash equity incentives. The swing came from a Q4 single‑quarter profit of RMB 104 million (1.04亿元), which erased the combined RMB 249 million deficit booked in the first three quarters (RMB 99m, 50m and 100m respectively). Fast growth rescued the headline number. But can growth withstand relentless price pressure in the automotive market?

Shipments and margins: volume up, ASPs down

RoboSense reported vehicle‑grade lidar shipments of 238,400 units in Q4, up roughly 37% quarter‑on‑quarter, driving revenue growth of 23.6% in the period — even as average selling prices in the car segment plunged from near RMB 3,000 in 2024 to about RMB 1,800 today. Robot lidar was the real winner: 2025 robot lidar shipments reached 221,200 units (22.12万台), with unit economics markedly stronger. Robot products delivered a Q4 gross margin around 37.3% and generated RMB 347 million in revenue, representing about 49% of the company’s sales, while vehicle lidar posted a slimmer 22% margin. Rapid unit growth masked weaker year‑on‑year vehicle revenue, exposing the tradeoff between scale and falling ASPs.

Technology push and strategic positioning

RoboSense has migrated to a VCSEL+SPAD‑SoC architecture and 3D stacking that integrates SPAD arrays with a processing SoC on one die, a move the company says reduces parts count and cost while boosting performance. SPAD (single‑photon avalanche diode) sensors operate at extreme sensitivity; coupled with in‑house SoC development, RoboSense claims step‑changes in line count, range (over 500m) and accuracy (sub‑centimetre to centimetre levels). The firm’s ADAS gross margin reportedly rose from 13.4% to 19.1% in 2025 as it shifted from external FPGAs to self‑developed silicon. Amid global semiconductor tensions and export‑control risks, many Chinese lidar makers are accelerating chip self‑reliance — can this combination of optics and domestic silicon give them lasting advantage?

Market implications and partnerships

The lidar market’s first major shakeout is underway, and two leaders — RoboSense and Hesai (禾赛) — have emerged as front‑runners in both vehicle and robot segments. Robotaxi commercialisation and scaled autonomous services are changing demand dynamics: performance, cost and reliability now dictate winners. It has been reported that Baidu (百度) will adopt RoboSense for an upcoming RT7 robotaxi platform, and it has been reported that BYD (比亚迪) is weighing customized RoboSense solutions for new product refreshes. If those deals materialise, RoboSense’s volume and roadmap would get a big boost — but the company still faces the twin pressures of rapidly falling ASPs in cars and the need to sustain high‑margin robot growth. Who can balance performance, cost and scale fastest? That question will decide the next phase of China’s lidar wars.