Hesai Gr Q3 2026 Earnings Call Transcript

Hesai Gr (NASDAQ:HSAI) reported third-quarter financial results on Tuesday. The transcript from the company’s third-quarter earnings call has been provided below.

This transcript is brought to you by Benzinga APIs. For real-time access to our entire catalog, please visit https://www.benzinga.com/apis/ for a consultation.

View the webcast at https://edge.media-server.com/mmc/p/ec9bmwyr/

Summary

Hesai Gr reported Q3 2026 net revenues of RMB 1,068 million, marking a 34% year-over-year increase, driven by strong LiDAR and actuation module shipments.

The company’s Strategic Growth Initiatives (SGI), including Cosmo and robotic actuation modules, exceeded revenue targets with RMB 135 million in cumulative revenue, surpassing their full-year goal.

Hesai Gr’s core LiDAR business maintained a 43% market share in China’s long-range ADAS LiDAR market, with future growth anticipated due to increased demand for high-end and multi-LiDAR configurations.

Cosmo, the company’s 3D world data engine, debuted globally and secured notable partnerships, including with Pixar’s RenderMan, contributing to its initial revenue generation.

The company expects SGI revenues between RMB 200-300 million for the full year and aims for SGI to reach break-even in 2027, with significant potential in robotics and physical AI markets.

LiDAR ASP and gross margins are expected to stabilize, driven by rational competition and increased content per vehicle, with 2026 full-year gross margin targeted at close to 40%.

Robotic actuation module shipments nearly doubled quarter-on-quarter, reflecting strong demand and positioning for future growth in the humanoid robotics market.

Hesai Gr is expanding its global footprint through partnerships with Chinese OEMs and integration with NVIDIA’s autonomous driving platform, anticipating growth in both domestic and international markets.

Full Transcript

OPERATOR

Hello ladies and gentlemen, thank you for standing by. Welcome to Hesai Gr’s third quarter 2026 earnings conference call. At this time, all participants are in a listen-only mode. Please note that today’s conference call is being recorded. I’ll now turn the call over to our first speaker today, Yuan Ting Shi, the company’s Head of Capital Markets. Please go ahead.

Yuan Ting Shi (Head of Capital Markets)

Thank you, operator. Hello everyone and thank you for joining Hesai Gr’s third quarter 2026 earnings conference call. Our earnings release is now available on our IR website at investor.itech.com as well as via Newswire services. Today you will hear from our CEO Dr. David Lee who will provide an overview of our recent updates. Next, our CFO, Mr. Andrew Fan will address our financial results before we open the call for questions. Before we continue, I refer you to the safe harbor statement in our earnings release which applies to this call as we will make forward-looking statements. Please also note that the company will discuss non-GAAP measures today which are more thoroughly explained and reconciled to the most comparable measures reported under GAAP in our earnings release and SEC filings. With that, I’m pleased to turn the call over to our CEO, Dr. David Lee. David, please go ahead.

Yifan Li (Chief Executive Officer)

Thank you, Yanting. Hello everyone and thanks for joining us. Last quarter we laid out our vision for Hesai Gr’s next chapter: building a full stack infrastructure platform for robotics and physical AI, enabling intelligent machines to see, understand and act. LiDAR gives these intelligent machines the eyes to perceive the world. Cosmo, the 3D world data engine, turns the real world into high fidelity 3D environments they can learn from, and robotic actuation modules, the robotic motion engine, give them the muscles to turn intelligence into action.

This quarter we turned that vision into real commercial traction, and we’re moving faster than expected. Cosmo made its global debut in October 2026, secured deployments with world-class players such as Pixar’s RenderMan, and has already started generating revenues. On the actuation side, shipments ramped up fast, with full body modules now entering mass production alongside robotic hands. The momentum is showing up in the numbers too. By the end of Q3 2026, cumulative SGI revenues reached RMB 135 million, already surpassing the RMB 100 million full year target we set at the start of the year.

As these new growth engines scale up, our core LiDAR business continues to lead in global market share, as recognized by Yo Group. Based on our current expectations, we believe pricing and margin profile for our cost efficient ATX LiDAR could remain relatively stable as the industry moves toward more rational competition. Together, these businesses are bringing our vision to life. LiDAR remains Hesai Gr’s rock solid foundation, a scaled profitable business generating strong cash flow. In ADAS, we continue to lead. According to Gasgoo, we held 43% of China’s long range ADAS LiDAR market in September, extending our number one position to 20 consecutive months. Looking ahead, several powerful trends are driving our next phase of growth: Level 3-driven upgrades to better and more LiDARs, and broader adoption from premium to mass market models and from China to overseas markets.

First, Level 3 demands greater robustness in autonomous driving. As automakers gear up for Level 3, demand is shifting toward high-end LiDAR with longer range and higher resolution. Our flagship ET Series LiDAR commands a premium price compared with our cost efficient ATX, reflecting its superior performance. It has already been selected by five major automakers: a top European automaker, Cadillac, Geely, Great Wall Motor, and a leading NEV automaker, representing an order book of over 1 million units.

Second, more LiDARs per vehicle. We recently secured new FT Series blind spot LiDAR design wins with Leapmotor and a top European automaker. These wins provide exciting validation of the multi-LiDAR opportunity in China and beyond, reinforcing our confidence in higher LiDAR content per vehicle and a larger addressable market. Third, LiDAR is reaching a broader market. Wuling, one of China’s most popular mass market automotive brands, has joined our customer roster for AT Series main LiDAR, opening up significant opportunities in high volume mass market models.

Meanwhile, we have secured design wins with seven leading OEMs, including Chery, for potential export vehicle programs across multiple overseas markets, positioning us to capture growing demand as Chinese automakers expand globally. As these trends expand our ADAS opportunity, we’re also pushing the boundaries of perception with Picaso, our full color ultra sensitive fixed dSPAD SoC. By capturing depth and color on a single chip under one timestamp, Picaso gives machines a unified view of the world. Our Picaso-equipped ET Series LiDAR is undergoing customer validation, and discussions are progressing toward potential programs with leading robotaxi operators and global automakers.

Beyond automotive, we also partnered with JD.com, China’s largest retailer by revenues, leveraging our 6D full color LiDAR and advanced perception capabilities in its ambition to build the world’s largest robotics fleet, with plans to deploy over 4 million robots, robovans, and other intelligent machines within the next five years. Taken together, these developments showcase how Picaso is creating whole new opportunities for the industry. Meanwhile, Hesai Gr ranked number one globally in the robotaxi LiDAR market with a 62% share, according to Yo Group.

Seeing the world is just the beginning. The next frontier is helping AI understand the world, and that’s where Cosmo, our 3D world data engine, comes in. Launched globally on October 14, 2026, Cosmo is an integrated hardware-software platform for turning real world spaces into high fidelity, AI-ready 3D assets. Think of Cosmo as both an eye and a brain. Cosmo1, the world’s first 8K AI spatial camera, captures real world spaces in photorealistic 3D with millimeter-level accuracy.

Cosmoworld then transforms these captures into fully editable 3D worlds, using our AIGC capabilities to identify individual objects, fill in missing areas, and generate 3D assets up to five times the speed of comparable 3DGS solutions and up to 200 times the speed of manual modeling. Why does this matter? Because real world data is one of the biggest bottlenecks in physical AI training. Robots in the physical world are slow and expensive, while simulation often struggles to replicate its complexity, particularly depth and spatial relationships.

This sim-to-real gap can lead to failures when robots perform tasks in real environments. Cosmo bridges this gap by turning these environments into high fidelity, simulation-ready 3D worlds with interactive objects. A single reconstruction can generate over 10,000 training trajectories, allowing robots to practice object interaction and manipulation at scale. We’ve established strategic partnerships with leading embodied AI companies such as Galbit and Dexmo to build Cosmo-powered 3D robotic training environments.

Looking ahead, we believe demand for these photorealistic 3D environments could reach 100,000 spaces per humanoid company. But here’s what makes Cosmo even more exciting: the same technology can serve markets far beyond robotics. In film and entertainment, Cosmo turns real locations into reusable 3D assets, allowing creators to modify foregrounds, backgrounds, camera angles, lighting, and objects long after filming. For example, a 10,000 square meter European streetscape that could take two months to build as a physical set could be captured and reconstructed in just one week, with cost savings of up to 80%.

The market response has been encouraging. Within a few months of delivering prototypes, we have attracted strong interest from leading studios in China and overseas. Most notably, we’ve established a landmark partnership with Pixar’s RenderMan, a cornerstone of Hollywood visual effects, bringing Cosmo’s photorealistic 3D environments directly into its production-grade rendering pipeline. In October, Pixar released an official tutorial for creators worldwide showcasing a workflow built around Cosmo-captured 3D assets for its upcoming version of RenderMan.

Cosmo was the only capture hardware named in the tutorial. We believe this collaboration positions Cosmo to help define a new standard for the film and entertainment industry. We’re also expanding into immersive experiences. Our partnership with Amap, China’s leading digital mapping and navigation platform under Alibaba Group, includes an initial program covering 1,000 pilot sites to bring next-generation spatial commerce to tourism and offline retail at scale.

Looking ahead, we believe Cosmo could transform how content is created across advertising, music, gaming, and beyond, bringing a new dimension to the creative industries. Cosmo generated its first revenue in the third quarter of 2026. Its business models span hardware sales, platform subscriptions, and its 3D asset library. Through integrations with NVIDIA Isaac Sim, Unreal Engine, Pixar’s RenderMan, and Houdini, Cosmo is building an open platform to bring real world 3D data to both AI and creative industries. Our vision is simple: make 3D creation as easy as taking a photo and turn the physical world into reusable digital assets.

For those interested in learning more about Cosmo, we also published our Spatial Intelligence Whitepaper, a practical guide to how Cosmo works and how it can be applied. You can download it from our official website. Finally, AI needs muscles to turn decisions into physical action. That’s where Hesai Gr’s robotic actuation modules, our robotic motion engine, come in. The opportunity is enormous. A single humanoid robot may require over 100 actuation modules, potentially accounting for more than half of its BOM cost. As humanoid robots scale toward mass deployment, we see a potential trillion-dollar market taking shape. The strongest validation comes from real world adoption.

We’ve achieved sustained commercial traction with Sharpa, our anchor customer and a global frontrunner in AI robotics. Sharpa’s newly unveiled D01 is a full body humanoid robot with integrated tactile sensing, and its W02 is a next-generation fully tactile, ultra-compact dexterous hand. Both are powered by Hesai Gr’s robotic actuation modules from hand to full body. This momentum is reflected in our shipments. In Q3, robotic actuation module shipments reached over 23,000 units, up 97% sequentially, bringing cumulative shipments to over 35,000 units.

What differentiates us is the depth of our engineering. We design key components in-house, including reducers and motors, giving us greater control over joint-level actuation, system integration, and performance optimization. As robots become more capable and their requirements grow more demanding, we believe this vertically integrated approach can be a meaningful competitive advantage. We also believe the industry is moving toward AI-defined hardware, where advances in AI models increasingly shape robot architecture and hardware requirements.

As humanoid joint architectures and actuation standards continue to evolve, companies that can translate emerging AI capabilities into scalable, production-ready hardware will be well positioned to lead. That’s what makes our collaboration with Sharpa particularly valuable. As many of you know, I co-founded both companies, where their capabilities are highly complementary. Sharpa develops AI models and deploys robots in real world environments, generating insights that help shape the next generation of robotic capabilities.

Hesai Gr provides the infrastructure: LiDAR to see, Cosmo to understand, and actuation modules to move and interact with the physical world. More capable hardware enables more advanced robots, while real world deployment and advances in AI inform the next generation of hardware. This creates a powerful feedback loop and the opportunity to accelerate innovation across our ecosystem. We’re excited about the opportunities ahead as we power the rise of robotics and physical AI.

With that, I’ll turn it over to Andrew for our financials and outlook. Andrew, please go ahead. Thank you, David. Hello everyone. We delivered another quarter of record revenues and shipments. Our core LiDAR business continued to scale profitably with impressive momentum, while our Strategic Growth Initiatives, or SGI, gained further commercial traction. Together, these businesses are broadening our revenue base and opening new avenues for long-term growth.

Let me begin with the consolidated results. Total net revenues for the quarter reached RMB 1,068 million, or US dollars 159 million, up approximately 34% year over year, marking our 10th consecutive quarter of year-over-year revenue growth. This performance was driven by shipments of nearly 730,000 LiDAR units and over 23,000 actuation modules, along with Cosmo’s first revenue contribution. Importantly, businesses beyond automotive LiDAR contributed more than 50% of our total revenues this quarter, reflecting our progress toward becoming a leading infrastructure platform for robotics and physical AI.

Gross margin remained healthy at 41%. On the expense side, we continue to invest for growth while keeping a close watch on spending. Total operating expenses increased 46% year over year, driven by disciplined investment in SGI. Excluding SGI, operating expenses for our LiDAR business declined year over year. On the bottom line, we delivered our sixth consecutive quarter of GAAP profitability. Despite increased investment in SGI, GAAP net income reached RMB 81 million, or US dollars 12 million.

Non-GAAP net income came in at RMB 119 million, or US dollars 18 million. The year-over-year decline reflects one-time gains of RMB 185 million recorded in the same quarter last year. Excluding those items, our GAAP net income increased 14% year over year. Now let me turn to our segment performance. Starting with LiDAR, we continue to scale this business with a clear focus on profitable growth. LiDAR segment revenues reached RMB 978 million, or US dollars 146 million. Segment income from operations rose 88% year over year to RMB 145 million, or US dollars 22 million. This growth was driven by broad-based demand across both ADAS and robotics. For the quarter, total LiDAR shipments were up 65% year over year, even as the broader automotive market faced headwinds in ADAS.

As David discussed, multi-LiDAR configurations and high-end products are increasing our LiDAR content per vehicle. Meanwhile, based on our current expectations, pricing and margins for our cost-efficient ATX LiDAR under contracts signed for next year could remain relatively stable as the industry moves toward more rational competition. Turning to SGI, the third quarter marked an important step forward for our second growth engine. Cosmo generated its first revenue, adding a new revenue stream alongside the robotic actuation modules. Together the two businesses delivered segment revenues of RMB 91 million, or US dollars 13 million, up 101% sequentially. This brought year-to-date SGI revenues to RMB 135 million, already exceeding the RMB 100 million full-year target we set at the start of the year.

This early traction reinforces our confidence in SGI’s growth potential as we move from initial commercialization towards scale. We continue to invest in SGI with discipline, directing most of it toward R&D to build core technologies and long-term competitive advantages. These investments were primarily reflected in operating expenses, resulting in an SGI operating loss of RMB 83 million this quarter. Our R&D-led, capital-efficient approach allows us to advance our technology and commercial readiness without requiring a material increase in capex.

Our recent business progress underpins our SGI outlook. We are reaffirming our full-year 2026 SGI revenue guidance of RMB 200 to 300 million. For 2027, we continue to target approximately US dollars 100 million in SGI revenues and expect the segment to reach break-even that year. For the fourth quarter of 2026, we expect total net revenues of RMB 1,100 million to RMB 1,150 million, reflecting more moderate downstream demand. 2026 is a transformational year for Hesai Gr. Our core LiDAR business continues to deliver strong growth while our SGI businesses are gaining validation on the technology, customer, and financial fronts, opening the door to a market opportunity potentially 100 times larger than LiDAR alone. Together, these businesses give us confidence in our ability to finish the year strong and build on it through 2027. With that, we conclude our prepared remarks today.

Operator, we are now ready to take questions.

OPERATOR

Thank you. If you wish to ask a question, please press star one on your telephone and wait for your name to be announced. If you wish to cancel your request, please press star two. If you are on a speakerphone, please pick up the handset to ask your question. Your first question comes from Tina Howe with Goldman Sachs. Please go ahead.

Tina Howe (Analyst)

Hi, management. Thanks for taking my question. I think it’s encouraging to see that despite the weak demand, our ASP as well as our gross margin have been stabilized and even improving on a sequential basis. So my question would be maybe just for 2026 and 2027, what is our latest guidance given the overall market and competition dynamics? If you could share some color on volume, ASP as well as margin and profitability. Thank you.

Andrew Fan (Chief Financial Officer)

Thank you for the question. This is Andrew. We expect fourth quarter revenue of RMB 1.1 billion to RMB 1.15 billion. For the full year, SGI is expected to contribute approximately RMB 200 to 300 million, with robotic actuation modules accounting for the majority of SGI revenue. We are seeing softer-than-expected demand in the broader automotive market, while robotics LiDAR revenue remains broadly on track, supported by continued demand across multiple robotics applications and favorable product mix.

Higher shipments of Robotaxi, Robovan, and humanoid LiDAR products, which generally carry higher ASPs and margins, are helping offset some softness in robotic lawn mower shipments, which typically carry lower ASPs than other robotics LiDAR products. On profitability, while quarterly gross margin may fluctuate with product mix, we remain confident in achieving a healthy full-year gross margin of close to 40%. We are also stepping up investment in SGI as customer validation and the commercial demand are developing faster than originally anticipated. As a result, we expect 2026 full-year group-level operating expenses to increase by approximately 20% year over year, with most of the incremental spending going into SGI-related R&D. Excluding SGI investment, operating expenses in our core LiDAR business, however, are actually declining year over year. This shows that our core LiDAR business is becoming more efficient while continuing to fund investment in our next growth engine.

Looking ahead to 2027, we expect our core LiDAR business to continue scaling profitably, with our cost-efficient ATX LiDAR pricing and the margins stabilizing further as the competition becomes more rational. Meanwhile, SGI is entering a much larger growth phase in an emerging market. We maintain our expectation for SGI revenue to reach approximately US dollars 100 million, or RMB 700 million, in 2027, representing more than 10% of total revenue.

We also see a potential path for SGI to reach break-even next year. As we move into 2027, we will provide more clarity on our expectations. Stepping back, Hesai Gr is evolving into a full-stack infrastructure platform for robotics and physical AI. While we may face some near-term headwinds in the automotive market, we remain very confident in the long-term adoption of LiDAR across both automotive and robotics, as well as the opportunity to build underlying hardware and the data infrastructure for physical AI. It’s an exciting journey, and we’re just getting warmed up. Thank you for the question.

OPERATOR

Thank you. Your next question comes from Jeff Chung with Citi. Please go ahead.

Jeff Chung (Analyst)

Hi, this is Jeff. Thank you, David. Thank you, Andrew. My question is about the LiDAR ASP. So it looks like the blended ASP and the blended GP margin both improved quarter on quarter. So this is very different from last year third quarter; both declined quarter on quarter. So could you give us some outlook into the fourth quarter, whether the auto LiDARs ASP trend is going to be stable, as well as the GP margin? And separately, could you confirm for us the factual numbers on the third quarter LiDAR ASP and blended GP margin separately between the autos and robotics?

Thank you.

Andrew Fan (Chief Financial Officer)

Good question. This is Andrew. We are pleased to see pricing competition in ADAS LiDAR becoming more rational. Based on the contracts we’ve already signed for next year, we expect pricing for our cost-efficient ATX LiDAR to stabilize further, supporting a relatively stable margin profile in 2027 compared with 2026. If you look back over the past decade, we’ve helped bring LiDAR from a luxury feature to a mass-market safety component. That’s a major milestone for the industry.

We’ve also brought most of our core components and ASICs in-house, which has already driven significant cost improvements. Naturally, the room for another major round of technology-driven cost reductions is becoming more limited at this stage. Cutting prices at any cost could come at the expense of product quality and reliability. That’s simply not a trade-off we’re willing to make. Another thing I’d highlight is that blended ASP doesn’t tell the whole story for our business because it can be affected by product mix. For ADAS LiDAR, what really matters is the content per vehicle. As Level 3 adoption progresses, we’re seeing a shift from a single main LiDAR toward configurations with a premium main LiDAR plus multiple blind-spot LiDARs. This creates opportunities for higher total LiDAR content per vehicle, which we see as an important long-term structural trend.

On group-level gross margin, we’re still targeting close to 40% for full-year 2026. Based on what we see today, we believe we are well positioned to maintain a relatively stable gross margin level in 2027. There are a few factors supporting this outlook. Our ADAS business is expected to maintain a relatively stable margin profile, while robotics LiDAR generally carries higher margins than ADAS. Overseas, these programs also tend to have a more favorable margin profile.

On the SGI side, robotic actuation modules are ramping up, while Cosmo combines hardware sales with software subscriptions and 3D asset licensing. Particularly, we believe Cosmo has the potential to achieve gross margins above 40% as the business scales, given its business model. I hope that answers your question. Thanks.

OPERATOR

Thank you. Your next question comes from Shelley Wang with Morgan Stanley. Please go ahead.

Shelley Wang (Analyst)

Thank you for taking my question. I have a question on the actuators. It’s good to see our robotic actuator volume is ramping up very fast, double quarter on quarter. Can the management share more on the actuators?

David Lee

Hi, this is David. That’s a great question. We are very excited to see the ramp already underway. That’s why, starting this quarter, we will begin disclosing our robotic actuation module shipments. And the numbers speak for themselves. In the third quarter, we shipped over 23,000 actuation modules, nearly doubling sequentially, bringing cumulative shipments to more than 35,000 units. That’s already a meaningful volume in today’s humanoid industry.

Actuation modules also account for the vast majority of our SGI revenue. This year, monthly production capacity has reached 10,000 units and our full body modules are now in mass production alongside our hand modules. Based on our current pipeline, we believe annual shipments could reach six digits in 2027. There are several key drivers behind this growth. First, continued expansion with Sherpa. Sherpa, as you may know, is working with some of the world’s largest technology companies. It remains our anchor customer and the largest contributor to current shipments. Its products are moving into commercial deployment, including the Zero retrofit ice cream project at Dairy Queen. More recently, Sherpa launched the Do1, its first full body humanoid with tactile sensing, and the W02, its next generation dexterous hand, both powered by Hesai Gr actuation modules.

Second, a much larger opportunity per robot. This is one of the most exciting growth drivers for 2027. A full size humanoid can require more than 100 actuation modules, accounting for over half of its total BOM cost. As the industry moves from hand level applications toward full body deployment, the content opportunity per robot increases dramatically. With our full body modules already in mass production, we are ready to capture this growth. Third, a broader customer base. This will not be a one-customer business. Our modules combine exceptional performance with high reliability, including approximately three times the torque and power density of leading products. Sherpa’s impressive robot performance has also attracted attention from leading AI companies that have recently unveiled their humanoid strategies. We are already making progress in discussions on additional programs with them.

While we cannot disclose customer names at this stage, we are very encouraged by the traction beyond Sherpa. And here’s another important point. If our technology can meet the demanding requirements of a humanoid, it opens the door to many other applications. We are seeing opportunities well beyond humanoid robotics. We look forward to sharing more as these programs progress. So we’re not simply shipping more modules. We’re increasing content per robot, expanding our customer base, and unlocking new applications. With production ramping up and our pipeline broadening, we believe this business is entering an exciting phase of growth. Hope that gives you a better sense of where we are. Thank you.

OPERATOR

Thank you. Your next question comes from Aaron Wong with Jefferies. Please go ahead.

Aaron Wong (Analyst)

Hi. Thanks, management, for taking my question. This is Aaron from Jefferies. My question is on Cosmo. As the market increasingly recognizes data as critical for physical AI, could you share more on the term and business model for our Cosmo product? And also our recent customer order wins. Thank you.

David Lee

Hello everyone, this is David and I’ll take this question. Well, let me start with the business model. Cosmo is an integrated hardware and software platform with three revenue streams: hardware sales through Cosmo One, our 8K AI spatial camera; software subscriptions and usage-based fees through CosmoWorld, our AI cloud platform for 3D reconstruction and editing; and 3D asset licensing where captured environments can be reused and monetized across applications.

What’s exciting is that the value goes well beyond the initial hardware sales. More devices generate more data, driving software usage and expanding our 3D asset library. Over time, we expect software and asset licensing to become increasingly important sources of recurring revenue. In 2026, hardware will account for the majority of Cosmo revenue. Looking ahead, we expect software- and asset-related revenue to grow meaningfully in 2027. We’re still at the early stage of commercializing this new category, so it’s too early to provide a specific revenue split.

In terms of market opportunities, we’re seeing early adoption across three major areas. First, Embodied AI: through partnerships with Galvan and Dexmo, we’re bringing Cosmo into robot training. A single reconstructed environment can generate over 10,000 training trajectories, allowing robots to train around the clock in realistic virtual environments before entering the real world. Looking ahead, we believe demand for these photorealistic 3D environments could reach 100,000 spaces per humanoid company, pointing to a significant market opportunity.

Second, film and entertainment. Our collaboration with Pixar’s RenderMan is an important industry validation, with Cosmo being the only capture hardware featured in its recent official tutorial. We’ve also demonstrated the potential to reduce visual effects production cost by up to 80%. Third, immersive experiences spanning tourism, cultural heritage, gaming, and spatial commerce. Our initial program with AMap, Alibaba Group’s mapping platform, covers 1,000 pilot locations.

Over time, the opportunity in this area extends to digitizing everyday experiences from shopping and music to advertising. We are already seeing customer traction in China and overseas, and Cosmo started contributing revenue in the third quarter 2026. Our focus now is to turn this initial traction into larger commercial deployments and grow software monetization alongside hardware sales. And these are enormous markets, as outlined in our Spatial Intelligence white paper we published today on our website. The humanoid robotics market could reach approximately US dollars 7.5 trillion by 2050, while the global media and entertainment industry represents an opportunity of roughly 2.8 trillion.

I’d encourage investors to read the white paper for a deeper look at Cosmo and the opportunities ahead. Ultimately, our ambition goes well beyond 3D reconstruction. When space can be recorded at scale, understood by intelligent agents, and reused continuously, Spatial Intelligence can mature from a frontier technology into a foundational AI capability. That’s the data infrastructure we’re building at Hesai Gr. Thank you for the question.

OPERATOR

Thank you. Your next question comes from Nora Min with UBS. Please go ahead.

Nora Min (Analyst)

Hi, this is Nora from UBS. I have a question for Andrew. Can you walk us through all the one-off items in third quarter results for this year and for last year so that we can have a normalized year-over-year net profit comparison? Thank you.

Andrew Fan (Chief Financial Officer)

Thanks for the question. This is Andrew and I’ll answer the question. Well, the shortfall was mainly on the ADAS side. The broader automotive market turned out to be weaker than expected at the beginning of the year. During the year, several automakers, including some of our customers, lowered their 2026 sales target and, naturally, when their production volumes come down, our ADAS LiDAR shipments are affected as well. That said, I think it’s important to distinguish between market conditions and our competitive position.

We came in a little below previous expectations on ADAS, but third quarter’s total revenue still grew 34% year over year while total LiDAR shipments increased 65%. According to Gasgoo, we held a 43% share of China’s long-range ADAS LiDAR market in September, ranking number one for the 20th consecutive months. Looking ahead, we continue to see strong opportunities as LiDAR penetration increases and the transition toward Level 3 drives demand for higher-performance products and multiple LiDARs per vehicle. Robotics was a bit different. One of our robotic lawn mower customers ran into some company-specific issues which led to slightly lower-than-expected JT series shipments, but we saw stronger-than-expected shipments for robotaxi and humanoid applications which helped offset the impact on robotics this quarter.

These applications also carry higher ASP and margin than lawnmower LiDARs, so overall robotics LiDAR revenue was broadly in line with our expectations. Now turning to net income, the year-over-year decline was mainly due to one-off gains in the third quarter last year. These included RMB 185 million from the disposal of an equity investment and the transfer of IP by the group. If we exclude these one-off items, GAAP net income actually grew 14% year over year, and we achieved that while stepping up our investments in SGI, particularly in R&D. Meanwhile, our core LiDAR business continues to improve, with LiDAR segment operating income up 88% year over year.

And here’s another important point I want to highlight in the third quarter this year. Our robotics-related businesses—let’s say robotics LiDAR, Cosmo, our 3D data engine, and robotic actuation modules—already accounted for over 50% of total revenue. So while we may face some near-term pressure from the automotive market, we’re also building a much broader business around physical AI. Our LiDAR business remains a scaled and profitable cash engine, while SGI is expected to be our next growth engine.

We’re investing today to build a much bigger opportunity for the future and we are confident in the long-term potential of both businesses. I hope that helps clarify things. Thank you.

OPERATOR

Thank you. Your next question comes from Dunn Lindren with CICC. Please go ahead.

Dunn Lindren (Analyst)

Hi David, Andrew and Zenting. Thanks for taking my question and congratulations on the strong momentum in your SGI. My question is about the robotics LiDAR. Could you please share some updates on the shipment of LOP and customer order pipeline for robotics? Thank you.

OPERATOR

Thank you. Your next question comes from Elizabeth Pang with DBS. Please go ahead.

Elizabeth Pang (Analyst)

Hi David. Hi Andrew and Yan Ting. Thank you for the opportunity. First of all, I’d like to congratulate the team on the strong set of results despite the challenging environment. I have two questions. One question on the export program. Could management share more information on export programs with the seven OEMs including Chery? When could we see these export design wins start production and what volumes could we expect going forward? And my second question is on the overseas LiDAR market.

How does management view the long term adoption potential of LiDARs in overseas ADAS markets by Western OEMs? We did see that over a couple of weeks ago Mercedes-Benz and BMW have announced that they will introduce Level 2 systems in Germany next year. However, the Mercedes Level 2 system does not incorporate LiDAR despite the existing partnership, Hesai Gr. Based on management’s discussions, do you and when do you expect Western automakers to eventually follow Chinese OEMs in adopting LiDAR-based solutions?

That’s my second question. Thank you.

David Lee

Good question. This is David. Well, we see three major growth drivers overseas. First, direct partnerships with global automakers. Our Mercedes-Benz programs remain on track as previously announced. Hesai Gr is a strategic LiDAR partner and confirmed supplier for its Level 3 programs across their vehicle platforms in Europe and China. Development and validation are progressing according to the customer’s timeline. Beyond this, we’ve secured major design wins with 10 global OEMs covering a broad range of models through their Chinese joint ventures.

These programs are expected to ramp up in bigger volume next year. Second, our cooperation with Nvidia is the reference LiDAR option integrated into Nvidia’s Hyperion autonomous driving platform. This gives OEMs adopting Nvidia’s solution a much easier path to integrate our LiDAR without having to start from scratch. It’s a powerful channel for expanding our reach globally. Third, Chinese automakers going global. We’ve secured export program design wins with seven leading Chinese OEMs, including Chery. As these automakers expand into international markets, we can grow alongside them. This is a true zero to one opportunity for us.

Looking at the bigger picture, we believe overseas ADAS is a very attractive opportunity. The addressable vehicle market is much larger and LiDAR ASPs and margins are generally higher. More importantly, some global automakers are moving directly toward Level 3 architectures with a high-performance front LiDAR and multiple blind-spot LiDARs per vehicle, as we’ve already seen in some of the new design wins we’ve secured. Ultimately, global automakers need more than a good sensor.

They need a partner that can deliver leading technology, automotive-grade quality, and global manufacturing and delivery capabilities. So we are very well positioned to be the first choice for that. We look forward to sharing more progress as these programs ramp up. Thank you for the question.

OPERATOR

Thank you. Your next question comes from Yawen Tan with BNP Paribas. Please go ahead. Pardon me. Yawen, your line may be muted. Thank you. There are no further questions. I’ll now pass back to management for closing remarks.

Yuan Ting Shi (Head of Capital Markets)

Thank you once again for joining us today. If you have any further questions, please feel free to contact our IR team. And this concludes today’s call. And we look forward to speaking to you again next quarter. Thank you and goodbye.

OPERATOR

Thank you. That does conclude our conference for today. Thank you for participating. You may now disconnect.

Disclaimer: This transcript is provided for informational purposes only. While we strive for accuracy, there may be errors or omissions in this automated transcription. For official company statements and financial information, please refer to the company’s SEC filings and official press releases. Corporate participants’ and analysts’ statements reflect their views as of the date of this call and are subject to change without notice.