Hesai Group Q3 FY2026 earnings call

Cosmo Platform Launch: Cosmo debuted globally in October 2026, securing partnerships with Pixar’s RenderMan and AMAP. It bridges the sim-to-real gap for physical AI by converting real-world spaces into high-fidelity 3D assets. Actuation Modules Ramp-up: Shipments exceeded 23,000 units in Q3 (up 97% sequentially), with full-body modules entering mass production. Cumulative shipments surpassed 35,000 units. LiDAR Market Leadership: Maintained #1 global share in Robotaxi LiDAR (62%) and #1 position in China’s long-range ADAS LiDAR market (43%) for 20 consecutive months. New Design Wins: Secured commitments from five major automakers for the ET Series (over 1 million unit order book) and blind spot LiDAR wins with Leap Motor and a top European OEM. Strategic Partnerships: Partnered with JD.com to support its deployment of over 4 million robots and collaborated with NVIDIA Hefei for autonomous driving integration.

Core LiDAR: Revenue reached RMB 978 million (USD 146 million), representing approximately 92% of total consolidated revenue. Operating income rose 88% year-over-year to RMB 145 million (USD 22 million). Strategic Growth Initiatives (SGI): Generated RMB 91 million (USD 13 million) in revenue, accounting for roughly 8% of total revenue. This segment recorded an operating loss of RMB 83 million due to disciplined R&D investments.

Q4 2026 Total Net Revenue: Expected between RMB 1.1 billion and RMB 1.15 billion. Full-Year 2026 SGI Revenue: Reaffirmed guidance of RMB 200 million to RMB 300 million. FY 2026 Gross Margin: Targeting close to 40%. FY 2026 Operating Expenses: Expected to increase ~20% year-over-year, driven by SGI R&D; core LiDAR opex is declining. FY 2027 SGI Revenue: Targeting approximately USD 100 million (RMB 700 million). FY 2027 SGI Profitability: Expecting the segment to reach break-even.

Automotive Market Softness: Broader ADAS demand was weaker than expected, with some customers lowering sales targets. Robotics Customer Issues: A robotic lawnmower customer faced company-specific problems, leading to lower-than-expected QT series shipments. Competitive Pricing Pressure: Industry competition remains intense, though management expects pricing for ATX LiDAR to stabilize as competition becomes more rational.

Q: Analysts asked about the sustainability of ASP and gross margins given soft automotive demand and increased SGI investment. / A: CFO Andrew Fan noted that while automotive demand softened, robotics LiDAR mix (higher ASP/margin) offset this. He confirmed that contracts for next year suggest ATX LiDAR pricing will stabilize, supporting stable margins in 2027. Full-year gross margin is targeted at ~40%, and core LiDAR operating expenses are declining despite overall group opex rising ~20% due to SGI R&D.

Q: Analysts sought details on the rapid ramp of robotic actuation modules and future volume potential. / A: CEO David Li disclosed Q3 shipments of >23,000 units (cumulative >35,000). Monthly capacity hit 10,000 units, with full-body modules in mass production. Annual shipments could reach six digits in 2027, driven by higher content per humanoid robot (>100 modules) and expanding beyond anchor customer Sharpa to other AI firms.

Q: Questions focused on Cosmo’s business model, TAM, and customer traction. / A: Li explained Cosmo generates revenue via hardware (Cosmo One), software subscriptions (CosmoWorld), and 3D asset licensing. Key partners include Pixar RenderMan and embodied AI firms like Galbit. He highlighted a potential market for 100,000 reconstructed spaces per humanoid company and emphasized that software/asset revenue will grow significantly in 2027 as hardware adoption scales.

Q: Analysts requested clarification on one-off items affecting net income comparisons and the contribution of non-ADAS businesses. / A: Fan clarified that YoY net income decline was due to RMB 185 million in one-off gains last year; excluding these, GAAP net income grew 14%. He stressed that robotics-related businesses (LiDAR, Cosmo, actuation) now contribute >50% of total revenue, signaling a successful pivot toward physical AI infrastructure despite ADAS headwinds.