HSAI

Hello, ladies and gentlemen. Thank you for standing by. Welcome to Hesai Group's First Quarter 2026 Earnings Conference Call. At this time, all participants are in listen-only mode. Please note that today's conference call is being recorded. I will now turn the call over to our first speaker today, Yuanting Shi, the company's Head of Capital Markets. Please go ahead.

Thank you, operator. Hello, everyone. Thank you for joining Hesai Group's first quarter 2026 earnings conference call. Our earnings release is now available on our IR website at investor.hesaitech.com, as well as via Newswire services. Today you will hear from our CEO, Dr. David Li, 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 press 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 over the call to our CEO, Dr. David Li.

David, please go ahead.

Thank you, Yuanting, and thank you, everyone. Today, we are incredibly honored and excited to announce that Hesai serves as strategic lidar partner and confirmed supplier for Mercedes-Benz models enabling Level 3 autonomy. We see this as a major milestone, not only a strong validation of our technology leadership, but also a clear signal that lidar is moving to core infrastructure in the global evolution of intelligent driving. More importantly, I am thrilled to announce that Hesai has officially entered a new era of growth and possibility. 2026 marks a transformative chapter for us as we initiate a strategic paradigm shift evolving from spatial perception to spatial intelligence. Beyond our lidar business, we are actively forging the eyes and muscles of physical AI.

This evolution underscores what has always been at the core of Hesai's DNA, a deep tech enterprise that leverages hardcore technological innovation and sustained R&D intensity to deliver the world's most cutting-edge products, ultimately unlocking profound long-term industry value. At our recent 2026 Tech Open Day, we unveiled several breakthrough innovations that we believe will redefine the industry landscape. I will come back to these in more details later in my remarks. Before turning to the broader opportunities in physical AI, let me first walk through our lidar business highlights for the quarter. Last year, we delivered a record 1.6 million lidar units. This year, we expect our total shipments to approximately double, reaching between 3 million-3.5 million units. We are confident that this level of scale firmly reinforces our leadership position in the global lidar industry.

The trend we highlighted before is only accelerating. LiDAR is fast becoming the invisible airbag deployed across vehicles at scale. This was clearly demonstrated at the recent Beijing Auto Show, where Hesai LiDAR was featured in 56 vehicle models across 24 leading automotive brands, ranking number one in LiDAR presence across exhibited models. At the show, our LiDARs were deployed across the automotive ecosystem. From leading brands such as Audi, Cadillac, Lotus, Li Auto, Xiaomi, BYD, Leapmotor, Geely, Great Wall Motor, Changan, Chery, to autonomous driving leaders like Pony.ai and WeRide. This broad adoption reflects our deep ecosystem penetration and the growing strength of the Hesai Inside effect. According to Yole Group, we ranked number one globally in long-range ADAS LiDAR shipments with a 43% market share in 2025.

Our leadership only strengthened in March 2026, when our China market share surged to 55%, according to Gasgoo, roughly triple that of the second-ranked player, marking our 14th consecutive month in the number one position. This momentum was fueled by our industry-leading product portfolio, with our flagship ATX continuing to scale across major platforms. In April 2026, we began SOP for the revamped version of ATX. OEM demand remained exceptionally strong, with backlog now exceeding 6 million units. As Level 2 scales, Level 3 is emerging as the next major growth driver for the lidar industry.

As responsibility shifts from the driver to the automaker, vehicle safety requirements are rising rapidly, driving architectures from typically one lidar per vehicle last year, representing roughly $200 of lidar content per car, to three to six lidars in Level 3 platforms. Increasing content to about $350 in entry-level multi-lidar setups and $500-$1,000 in more advanced configurations. We are well-positioned for this transition with multi-lidar design wins secured across leading OEMs such as Li Auto, Xiaomi, and Changan. Notably, Li Auto officially commenced deliveries of its multi-lidar model on May 15th, 2026, marking the first mass production deployment of our FTX blind spot lidar. Level 3 is not just about more sensors. It fundamentally elevates the role of lidar from a passive backup to an active decision enabler in safety-critical driving scenarios.

To lead this shift, we introduced Picasso, the world’s first 6D full-color ultra-sensitive SPAD-SoC. It fuses RGB color and precise 3D geometry at the true chip level, generating real-time colorized point clouds. In plain terms, it enables systems to see and understand the world more like humans do and directly addresses the long-standing lidar versus camera debate. For the first time, we are delivering a unified perception stack in a single chip, bringing both worlds together and taking a real leap in how machines perceive and understand the physical world. For example, through native pixel-level fusion of color and depth, Picasso-equipped lidars help address challenging real-world edge cases across intelligent systems, not only in autonomous driving, but also in robotics and industrial applications, such as distinguishing overlapping traffic lights at adjacent intersections or differentiating a yellow signal from sunlight, scenarios where vision-based systems can struggle.

Powered by the Picasso, our flagship ETX now supports up to 4,320 channels and delivers a maximum 600 m range, enabling ultra-high resolution, full-color, 3D imaging. With enhanced small object detection, ETX pushes safety margins to unprecedented heights. Commercial momentum is already building for this industry-first technology. We secured an exclusive design win with KargoBot for its Phase 2 transport robots, deploying our 4,320 channel ETX together with FTX blind spot lidars. This marks the first commercial vehicle contract for our 6D full-color lidar. ETX is expected to enter mass production in the second half of 2026 across multiple flagship programs, with additional engagements actively progressing among both ADAS and robotics clients.

Looking ahead, we see lidar expanding far beyond today’s boundaries, with many camera-dominated applications naturally shifting toward lidar over time, positioning us for the next wave of physical AI. On the global side, we are incredibly honored and excited to announce that Hesai serves as strategic lidar partner and confirmed supplier for Mercedes-Benz models enabling Level 3 autonomy. The new supply agreement supports Mercedes-Benz programs in Europe and China, with lidar production supported by Hesai’s new Galileo manufacturing center in Thailand. In parallel, we continue to deepen our footprint with Chinese joint ventures, as well as fast-growing Chinese automakers expanding overseas. We have been selected for GAC Toyota’s 2026 bZ3X model, marking our first entry into the Japanese automotive ecosystem. We have also secured new overseas design wins with Xiaomi, one of our largest customers, with SOP expected to begin in 2027.

Looking ahead, we see strong and sustained momentum driven by both global OEM partnerships and the continued international expansion of Chinese automakers. Together, these two forces are becoming the key structural growth drivers for Hesai in the years ahead. Beyond ADAS, robotics is emerging as an addressable market roughly 10 times larger. We are already a leader across key segments, including humanoid and quadruped robots, robotaxis, robovans, and robotic lawnmowers. Ranking number one, according to GGII, Yole Group, and Frost & Sullivan. To share a few recent highlights, in humanoid robotics, our JT128 lidar-powered Honor's Lighting robot to deliver a championship-winning performance and break the human world record at the world's first humanoid robot half-marathon. In robovans, we secured an exclusive design win with Zelos for 200,000 lidar units and deepened our partnership with Neolix as its largest lidar supplier.

We are also expanding into smart mobility with an FTX design win from NIU Technologies for next-generation electric two-wheelers, a market with more than 60 million units sold annually in China. Having established ourselves as a global leader in 3D perception solutions, one thing is clear: bringing AI into the physical world takes more than just lidar. We are taking our most exciting step yet, expanding from spatial perception to spatial intelligence. To build the uncompromised digital foundation for the physical AI era, we introduced Kosmo, an AI algorithm integrated spatial intelligence device. As physical AI accelerates, we believe richer, higher fidelity spatial data is becoming an increasingly scarce strategic resource, arguably even more critical than computing power itself. Kosmo is purpose-built to break this bottleneck.

Unlike traditional 3D reconstruction solutions that rely on bulky hardware and costly workflows, Kosmo combines Hesai's ultra-high precision lidar with proprietary 3DGS and AIGC algorithms to rapidly generate photorealistic production-grade 3D environments at scale. For a 200 sq m space, Kosmo requires only one-fifth the time of existing 3DGS solutions and roughly one-fiftieth the time of traditional methods with dramatically lower labor costs. More importantly, Kosmo is far beyond a hardware AI integrated product. It is the starting point of a scalable new business model. Combining hardware, AI software, spatial data, and future platform services, creating recurring revenue streams, strong ecosystem effects, and a durable long-term moat. We believe Kosmo unlocks RMB 1 trillion downstream opportunities across robotics, simulation and training, immersive media, 4D entertainment, and beyond. Commercial traction is already building rapidly.

We are actively engaging with a powerhouse roster of industry leaders, spanning global pioneering embodied AI companies and top-tier entertainment studios to market-leading tech giants, AI-driven industrial titans, and global luxury brands, with some customers already having placed orders. We are incredibly excited about the long-term growth potential ahead. Beyond perception, bringing AI into the physical world also means enabling it to act, giving it the muscles to truly come to life. This leads to our next strategic pillar, robotic actuation modules. This is a natural extension of our systemic expertise in materials, simulation, physical design, and the manufacturing of delicate components. We also bring proven know-how gained from millions of in-house developed automotive-grade motors and encoders deployed in our LiDAR products. Consider this a first glimpse, an Easter egg. More exciting updates will follow in the near future.

To wrap up, let me bring us back to the core theme today, our evolution from spatial perception to spatial intelligence. We are executing this across three clear dimensions. First, perceiving the world powered by our chip-based lidar solutions, giving robots the eyes to see the physical world in high fidelity. Second, understanding the world with Kosmo and its ecosystem, transforming rich spatial data into actionable intelligence for real reasoning and decision-making. Third, changing the world through our robotic actuation modules, the muscles that allow machines to truly interact with the physical world and serve humanity. Together, these are unlocking powerful new growth engines and positioning us as a core enabler of physical AI, digitizing the real world and redefining how humans and robots perceive and act. With that, I will now turn the call over to Andrew to discuss our financial performance and outlook. Andrew, please go ahead.

Thank you, David. Hello, everyone. Before diving into our first quarter financial and operating performance, I would like to provide an important update on our financial reporting structure. As David highlighted, 2026 marks a year of strategic evolution for Hesai as we expand our boundaries to become the key enabler of physical AI. To better reflect this shift and further enhance our disclosure transparency, starting this quarter, we are updating our segmentation into two distinct categories. Our core lidar business and our strategic growth initiatives, or SGI, which includes our new AI algorithm integrated spatial intelligence device, Kosmo. We believe this updated reporting framework will help investors better appreciate the distinct financial profiles, independent growth trajectories, and underlying value drivers of these two segments.

By separately reporting our robust, rapidly scaling core lidar business from our highly transformative SGI segment, we aim to provide you with clearer visibility to better model our business and track how each engine contributes to our long-term shareholder value. With that, let me walk you through our first quarter operating and financial performance, followed by our outlook for the second quarter and the rest of 2026. To be mindful of the length of our call, I encourage listeners to refer to our earnings release for further details. We kicked off 2026 with strong momentum and delivered another strong quarter of robust growth, improving scale, and disciplined profitability. Starting with the top line, total net revenues reached RMB 681 million, or $99 million, representing an increase of 30% year-over-year.

This marks our eighth consecutive quarter of year-over-year revenues growth, reflecting both sustained demand across our core markets and our ability to scale with consistency and operational rigor. This performance was supported by total LiDAR shipments of over 471,000 units for the first quarter, with both broad-based ADAS and robotics shipments more than doubling year-over-year. Our gross margin for the quarter stayed healthy at over 39%. On operating expenses, we maintained a disciplined approach while continuing to invest strategically in long-term growth. During the quarter, total operating expenses increased by 9% year-over-year, primarily reflecting targeted R&D investments in the new initiatives that David outlined earlier, partially offset by savings from scaled adoption of AI tools across the company. Excluding these investments, our LiDAR operating expenses declined year-over-year, driven by ongoing efficiency gains and cost discipline.

This brings us to the underlying profitability of our two segments. For the first quarter, our LiDAR business delivered a robust operating profit of RMB 42 million, demonstrating its strong self-sustaining profitability. Meanwhile, our SGI business recorded an operating loss of RMB 51 million, representing our deliberate and forward-looking investment in building the infrastructure of physical AI. As a result, we delivered GAAP net income of RMB 18 million, or $2.7 million, and non-GAAP net income of RMB 48 million, or $6.9 million, for the quarter. This extends our track record to four consecutive quarters of GAAP profitability and six consecutive quarters on a non-GAAP basis. The sustained profitability reflects the strength of our operating model and the benefits of scale as our business continues to expand, firmly positioning us at the forefront of the industry.

Looking ahead, we expect to build on this strong momentum. For the second quarter of 2026, we expect net revenues to be in the range of RMB 850 million-RMB 900 million, or $123 million-$130 million, representing year-over-year growth of approximately 20%-27%. We also expect revenue momentum to strengthen progressively each quarter throughout the year. From a broader perspective, we believe our business is entering a more sustainable growth trajectory. Our core LiDAR business will continue to drive growth, scale, and cash generation, while SGI is beginning to contribute and significantly expand our long-term growth potential.

Supported by our strong customer base and the momentum from early adoption, we expect SGI to generate approximately RMB 100 million in revenues this year, with revenue contributions expected to start in the second quarter of 2026. Looking further ahead, we aim to scale this to around RMB 500 million by 2027, and over time, establish it as a meaningful and long-term growth driver for Hesai. Together, these developments are accelerating our evolution into a key enabler of physical AI, unlocking the next phase of growth with greater scale and global impact. This concludes our prepared remarks today. Operator, we are now ready to take questions.

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're on a speakerphone, please pick up the handset to ask your question. For the benefit of all participants on today's call, if you wish to ask your question to the management in Chinese, please immediately repeat your question in English. For the sake of clarity and order, please ask one question at a time. Management will respond, and then feel free to follow up with your next question. Your first question comes from Tina Hou with Goldman Sachs.

Hi, management. Thank you very much for taking my question, and congrats on a strong set of results. Very excited about our new opportunities here. My question is more related to our announcement with Mercedes-Benz and supplying to their L3 models going forward. Wondering how should we think about the opportunity in terms of volume, revenue, and margin contribution over the next, let's say, one to three years, and then longer term from this collaboration? Thank you.

I'll take this question. This is David. We are pleased to announce that Hesai serves as strategic lidar partner and confirmed supplier for Mercedes-Benz models enabling L3 autonomy. The new supply agreement supports Mercedes-Benz programs in Europe and China, with lidar production supported by Hesai Group's new Galileo manufacturing center in Thailand. We will not be able to disclose further details on the strategies of the customer, we do confirm the new contracts are a continuation of the existing collaboration with additional volumes and car models. Our global expansion is fully supported by our new Thailand manufacturing center, unified product architecture, and strategic NVIDIA partnership, uniquely positioning Hesai to lead the global ADAS market. We are expanding our overseas reach with Chinese joint ventures and fast-growing domestic OEMs going global.

Winning a design win on GAC Toyota's 2026 bZ3X model marks our breakthrough into the Japanese auto ecosystem. We've also locked in new overseas design wins with Xiaomi, a top customer targeting a 2027 SOP. Looking ahead, our growth will be structurally driven by two powerful engines: international OEM partnerships and a global expansion of Chinese automakers. That's my response to your question. Thank you.

Thanks, David.

Your next question comes from Tim Hsiao with Morgan Stanley.

Hi, this is Tim from Morgan Stanley. Thanks for taking my questions. Switching gears to the spatial intelligence device, because we noticed Kosmo has already received early orders, and it's also highlighted in the announcement that as a scalable business model. Wondering if you could elaborate a bit more about how we should think about the fundamental difference or synergy between the Kosmo and the conventional lidar business. Are there any comparable products currently available in the market? How does Kosmo stand out against such alternatives, if any? Additionally, it'd be great if the management team can share more details regarding if the existing client base. I think David also shared a lot of information.

In the meantime, timeline for consumer-facing rollout as well as, your expectation for financial contribution from Kosmo. Yeah, that's a few follow-ups. Thank you.

Thank you for raising the question. Well, you know, Kosmo is much more than a sensor or a hardware device. I would describe it as the world's first AI algorithm integrated spatial intelligence device. It combines our custom LiDAR, multiple sensors, and our in-house 3DGS and AIGC algorithms together into one system, like allowing us to capture and reconstruct the physical world in a very seamless and photorealistic way. We believe in the era of physical AI, high-quality spatial data is becoming a scarce strategic resource. Arguably, I mean, even more important than compute itself. Kosmo is designed to provide what we call the uncompromised digital foundation for this new era, right? One key difference between traditional LiDAR and Kosmo is really the output. Traditional LiDAR gives you raw point clouds, which are almost like the wireframe or skeleton of the world.

Kosmo goes much further. Using AI algorithms, it reconstructs a fully photorealistic 3D model with color, texture, and shadow, much closer to how humans actually experience the real world, right? We think we have two very important advantages here. First is hardware. Since we use our own custom high-performance lidar, the raw input quality is already extremely strong compared to, you know, most third-party solutions. Second is software. Our proprietary reconstruction and AIGC algorithms are a major differentiator. Based on early customer feedback, you know, our models consistently deliver higher quality results compared with other software solutions. As a result, efficiency improvement is actually very significant. For example, for a 200 sq m space, Kosmo can complete reconstruction in roughly one-fifth of the time of existing 3DGS solutions and around one-fiftieth of traditional workflows, while also reducing labor costs substantially, which is huge, right?

The market opportunity here is still very early and evolving, so it's difficult to define a precise TAM today. Directionally, we think this could become a very large multi-industry opportunity spanning robotics, simulation and training, immersive media, 4D entertainment, industrial applications and beyond. What's even more important, in my view, is the business model we are building around Kosmo. The AI-integrated Kosmo is really just the front door. It gets us into the ecosystem, but recurring revenue is, you know, what we believe becomes the long-term engine over time. Today, we are already engaging with a very exciting group of customers, ranging from pioneering embodied AI companies and top-tier entertainment studios to, well, global technology leaders, industrial companies, and luxury brands. We also already have early orders in place.

Beyond the device itself, customers also pay for software and platform usage, whether it's data processing, model training, or cloud-based services through licensing and subscription models. Compared with traditional LiDAR, I mean, this business naturally carries a more software-driven and potentially higher margin profile. Longer term, like in a couple of years, we see two major paths ahead for Kosmo. First is 4D V, essentially 3D plus time, where we believe a single lightweight Kosmo device can eventually replace today's bulky, expensive camera-based 4D capture systems. Second is a broader consumer-facing opportunity. Over time, we believe even ordinary consumers will be able to use Kosmo to create immersive 4D content themselves, much like how smartphones democratized photography and video creation over the last decade, right?

Overall, we see Kosmo as the starting point of a scalable new business model and a very important expansion opportunity for Hesai beyond traditional lidar. We will soon begin demo shipments, and we believe it can become a meaningful growth driver for us in the physical AI era, with initial eight-digit revenues expected this year. That's my answer for your question. Thank you.

Thank you very much for sharing all the details, David. Thank you.

Your next question comes from Nora Min with UBS.

Hi, this is Nora Min from UBS. Thank you for taking my question. My question is about your progress of the humanoid robot actuator business, and what do you think is the current technical bottleneck, and who are your potential customers? Also, are the actuators for the body or hand? Thank you.

Thank you for the question. Well, people often think LiDAR is simply a sensor, but in reality, you know, it's much more like a precision machine. Inside a LiDAR system, I mean, you have optics, electronics, software, and sophisticated mechanical engineering all working together in tight synchronization, right? For both our mechanical and semi-solid state LiDARs, the moving components are extremely demanding from an engineering standpoint. They must meet automotive-grade standards and survive years of continuous vibration, heat, cold, and harsh real-world driving conditions while still maintaining extremely high precision every single second, right? I mean, it's a bit like asking a professional athlete to run a marathon every day for years without ever losing form or accuracy. Over the past decade, you see, we have already developed and shipped more than 1 million automotive-grade motors and encoders into our LiDAR systems.

Fundamentally, like, the physics behind LiDAR scanning modules and robotic actuation modules are very similar. Both require extremely precise control of velocity, position, and force. In many ways, moving from LiDAR into robotic actuation was a very natural extension for us, really. The same capabilities we built for LiDAR, such as material science, simulation, proprietary ASICs, precision manufacturing, and system-level engineering, can now be extended from helping robots see the world to helping them move within it. Now, on robotic actuation, we haven't yet disclosed detailed information, but early signals are very encouraging. Key performance metrics are already reaching what we believe is, well, among the industry's highest level. We look forward to sharing more details in the coming quarters as this business develops, of course. That's my answer. Thank you.

Thank you, David.

Your next question comes from Jessie Lo with Bank of America Securities.

Hi. Thank you for taking my question. My question is around the strategic growth initiatives as well. Could you talk about the ASP revenue and also margin on both the Kosmo and also robotic actuator modules?

Thank you for the question. This is Andrew. Both the robotic actuation modules and Kosmo are part of our new strategic growth initiatives or SGI segment. Think of this as an early-stage innovation layer sitting alongside our core lidar business. At this stage, because these products are still in early commercialization, we are reporting them on a consolidated basis within SGI rather than splitting them out individually. On the revenue side, we expect SGI to start contributing in the second quarter this year. With strong early customer interest and initial adoption, we currently anticipate around RMB 100 million in full year 2026 revenues. Looking further out, we see a clear scaling path, potentially reaching around RMB 500 million in 2027.

Over a longer horizon, within five years, we believe this business could grow to a scale that is comparable to our core lidar business. On profitability, SGI is expected to be accretive to group gross margin over time as this segment scale and matures. Kosmo, in particular, should carry a structurally higher margin profile because it is not just a hardware product. It integrates hardware, software, and data workflows. Over time, we also expect part of Kosmo to evolve toward a recurring ARR model through software and platform services, which naturally improves profitability as the installed base expands. On ASP, it's important to note that we are still in a very early deployment phase, so current pricing is highly customized and driven by pilot programs. At this stage, these early data points are not representative of long-term economics.

As we move into mass production and the product lines mature, we will provide more normalized unit economics to the market. That's my answer to your question. Thank you.

Sure. Thank you very much, Andrew.

Your next question comes from Danlin Ren with CICC.

Hello, everyone. This is Danlin Ren from CICC Auto Team. Thanks for taking my question, and congratulations on your strong results and exciting partnership with [Mercedes]. My question is regarding your Picasso chip. We think the market has came to understand how your product interacts with cameras. To what extent can 6D lidar replace certain camera functions? Additionally, some of your peers are also promoting camera lidar fusion solutions. How does your approach differ from theirs? What has the customer feedback been so far? Thanks.

Thank you for the question. First, let me set the stage here. People often debate cameras versus LiDAR, but I think that's a little bit like asking whether humans should choose between seeing colors or seeing depth. In the real world, we naturally use both together. That's exactly what our 6D full-color LiDAR does. It helps machines not only detect objects, but truly understand the physical world more like humans do. This is why we believe, in many future applications, your next camera may not even be a camera anymore. It could be a 6D full-color LiDAR. The opportunity is massive across autonomous driving, robotics, industrial automation, infrastructure, and even electronics like Kosmo. What makes us different is that many existing solutions are still basically a camera and a LiDAR packaged into one box with software trying to combine the information afterward.

That's still system-level fusion. With Picasso, our world's first 6D full-color SPAD SoC, we fuse RGB color and precise 3D geometry directly at the chip level from the very beginning, almost like the human brain naturally combining both eyes in real time. You can think of it as early fusion. The benefit is very clear. The system instantly understands both depth and color together in fully aligned real time, improving perception capability, reducing edge case errors, and enhancing safety. We understand some peers are moving in a similar direction now, which we see as validation of the trend, but we believe we are roughly about one year ahead in R&D. Customer feedback so far has been extremely encouraging across both ADAS and robotics. For example, with overlapping traffic lights at adjacent intersections, traditional cameras struggle with depth, while traditional LiDARs cannot understand colors.

Our 6D full-color lidar solves both simultaneously. We believe this creates a strong differentiator for us and supports premium pricing given the very clear value add. Importantly, customers also do not need to completely rebuild their perception stack. The data from 6D full-color lidar and traditional lidar are compatible. Commercially, the momentum is already starting. We recently secured an exclusive design win with KargoBot for its transport robot, which marks the industry's first mass production deployment of 6D full-color lidar on commercial vehicles. We are also actively engaging with both Robotaxi and ADAS customers. Looking ahead, we expect the Picasso-equipped ETX to enter SOP in the second half of 2026, with broader global adoption ramping through 2027 and 2028.

We really believe this could become one of the key technology transitions for the entire perception industry over the next several years. That's my response to your question. Thank you.

Yeah. Thanks, David. It's very clear. That's all from me. Thank you.

Your next question comes from Jeff Chung with Citi.

Hey. Hi, David. What's the updated [2026] revenue forecast right now, and how much SGI contribution to the full year revenue? First question. Secondly, if we take out the RMB 100 million R&D revenue from SGI in the second quarter, the core revenue excluding SGI Q-on-Q growth will become 18%. Would that be a little bit conservative? Finally, any volume guidance for the second quarter. Thank you.

Thank you. It's Andrew here. For the second quarter, we expect total revenues to be in the range of RMB 850 million-RMB 900 million, with around 650,000 LiDAR shipments. Importantly, the second quarter will also mark the initial revenue contribution from SGI. Gross margin has remained resilient even as our ADAS business continues to scale rapidly. We've now delivered a GAAP net income for four consecutive quarters and non-GAAP net income for six consecutive quarters. We expect to maintain this momentum. If you look at the pattern of our business and OEM production cycles, we also expect financial performance to continue improving sequentially through the year, with SGI expected to contribute around RMB 100 million in 2026 during this early ramp stage.

For the full year 2026, we are reaffirming our LiDAR shipment guidance of a record 3 million-3.5 million units. We remain highly confident in our top line and delivery trajectory, supported by several strong and accelerating drivers, including higher LiDAR penetration across the market, increasing LiDAR content per vehicle as the industry moves toward Level 3. The early ramp of global ADAS programs, including the overseas expansion of Chinese OEMs, as well as broader robotics adoption across increasingly diverse applications. Overall, we believe Hesai is executing strongly across all fronts. Our LiDAR business continues to serve as a very solid foundation with clear market leadership and global scale. At the same time, we are extremely excited about the future of SGI, and we look forward to sharing more updates on these new businesses very soon.

That's my answer to your question. Thank you.

Your next question comes from Zhang Yu with Huatai Securities.

Hi, management. This is coming from [audio distortion]. I have a question about ASP. How do we interpret the ongoing decline in LiDAR ASP? What's your long-term outlook for pricing as the market matures?

Thank you for the question. We think it's important to look beyond headline blended ASP because what we are really seeing today is a reflection of strong market expansion and a clear product mix shift. As we scale into the mass market, ADAS LiDAR naturally takes the largest share of volume. These products usually carry lower unit prices than robotics LiDARs, blended ASP comes down. Fundamentally, this is a healthy signal. It reflects LiDAR becoming a mainstream standard across both automotive and robotics. Looking ahead, we see a clear structural catalyst as the industry moves towards Level 3 autonomy. Level 3 raises the bar on safety, redundancy, and sensing performance, which directly drives higher sensor content per vehicle. At Level 2, LiDAR content per vehicle last year was roughly around $200, typically anchored by one long-range ATX.

At Level 3, it already steps up to about $350 as a starting point. For example, 1 ATX plus 2 FTX for blind spot coverage. As systems evolve further with ATX high-end LiDAR and additional blind spot FTX, contents can increase to roughly $500-$1,000 per vehicle. This is already emerging in early customer adoption and is increasingly reflected in ongoing engagements. What's more, with our newly launched Picasso 6D full-color SPAD SoC, LiDAR is also evolving from a passive safety sensor into an active intelligence layer, enabling richer perception through real-time colorized point clouds. That shift from safety only to active functionality creates a clear upgrade cycle, and we do see this value add potentially supporting higher pricing for advanced LiDAR configurations. Ultimately, we see the market evolving very much like smartphones.

Once users experience a better smartphone, they don't go back because they truly appreciate the value it creates in daily life. In the same way, as automotive and robotics applications become more intelligent over time, end users will increasingly recognize and cherish the value enabled by our technology, not as a cost, but as a capability upgrade in how they move, work, and interact with the physical world. That's my answer to the question you just raised.

Well, thank you, Andrew. Thanks for your insights. Thank you.

Your next question comes from Jia Lou with BOCI.

Hello, management. Thank you for taking my questions. I have a question regarding the competition dynamics. Over the past few years, the LiDAR industry has been focused on cost reduction. However, since beginning of this year, we've observed a shift towards performance race, such as escalation in channel counts, while the technology roadmap have begun to convert. How does management envision the evolution of the competitive landscape of LiDAR industry? What will be the key differentiator for Hesai to maintain its leadership? Thank you.

Thank you. That's a very good question. When people talk about LiDAR competition, the discussion often goes straight to price. Yes, the price of a single LiDAR has come down over the years, but that's really the result of strong in-house ASIC capability and system integration, which helped bring LiDAR from a luxury technology into mass adoption without sacrificing performance. That is exactly why today LiDAR can be deployed not only in $10,000 vehicles, but even in affordable home appliances. We are seeing penetration rapidly expand across sectors, including categories many people did not initially expect, such as robotic lawnmowers and two-wheelers. Once the cost performance curve crosses a certain inflection point, adoption doesn't grow linearly. It starts to expand across industries, and ultimately, the results speak for themselves.

According to Yole Group, we ranked number one globally in long-range ADAS LiDAR shipments in 2025, with 43% market share. In March this year, Gasgoo data shows our share further increased to 55%, roughly 3x that of the second player. Beyond automotive, we are also seeing strong momentum in robotics, where we believe we are already the number one global player in robotics LiDAR revenues across robo-taxis, robo-vans, lawnmowers, humanoids, and other emerging applications. The more important question is not price. It is who can lead across multiple technology cycles. Today, the industry is moving toward higher performance, higher resolution, new architectures, and new types of demand. We are already at 4,320 channel level products.

At the same time, we are expanding into blind spot LiDAR with full 360 degree coverage for large-scale deployments like the Li Auto's L series platform. We are moving towards deeper fusion of LiDAR and imaging with Picasso, our 6D full-color SPAD-SoC, as well as into physical AI with Kosmo for high fidelity spatial intelligence. Across all these shifts, our key advantage is the engineering capability and talent we have built over the past decade. We have built multiple distinct product architectures. From gas detection to mechanical LiDAR, semi solid-state, and the fully solid-state systems. While they all fall under LiDAR, they are fundamentally different in architectures. Today, we are extending that same capability into physical AI, including Kosmo and robotic actuation modules. At the end of the day, whether competition is in technology, products, or market share, we welcome it.

We remain confident in our long-term structural advantages and our ability to continue launching category-defining products with a strong pipeline of major launches in 2026. Our leadership in physical AI is, in our view, an important differentiator versus peers. Ultimately, this is a deep tech business. That's my answer to your question. Thanks for that.

There are no further questions at this time. I'll now hand back to Yuanting Shi for closing remarks.

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

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

Hello, ladies and gentlemen. Thank you for standing by. Welcome to Hesai Group's Fourth Quarter and Full Year 2025 Earnings Conference Call. [Operator Instructions] Please note that today's conference call is being recorded. I will now turn the call over to our first speaker today, Yuanting Shi, the company's Head of Capital Markets. Please go ahead.

Thank you, operator. Hello, everyone. Thank you for joining Hesai Group's Fourth Quarter and Full Year 2025 Earnings Conference Call. Our earnings release is now available on our IR website at investor.hesaitech.com as well as via Newswire services. Today, you will hear from our CEO, Dr. David Li, 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 press 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 over the call to our CEO, Dr. David Li. David, please go ahead.

Thank you, Yuanting, and thank you, everyone, for joining our call today. I'd like to start by taking a step back and looking at what we accomplished over the course of the year. 2025 was a defining year for Hesai. We achieved a milestone no other lidar company has reached, industry first full year GAAP net income of RMB 436 million. This was not just a year of growth, it was the year our technology leadership, operational scale and execution converged to set new standards for the industry. On the product front, we continue to lead the way. According to Gasgoo, ATX, our flagship ADAS lidar, largely contributed to our #1 position in 2025 with over 40% share of the long-range automotive lidar market. Meanwhile, our JT series entered mass production and shipped over 200,000 units in its first year alone, establishing clear leadership in Robotics as well. At the same time, we reinforced our financial position through a successful USD 614 million dual primary listing in Hong Kong, further strengthening our robust balance sheet and enhancing our capacity to support long-term growth. As we enter 2026, we are carrying significant momentum across markets. With demand accelerating across various key applications, we are raising our 2026 lidar shipment outlook to between 3 million and 3.5 million units. This reflects the massive scalability and resilience of our business. Now let's take a closer look at our business highlights, starting with our progress in the ADAS market. Currently, lidar is rapidly becoming what we call the invisible airbag, essential, affordable and increasingly standard. Over the past year, we have been a key force behind the broader rollout of lidar across the industry. We achieved 100% lidar adoption on best-selling models from partners, including Li Auto and Xiaomi, while also breaking into the sub RMB 100,000 price segment with Leapmotor. This marks a fundamental shift. Lidar is no longer a premium add-on, but a core safety feature in mainstream vehicles. Our momentum is also reflected in the strength and breadth of our partnerships. we have secured 2,026 design wins with key partners, including Li Auto, Xiaomi, BYD, Leapmotor, Great Wall Motors and Changan, many on an exclusive basis. Additionally, leading automakers such as BAIC and FAW Bestune are joining our SOP roster. Altogether, we have now secured ADAS orders from every one of the top 10 OEMs in China and have secured ADAS design wins with 40 automotive brands across more than 160 vehicle models, reinforcing our position as the partner of choice for world-class automakers. This leadership allowed us to go beyond a key milestone we first envisioned almost a decade ago, enabling 1% of all vehicles worldwide with 3D perception. With over 2 million cumulative ADAS lidars delivered, we are capturing over 40% of ADAS long-range lidar demand. This gives us significant manufacturing leverage and drives a powerful flywheel of innovation. To support accelerating growth at scale, we launched our revamped version of ATX lidar last November at our Tech Day event. Powered by our in-house FMC500 500 SoC, integrating MCU, FPGA and ADC; the revamped ATX features up to 256 channels, delivering enhanced performance, reliability and cost efficiency. With an order backlog exceeding 6 million units, it positions us strongly for the next phase of mass adoption and is expected to begin SOP in April 2026. While Level 2 drives volume, Level 3 is the value multiplier. In China, the regulatory environment has reached a pivotal inflection point. With Level 3 models now approved for public road deployment in cities such as Beijing and Chongqing, the industry is moving decisively from testing into real-world deployment. As responsibility shift from the driver to the OEM, zero failure has become a mandatory requirement. To manage complex driving scenarios, Level 3 systems need broader coverage with more lidars. This is where our FTX blind spot sensors come in, enabling full 360-degree perception. At the same time, Level 3 also demands better lidars, raising the bar on performance and reliability. Our ETX ultra high-performance long-range lidar is purpose-built for these demands. It offers around twice the detection range of ATX and will incorporate our proprietary SPAD, which eliminates the false triggers commonly seen in traditional SPAD architectures. ETX is expected to begin SOP by 2026. With recent multi-lidar design wins from Li Auto, Xiaomi and Changan, with SOP planned for 2026 to 2027, along with several late-stage Level 3 discussions underway with additional leading Chinese OEMs; we are seeing a meaningful increase in lidar content per vehicle as multi-lidar models typically feature 3 to 6 lidars per vehicle. This mirrors the evolution we saw in smartphone cameras, where increasing sensor count drove a steady expansion in total system value. We believe ADAS lidars is now entering a similar value creation cycle. Internationally, our business has also reached a critical inflection point. We are pleased to announce a strategic partnership with Grab, Southeast Asia's leading super app. With Grab as our exclusive regional distributor in Southeast Asia, we are combining Hesai's global lidar leadership with Grab's unparalleled local network to aggressively scale our footprint across the region. More significantly, we have been selected as the primary lidar partner for NVIDIA's DRIVE Hyperion 10 platform, which we view as a true game changer in how we scale globally. Historically, international expansion in automotive was a slow OEM by OEM process, often taking years of validation and negotiations. Integration into the Hyperion ecosystem enables a fundamental shift in our go-to-market approach from individual engagements to a scalable turnkey model. This positions Hesai as the default gold standard lidar choice for OEMs building autonomous driving systems on the NVIDIA platform. Additionally, we have joined NVIDIA Halo AI Systems Inspection Lab to further advance safety in autonomous vehicles and robotics. Building on our momentum, our exclusive multiyear design win with a top European OEM is progressing well, with sample deliveries firmly on track. More importantly, we've achieved a key breakthrough, unifying our high-performance lidar architecture across China and global markets with the ET series as a prime example, enabling a single platform to scale seamlessly worldwide. This unified architecture eliminates redundant development while combining China's operational agility and cost advantages with the most stringent global quality standards. In fact, Hesai is the only Asian lidar manufacturer with German VDA 6.3 process audit certification, a globally recognized benchmark for the industry's most rigorous production and quality standards. The result is a structurally advantaged one platform model, delivering superior cost, speed and global scalability that is extremely difficult to replicate, putting us firmly in the driver's seat of global expansion. Looking ahead, 2026 is going to be a pivotal year for the evolution of intelligence. As NVIDIA's CEO, Jensen Huang, described at this year's CES, we are entering the ChatGPT moment for physical AI, a shift from digital chatbots to kinetic work bots operating in our factories, streets and homes. If 2025 was the year AI learned to reason, 2026 is the year AI gains a body. However, for AI systems to truly reason about the physical world, it requires a grounding in geometric truth. While cameras provide the context or the what, lidar provides the sub-centimeter spatial accuracy, the where. This makes lidar an indispensable bridge between the carbon-based world and silicon-based intelligence. Without the spatial intelligence, physical AI remains blind to the loss of physics. This structural shift plays directly to our strengths and the results are already very encouraging. According to GGII, Yole Group and Frost & Sullivan, we now rank #1 across multiple major robotics lidar submarkets, spanning humanoid and quadruped robots, robotaxis, robovans and robotic lawn mowers. For example, our JT128 lidar showcased this leadership at the 2026 Spring Festival Gala. During China's largest broadcast, which peaked at 400 million viewers, dozens of unitary humanoid robots delivered a complex synchronized [ kung fu ] performance. By providing 360-degree blind spot-free precision perception and ultra-high reliability, JT128 lidar outperformed competing offerings, seamlessly integrating with [ Unitree ] AI algorithms to achieve ultra-low latency and eliminate cumulative motion errors, ensuring absolute stability. Beyond human robotics, we have also established a strong market position in robotic lawn mowers. We have secured orders from clients, including Dreame and MOVA, representing a backlog of over 10 million lidar units with strong follow-on potential as deployments scale. In Robotaxis, we now work with nearly every leading player, including Pony.ai, WeRide, Baidu Apollo Go, DiDi and others across North America, Asia and Europe. In Robovans, we have almost achieved full coverage of key players like Zelos, Neolix and Meituan. Beyond these segments, we are actively expanding lidar applications. Recently, we secured a design win for NIU Technologies next-gen electric 2-wheel model featuring our FTX lidar. With over 10 million electric 2-wheelers sold annually in China, this brings automotive-grade 3D perception to a massive market and unlocks a new intelligent category. Together, these fast-growing segments put us right at the heart of the robotics ecosystem, helping bring physical AI from concept to real-world action. After shipping nearly 240,000 robotics lidar units in 2025, we expect that volume to at least double in 2026. Lastly, I'd like to share what's next for Hesai over the coming decade and why we are genuinely excited about the opportunities ahead. The physical AI revolution is accelerating at an unprecedented pace, but many of its critical building blocks are still in their early stages, such as sensing, motion control, integrated AI-driven decision-making and full system orchestration. These gaps represent enormous white space opportunities, and they are exactly where we believe the next wave of transformative growth will unfold over the coming decade. Hesai is uniquely positioned to lead this next phase. We bring decades of expertise in lidar, automotive and robotics-grade hardware. Today, we are doing far more than building components. We are evolving into the key enabler of physical AI, digitizing the real world and redefining how humans and robotics perceive and act. This positions us at the forefront of the AI-driven fourth industrial revolution and perhaps more importantly, opens the door to a decade of exponential opportunity. Let's now move on to something more immediate. In the next few months, we will launch two groundbreaking products, each targeting an addressable market worth trillions of RMB. One is the eyes of physical AI, enhancing perception and situational awareness beyond what is currently possible. The other is the muscles, delivering precise powerful motion control for robots and autonomous systems operating effectively in the real world. Together, these products are expected to become Hesai's second growth engine. We anticipate initial revenue contributions beginning as early as 2026. Within 5 years, this business has the potential to rival or surpass our lidar segment and within a decade, to scale another tenfold. This is more than a product portfolio expansion. Guided by our mission to empower robotics and elevate lives, we are entering the next chapter of our growth story to become the key enabler of physical AI. If 2025 was a year of market validation and record performance, 2026 will be a year of acceleration and transformation. The opportunity ahead is massive, and we are ready to lead the way. With that, I will now turn the call over to Andrew to discuss our financial performance and outlook. Andrew, please go ahead.

Thank you, David, and hello, everyone. Let me start by walking you through our full year operating and financial performance and share our thoughts and outlook for 2026. To be mindful of the length of our call, I encourage listeners to refer to our earnings release for further details. 2025 was a pivotal year for Hesai, marked by remarkable progress in both our financial performance and operational execution. We delivered record net revenues of over RMB 3 billion or USD 433 million, representing an increase of 46% year-over-year. This performance was underpinned by a substantial ramp in our production volumes, with total shipments exceeding 1.6 million units, more than tripling from last year, including nearly 240,000 units from robotics lidar. This expansion reflects both robust demand across markets and our ability to execute consistently and reliably at scale across a broad range of applications from passenger vehicles, humanoid and quadruped robots to robotaxis, robovans, robotic lawn mowers and many more. Together, these have reinforced our position not only as a global volume leader, but also as the partner of choice for high-value, mission-critical applications. Beyond strong top line growth, we also significantly improved the quality of our financial performance. Gross margin remained healthy at over 40%, while operating expenses, excluding other operating income, came down RMB 88 million or USD 13 million despite substantial revenue growth. This reflects strong operating leverage supported by our disciplined cost management as well as efficiency gains enabled by AI across R&D, manufacturing and operations. These improvements flowed directly to the bottom line, enabling Hesai to achieve industry-first full year GAAP profitability with net income of RMB 436 million or USD 62 million. Full-year GAAP net income, excluding after-tax gains from equity investments of RMB 148 million or USD 21 million was RMB 288 million. or USD 41 million. On a non-GAAP basis, full year net income reached RMB 551 million or USD 79 million, with the difference from GAAP net income mainly driven by stock-based compensation. Excluding after-tax gains from equity investments, full year non-GAAP net income was RMB 403 million or USD 58 million. Kindly note that we have already delivered GAAP net income for 3 consecutive quarters and non-GAAP net income for 5 consecutive quarters, demonstrating the sustainability of our earnings performance. Just as importantly, this profitability was paired with strong cash generation. We delivered positive operating cash flow of RMB 117 million or USD 17 million during the year, marking our third consecutive year of positive operating cash flow, while our net assets grew to around RMB 9 billion or USD 1.3 billion. Today, we operate with the most robust income statement and balance sheet in the global LiDAR industry, reflecting our ability to scale technological leadership while maintaining a solid financial foundation. Building from this position of strength, we are entering 2026 with a dual focus, scaling lidar leadership while proactively expanding into new growth opportunities. We expect our core lidar business to deliver shipments of 3 million to 3.5 million units in 2026. This expanded scale will reinforce our operating leverage, supporting sustainable profitability and steady cash generation. At the same time, we expect to maintain resilient gross margins through ongoing innovation and disciplined operations. Additionally, and perhaps most excitingly, 2026 marks the beginning of commercialization for our new state-of-the-art products, which we believe will become the second growth engine for Hesai in the next decade. As we invest to advance these strategic priorities, we expect to drive a strong and resilient bottom line as we scale in 2026. For the first quarter of 2026, we expect net revenues to be between RMB 650 million and RMB 700 million or USD 93 million to USD 100 million, representing year-over-year growth of approximately 24% to 33%. We also expect revenue momentum to strengthen progressively each quarter throughout the year. To conclude, 2025 was a pivotal year that enhanced the quality and scale of our business. Building on this momentum, we are positioning to become the key enabler of physical AI, digitizing the real world, redefining how humans and robotics perceive and act. As we scale, our goal is clear: to build a globally competitive technology leader grounded in innovation and financial rigor, creating sustainable compounding value to our shareholders and the broader ecosystem. This concludes our prepared remarks today. Operator, we are now ready to take questions.

[Operator Instructions] Your first question comes from Tina Hou with Goldman Sachs.

Congratulations on raising the volume guidance. And also, look forward to the new product launch. So my question is mainly focused on the Robotics business. Wondering if management can give us more details about the different verticals, including robotaxi, robovan as well as humanoid robot. How do you see the businesses pan out in 2026 and then beyond?

Thank you, Tina. It's Andrew here. I will take this question first. As David just quoted Jensen Huang's speech at CES, 2026 marks the ChatGPT moment for physical AI, where our lidar provides the crucial sub-centimeter special accuracy, serving as the indispensable bridge between the carbon-based world and silicon-based intelligence. Because of this structural shift, our Robotics business is truly blooming everywhere. We are incredibly proud to share that according to industry trackers, Hesai is now ranked #1 across major robotics lidar submarkets. Let me take a moment to walk you through the key Robotics verticals that may be of interest to our investors. First, humanoid and quadruped robot. We see humanoid and quadruped robotics as a significant long-term opportunity. At the core of this vision is the need for precise perception and action as any robot interacting dynamically with the physical world relies on accurate sensing. This makes lidar a critical and ultimately standard component for positioning, navigation and obstacle avoidance. We are currently ranked #1 in humanoid and quadruped robot segment according to GGII and have secured orders from leading players, including Unitree, HONOR Robot, Galbot, Magiclab and Vita Dynamics. We expect annual shipment in this segment to reach 5-digit levels in 2026. Our JT128 lidar was deployed across Unitree's robot at the 2026 Spring Festival Gala and was selected for its superior range and reliability, enabling large-scale synchronized movements with high precision and stability. Second, robotaxi. Hesai is the world's largest robotaxi lidar supplier according to Yole report. Our main and blind spotting lidars are widely deployed among Chinese leading players, including Pony.ai, WeRide, Baidu Apollo Go, DiDi and Hello. Globally, we have secured a supply agreement with a wide area of top autonomous driving companies across North America, Asia and Europe. In short, we collaborate with nearly every key player worldwide, an important differentiator from our peers. Whether ADAS or mechanical lidar solutions are selected by robotaxi customers, our revenue model scales with their fleet size, number of lidars per vehicle and ASP. As leading operators accelerate large-scale deployments, we expect exponential fleet growth to drive rapid revenue expansions for Hesai. For robotaxis, we anticipate 5 to 10 lidars per vehicle to ensure a full 360 degrees coverage. Thirdly, robovan. The robovan sector is undergoing a major transformation. No longer limited to closed campuses, robovans are increasingly operating on complex urban rails. Supported by favorable government policies and proven business models, the market is projected to scale from 5 digits to 6 digits of robovans in 2026. Each robovan typically features 2 to 6 lidars. Hesai is ideally positioned to capture this growth. We are the core LiDAR supplier for leading robovan players globally, including Zelos, Neolix, and Meituan and DoorDash, serving as the sole supplier for many. GII recently ranked Hesai #1 in lidar design wins for this sector. Several players that previously relied on competitors' products are switching to Hesai this year, underscoring our role as the go-to hardware partner in the accelerating commercial robovan market. Fourthly, robotic lawn mowers. The robotic lawn mower market is a major growth opportunity for our Robotics business. Global annual lawn mowers sales reached about 20 million units, yet lidar-equipped robotic lawn mowers account for just 1% to 2%, highlighting a huge untapped market as consumers adapt smarter, hands-free yard care. Hesai is moving aggressively to capture this space. Since launching the JT Series 3D lidar at CES 2025, cumulative deliverables have already exceeded 200,000 units by 2025, supported by strong global partnerships with leading brands, including Dreame, MOVA and Nexlawn. We recently secured a milestone agreement to exclusively supply 10 million JT lidars to Dreame and MOVA, ranked #1 globally by Frost & Sullivan for lidar robotic mowers in 2025. This record-setting order signals a fundamental industry shift, establishing lidar as the standard for high-end smart yard products and ushering in a new era of outdoor robotics perception. As a quick summary, across these diverse applications, our Robotics business sits at the heart of the ecosystem, consistently delivering relatively higher ASPs and strong margins. After shipping 200,000 Robotics lidar units in 2025, accelerating momentum across these segments gives us full confidence that volumes will at least double in 2026. In the long term, new types of robots will begin to adopt lidar. For example, new technologies, 2-wheel scooters recently integrated our FTX lidar for autonomous operation. The robotics market could have a TAM several times larger than ADAS. After all, you can drive only one car, but in the future, 10 robots could be working alongside you. Tina, that's my answer to the question just raised.

Your next question comes from Tim Hsiao with Morgan Stanley.

This is Tim from Morgan Stanley. Congratulations on the strong results and sustained industry leadership. I just want to have a quick follow-up questions also about robotics market because the market is apparently very interesting, exciting and highly focused by investors. But we noticed that the founders of Hesai have also invested in a company called Sharpa, which has been gaining a lot of attention recently. So just want to understand how should we view the relationship between Hesai and Sharpa? And is there any opportunity for business cooperation with Hesai within that year? And how does management view the future technology and supply chain synergies between the two entities? That's my question.

Thank you, Tim. Thank you for the question. And it's actually a great topic, and I also wanted to offer from my side. First, I want to clearly define the structural relationship. Hesai and Sharpa are two fully independent operating entities. There is no relationship of equity subordination or operational control between them. What were the co-founders of Sharpa were responsible solely for strategic guidance at Sharpa as a core shareholder role, and we do not hold executive position for actual operational growth. Our primary and full-time identities remain the CEO, CTO and the Chief Scientist of Hesai, and our focus and energy are dedicated to Hesai. Sharpa possesses its own mature and independent team. While looking ahead, we remain open to future collaborations where it makes strategic and commercial sense as it can create an actually compelling win-win dynamic. Both companies can apply their technologies in real-world scenarios while benefiting from shared insights and industry-leading expertise. For example, as an AI robotics company, Sharpa may utilize Hesai's products while Hesai as a hardware innovator may explore deploying humanoid robots in its automated production line over time. At the same time, Sharpa's progress in AI could broaden the perspective of founders and the Hesai team and potentially inform our long-term innovation road map. That said, I want to emphasize that the coordination, if any, in the future; will be conducted strictly on fair and market-based terms with the objective of maximizing long-term value for Hesai's shareholders. Based on our preliminary estimate of the future humanoid robotics market and the growth of Hesai, such operation will only contribute a small portion of our business. As Hesai continues to evolve into a key enabler of physical AI, digitizing the real world, redefining how human and the robotics perceive and act, we remain focused on executing our core strategy. This includes strengthening our leadership in lidar, advancing our next-generation eyes and muscles product portfolio and driving sustainable long-term growth by building out the Hesai ecosystem. We believe the addressable market we're targeting over time expand well beyond the traditional lidar segment, and we're truly excited about the journey ahead. This is hopefully helpful information to help you understand what Hesai and Sharpa each are trying to do and the possible synergies and the collaborations between the two entities. Thank you, Tim, for the question.

Your next question comes from Jeff Chung from Citi.

This is Jeff from Citi. First of all, congratulations, fantastic results. So my first question is that we have the first quarter revenue guidance. So could you give us more color on the first quarter volume guidance? And separately, we recognize Hesai did a great job with the sequential OP margin improvement in the past 4 quarters. Could you give us more color on the first quarter and the full-year GP margin and OP margin guidance?

Thank you, Jeff. I will take this question, and I'll try to address our 1Q and the full year guidance for 2026 to the extent I can. For the first quarter of 2026, we expect the total revenues to be between RMB 650 million to RMB 700 million, representing a solid year-over-year growth of approximately 24% to 33%. On the volume side, we anticipate total shipments to be in the range of 400,000 to 450,000 lidar units, including around 100,000 units from Robotics. We have delivered GAAP net income for 3 consecutive quarters and non-GAAP net income for 5 consecutive quarters and expect to maintain this momentum. It is important to note that due to typical automotive industry seasonality and the timing of the holidays, we do expect a sequential decrease in deliveries compared to the seasonal high we saw in the fourth quarter last year. This is entirely consistent with our historical patterns. However, the fundamental demand for our lidars remains exceptionally strong in 2026, and we expect both revenues and shipment volumes to increase sequentially over the course of 2026. We are highly confident in our accelerating momentum and our ability to maintain a healthy financial profile as we execute our 2026 road map. Looking ahead to 2026, we see it as a true inflection point. On one hand, we anticipate strong demand for lidar in both passenger vehicles and robotics, which is expected to drive meaningful increase of our full year 2026 revenues. Correspondingly, we are raising our shipment guidance to a record 3 million to [ 3 ] million units for this year, with both ADAS and Robotics lidars expected to roughly double year-over-year. While volume is scaling rapidly, we do anticipate a potential decrease in blended ASP. That's mainly due to, first, modest volume-based pricing and standard annual decline for our larger order strategic OEM customers. That's mainly for the ADAS products. And secondly, a shift in product mix towards certain lidar products with a relatively lower unit prices, such as the AT series, FT series and JT series, typically around 1 to couple of hundred U.S. dollars each. Though these products will account for a larger share of deliveries and revenue compared with our traditional high ASP Robotics products such as Pandar and XT Series. That said, we are highly optimistic about our top line and margin resilience because of several strong positive catalysts accelerating in 2026 and 2027. First, lidar is rapidly transitioning from an optional add-on to a standard configuration, successfully penetrating the mass market for vehicles priced between RMB 100,000, which will continuously drive up overall the penetration rate. Second, Level 3 vehicle deployment in China will drive multi-lidar setups, pushing lidar content per vehicle to as high as $500 to $1,000 range. We have already secured multi-lidar design wins with our core customers, including Li Auto, Xiaomi and Changan, featuring 3 to 6 lidars per vehicle, with SOP planned for 2026 to 2027. Third, our overseas ADAS business is expected to start contributing in as early as 2026, marking the beginning of global ADAS lidar mass adoption with international ADAS programs typically carrying higher ASPs. We expect our partnership with NVIDIA to roll this game forward. Fourth, our Robotics business continues to gain momentum across diverse applications and customers, and it typically carries a relatively higher ASP and margin compared to ADAS. Finally, our newly second growth engine, the eyes and the muscles of physical AI, will serve as a powerful new driver for our long-term growth. Stay tuned for two new products that we plan to launch in the coming months, each targeting at RMB 1 trillion TAM. On the profitability front, through continued cost optimization across ASIC design, supply chain and manufacturing and with the launch of our FMC500 SoC to improve cost structure in ADAS products, we expect our group blended gross margin to remain resilient in 2026, despite a strong increase in ADAS lidar shipments. As a result, we are confident that the profits from our core lidar business will continue its solid growth trajectory. Additionally, and perhaps most excitedly, 2026 marks the beginning of commercialization for our new state-of-the-art products, which we believe will become the second growth engine for Hesai in the next decade. In short, we are entering 2026 with accelerating shipments, robust revenue growth, a highly disciplined margin profile, solid bottom line increase and exciting new growth engines. Jeff, that's my answer to your question. Thanks for that.

Your next question comes from Aaron Wang with Jefferies.

This is Aaron from Jefferies. I just have a quick question on our guidance. Last year, we had a profit guidance for the full year. I was wondering if the company will also provide a full year net income guidance for 2026.

Thank you, Aaron. Given the differences in compliance requirements and listing rules between the U.S. and Hong Kong market, as you know, we just listed in Hong Kong, and to align with the best disclosure practices for dual listed companies; we have decided not to provide specific full year net income guidance at this time. However, I want to emphasize that this adjustment in disclosure does not reflect any lack of confidence in our business. On the contrary, underpinned by our solidified customer base, undisputed industry dominance and a disciplined cost structure, we are fully confident in maintaining our growth trajectory of revenues, shipments and profits in 2026. At the same time, we highly encourage investors to look forward to our new business initiatives. We are investing strategically in these areas, and they are expected to become Hesai's second growth engine. Aaron, that's my answer to your question.

Your next question comes from Nora Min with UBS.

This is Nora from UBS. Thank you for trusting me with the most exciting question. So what is the master plan behind this non-auto, non-lidar new product? Would you share with us a bit of timeline, a bit of progress, a bit of more detail?

Thank you, Nora. Our guiding mission has always been to empower robotics and elevate lives. We have never defined ourselves solely as a lidar company. So as Jensen Huang and David repetitively mentioned, we believe 2026 will be the ChatGPT moment for physical AI. We are entering an era where AI will truly understand the rules of the physical world and learn to interact with it, which will trigger a big bang in robotic applications. The physical AI revolution is accelerating at an unprecedented pace, but many of its critical building blocks are still in their early stages. These gaps represent enormous white space opportunities. Hesai is repositioning its role in the new era as a key enabler of physical AI, digitizing the real world, redefining how humans and robotics perceive and act. In the next few months, we will launch two groundbreaking products, each targeting a mass trillion RMB market. First, the eyes, enhancing perception and situational awareness beyond what is currently possible; and second, the muscles, delivering precise, powerful motion control for robots and autonomous systems. Regarding the financial outlook of these two new products, we anticipate initial revenue contributions from these new products beginning as early as 2026. Within 5 years, we expect this business to rival or even surpass the scale of our current lidar segment. Within a decade, it has the potential to scale another tenfold lidar -- larger. As we expand into the broader physical AI ecosystem with our upcoming eyes and muscles products, our core advantages come from strategic foresight and a decade of lidar mass production experiences. We tackle challenges head on, refining our products to lead in performance, quality and cost simultaneously. Building the muscles of physical AI naturally extends our expertise in materials, simulation, design and precision manufacturing, backed by our proprietary ethics, in-house production and rigorous quality management to deliver reliability, scalability and extreme performance at scale. For the eyes, our strength comes from world-class software and algorithm capabilities seamlessly integrated with our hardwares. Years of R&D in 3D risk construction and rendering recently earned us an award at the September 2025 SIGGRAPH Challenge, a premium global event showcasing the pinnacle of computer graphics. We are truly excited to bring these best-in-class algorithms to life in our soon-to-launch hardware tools. Together, all these capabilities let us push physical boundaries and raise performance ceilings, supporting our new positioning as Hesai evolves into a key enabler of physical AI, digitizing the real world, redefining how humans and robots perceive and act. Nora, that's my response to your question.

Your next question comes from Jessie Lo with Bank of America Securities.

This is Jesse from Bank of America Securities. Congrats on the great results. So my question is surrounding the NVIDIA cooperation. So following the announcement of us selected as the partner for NVIDIA DRIVE AGX Hyperion 10, what are the next steps? Or what could we expect to see in the coming years? And what differentiates us from the peers in this collaboration?

Yes. This is David. Maybe I will offer some insight and our interpretation of such a collaboration. So first of all, I wanted to just help people understand the NVIDIA -- the platform is beyond only the computational hardware. It's the full stack solution, meaning it's the hardware, the software and the data. And then we are the selected partner for lidar. What that means is that, obviously, in the end, NVIDIA has customers and the customers will ultimately decide the vendor. But as we are already selected by NVIDIA. It just makes this process a lot easier in the following way. And the first is that the system, the sensor setup, the computation is already a complete system, and that's proven. Obviously, let's say, somehow you wanted to pick a different vendor who's not on the list, it just make it harder for you to verify that. And then that's actually the smaller part of the problem. The much bigger problem is the rest, including training the model and also data collection and the verification of such a system, right? You can imagine for NVIDIA to provide such a full stack solution to all the robotaxis and the Western OEMs that they're working with you need a large amount of data they already have for the project we already have with NVIDIA, that's with Hesai lidar. So now let's say, for whatever reason, obviously, I will not understand or support, but they have to use a different lidar. And then you immediately have to face the challenge that what do you have to do with the model we train and the data we collected for the past actually years of collaboration. So it's not impossible, it just make it very inefficient if you had to do that. So -- and that's one of the reasons that we are super excited and definitely honored to be in this program. And we're also motivated to work alongside with NVIDIA when they are working with customers globally, promoting such a unified solution. And to me, this actually is the smartest way to push the autonomy because in the end, it's less important if you have different components. It's important that if you have one solution that works and then try to utilize and not reinvent, we will utilize the same solution with all those customers. And I do believe NVIDIA also shared the same vision, and that's why we are now working very closely with them in supporting them with our latest sensors and in qualifying them in different parts of the world and try to develop and accelerate the programs they already have and the new programs that they will be signing.

Your next question comes from [indiscernible] with CICC.

This is Dani from CICC. Congrats on your strong results last year. I have two questions for you. The first one is about the price. What's your outlook on the trend of your ASP decline? And my second question is, could you please share more color on your methods for further cost reduction?

Thank you, Dani. We wouldn't suggest our investors read too much into total and blended ASPs. It's really just a simple math, and the decline is mainly driven by product mix. ADAS lidars, which are generally lower priced than our Robotics lidars, are taking a bigger share. As Level 3 ramps up, our blind spotting FTX lidars, which are lower priced than long-range ATX lidars, will push blended ASP down further. So this isn't really about price, it's mostly about mix. That said, ADAS lidars follow the typical annual automotive price declines. For example, ATX is expected to carry a price tag around $150 in 2026, which is already near an optimized cost structure. So we expect the future declines to narrow. Over time, this will be offset by structural growth, more lidars per vehicle and high-performance, higher-priced L3 products like ETX as well as expansion beyond China. Looking ahead, we see clear and durable pathways to further reduce lidar costs driven by scale, technology and manufacturing excellence. First, scale is a powerful lever. After delivering 1.6 million units in 2025, tripling year-over-year, we are guiding 3 million to 3.5 million units in 2026. This step-change in volume will meaningfully dilute fixed costs and strengthen our supply chain leverage. Secondly, our proprietary technology, chip technology is structurally lowering our BOM. With 100% in-house development of core modules and our FMC500 SoC integrating MCU, FPGA and ADC functions, we are replacing costly discrete components with a highly efficient single-chip solution, reducing cost while improving performances. Meanwhile, our in-house fab integration is expected to ramp by 2026, further improving our cost structure. Finally, our highly automated in-house manufacturing drives compounding efficiencies. By standardizing core architectures and continuously improving yields, we expect to maintain a healthy margin profile in the future. Thank you, Dani.

Your next question comes from Frank Tao with CMBI.

I'll add my congrats on the upbeat shipment volume guidance as well. Could management share with us your outlook for the operating expenses in the year of 2026?

Thank you for raising this question. We are very pleased with our progress in expense management, as we guided at the beginning of last year. For year 2025, our operating expenses actually came down by RMB 88 million despite our substantial revenue growth. This clearly reflects the strong operating leverage in our business, supported by our highly disciplined cost management. A key driver behind this is that AI is at the heart of how we work. We firmly believe that any company not fully embracing AI in 2026 will inevitably be left behind by the market. Because of this, we will continuously and aggressively embrace AI to boost our operational efficiency, transform our workflows and strengthen profitability. So far, this proactive approach has already delivered tens of millions of renminbi in measurable cost savings and significantly improved our productivity. Looking ahead to 2026, we anticipate a modest mid-teen increase in overall OpEx, primarily due to RMB 200 million invested in new eyes and muscles products in R&D. Otherwise, excluding new business spend, OpEx is expected to be well managed, flat or even down in single digits in 2026, demonstrating our discipline and AI adoption. Thank you.

That concludes our question-and-answer session. I'll now hand back to Yuanting Shi for closing remarks.

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