業績會第一現場

was live · ·

地平線機器人2025年全年業績直播

Key Takeaways (AI-Generated)

财务表现

- 2025年总收入达人民币37.6亿元，同比增长57.7%

- 汽车业务收入占总收入近95%，毛利率67.2%

- 全年整体毛利率维持在64.5%

- 产品解决方案收入激增144%至16亿元，受交付量增38%和单价增75%推动

业务进展

- 以47.7%市场份额成为中国本土品牌ADAS解决方案提供商第一名

- 汽车级SoC解决方案年出货量超400万颗，同比增长39%

- 支持NOA功能的SoC出货量占总量45%，为2024年同期近5倍

- HSD于2025年11月正式量产，一个月内出货量超2.2万颗

下一季度业绩指引

- 未来几年平均收入增长率有信心达到60%

- 2026年汽车业务收入增长预计加速至约60%

- 2026年SoC出货量预计增长约35%，AD SoCs占比增至55%以上

- 2026年HSD出货量预计达约40万颗

机会

- 中国乘用车市场智能辅助驾驶渗透率达68%，高阶车型占43%

- 推出中国首款全车AI智能体SoC用于座舱驾驶融合

- 与大众集团Cariad深度合作及合资公司成立

- 座舱驾驶融合方案预计显著降低系统级成本

风险

- 在高阶智能辅助驾驶市场面临英伟达和华为竞争

- 内存价格上涨给汽车制造商带来成本压力

Full Transcript (AI-Generated)

Operator

Greetings ladies and gentlemen. Welcome to Horizon Robotics 2025 Annual Results Conference Call. At this time all participants are in a listen only mode. A question and answer session will follow the formal presentation. As a reminder, this conference is being recorded.

I would now like to turn the call over to your host, Miss Zhuan Jiang, VP and Board Secretary and Head of Investor Relations of Horizon Robotics. Thank you, Zhuan. You may begin.

Zhuan Jiang

Thank you, ladies and gentlemen. Welcome to the investor conference call and audio webcast hosted by Horizon Robotics. On today's call, we also have Doctor Kai Yu, Founder, CEO and the Chairman of the Board of Horizon Robotics and Mr. Lei Wang, CFO and Head of Capital Market of the Company.

We will begin with Doctor Yu sharing our latest strategic and business developments. This will be followed by latest review of the company's financial performance in 2025. We will then proceed to the Q&A session.

Before we start, we want to remind you that this call may include forward-looking statements, which are underlined by a number of risks and uncertainties and may not be realized in the future for various reasons. Information about general market conditions come from a variety of sources outside of our robotics. This presentation also contains some unaudited non FRS financial measures that should be considered in addition to the company's financial prepare.

I will now turn to the call over to our CEO, Doctor Yu.

Doctor Kai Yu

Good evening, dear investor friends. I am Yukai, Founder and CEO of Horizon Robotics. I'm very glad to meet everyone online again and report to you on our business performance for the full year of 2025.

First, I would like to share with you the latest trends in the entire intelligent assisted driving industry. The penetration rate of intelligent assisted driving in China's passenger car market reached a new high of 68%. Among these, advanced intelligent assisted driving models represented by highway NOA and urban NOA already accounted for 43% of all intelligent assisted vehicles.

Compared with 2024, the penetration rate of various NOA intelligent assisted driving functions doubled within just one year. To put it more intuitively, this means that for every 3 passenger cars sold in the market last year, 2 featured intelligent assisted driving and 1 was equipped with NOA intelligent assisted driving system.

In the continued rise of the intelligent driving industry, Chinese domestic brands and Horizon Robotics have played a core leading role. Last year, among the intelligent vehicles sold by Chinese domestic brands, advanced intelligent assisted driving models accounted for as high as 62% of sales, far higher than the 13% level of joint venture brands.

Everyone knows that the mainstream market for cars priced under CN¥200,000 forms the foundation of China's auto market, accounting for 65% of total vehicle sales in China. It is within this price range that Chinese domestic brands have experienced an explosive growth in mid to high level intelligent assisted driving systems, with a sales share of NOA level models rising from 5% at the beginning of 2025 to over 50% by the end of the year.

In this largest sales segment, where the proportion of intelligent assisted driving is steeply increasing, Horizon is undoubtedly the driving engine. Based on the new generation Journey 6 series SoC advanced intelligent assisted driving solutions, we have captured 44% of the market ranking first in market share in the first year of mass production, which has also filled us with confidence to further break into higher price ranges in the future.

Further, in the price range around CN¥150,000, Horizon has promoted the mass production of multiple models equipped with urban NOA functions. This shows that the Chinese auto market has not only achieved democratization of intelligent assisted driving, but with joint effort by us and car makers has begun to stride into the new stage of democratization of advanced intelligent assisted driving.

From Horizon's perspective, this leap is of great significance. On one hand, the equipped models are penetrating into the mass market, bringing an exponential expansion of the potential market size for urban NOA with a market space of tens of millions of vehicles in China alone. Meanwhile, by replacing basic ADAS SoC that cost only 20 to 30 U.S. dollars per vehicle with our high value added intelligent assisted driving software system, we have gradually increased the per vehicle value to the scale of hundreds of U.S. dollars.

The market is cascading down and the per vehicle value is rising. This exciting trend is unfolding fully under our promotion. The urban NOA intelligent assisted driving experience that was once limited to luxury cars is now becoming a standard right for mass users.

Against this background, we are pleased to report to our investor friends that Horizon achieved the following major milestones during the reporting period. First, in 2025, with a market share of 47.7%, we secured the absolute lead as the number one Chinese domestic brand ADAS solution provider continuing to lead the industry.

At the same time, in the rapidly booming domestic brand advanced intelligent assisted driving market with a market share of 14.4%, we are almost tied for second place with Huawei which has a market share of 15.2%. Together with NVIDIA we formed the first tier of the market, with a combined market share of the three companies reaching 89%, showing significant concentration at the top and forming a one strong player and two powerful players market pattern.

In the largest space and fastest growing intelligent assisted driving segment under the mainstream price range of CN¥200,000 which I briefly introduced, Horizon has already captured 44% of the Chinese domestic brand NOA market firmly holding the industry's top position.

Second, in 2025, our automotive grade SoC solution annual shipments exceeded 4 million units with a year on year growth of about 39%. The biggest highlight was the optimization of the structure benefiting from the promotion of democratization of intelligent assisted driving by downstream automakers. Our SoC shipments supporting NOA functions have already reached 45% of the total, nearly five times the shipments during the same period in 2024, which can be described as explosive growth.

This product upward structural change while optimizing our SoC product portfolio has also driven our ASP to increase by more than 75% compared with 2024. Once again achieving high quality growth with both volume and price rising. I want to emphasize that this is rarely seen in the industry.

From another perspective, during the reporting period, more than 95% of the SoC shipped that support NOA function were jointly delivered through ecosystem partners, which fully highlights our position as the ecosystem leader of the industry and proves the success of our open ecosystem strategy.

Third, our all scenario urban assisted driving solution Horizon Super Drive or HSD officially entered mass production in November last year, first launched in the mainstream market around CN¥150,000, becoming China's first mass produced intelligent assisted driving large model based on one stage end to end technology. Within just over a month's time, it quickly surpassed 22,000 units in shipments, successfully bringing high level intelligent assisted driving experiences to mass users.

In addition, we have observed a trend indicator. Horizon's HSD has already become an important factor influencing end consumers' car buying decisions. Everyone knows that the traditional sales structure of the automotive industry often takes the shape of a pyramid, with mid and low tier configurations responsible for volume and top tier models selling less. But for the models in 2025 that first launched with our HSD, the top tier models with HSD as a core configuration accounted for as much as 83% of sales, breaking this pattern.

What's even more exciting is that not only are users willing to pay for it, but the usage stickiness is also very high. During this Chinese New Year, the ratio of intelligent assisted driving mileage among intelligent driving users of this model reached 41%, meaning HSD driving mileage is nearly equal to that of human drivers. That is a strategically significant number.

In the past, people ask me under what conditions Horizon might consider charging fees through a subscription model. In my view, once the proportion of intelligent assisted driving mileage surpasses the critical point of 50%, meaning mileage dominated by machines exceeds that of humans, users' reliance on intelligent assisted driving will become irreversible. This high product stickiness provides conditions for us to launch SaaS subscription charging model in the future.

Currently, the company's revenue is mainly driven by sales of new cars equipped with our products. In the future, we expect to leverage the large installed base of HSD equipped vehicles to collect service subscription fees through the full life cycle.

Fourth, regarding our global strategy, I think it can be summarized in three sentences: Help Chinese domestic brands go global, help joint venture brands stay, and help overseas brands step into the intelligent era. Regarding going global, by 2025, we have already cumulatively secured mass production nominations from 11 automakers and over 40 overseas models with a cumulative life cycle export nominations reaching 2 million units.

Well known top five Chinese automakers going global have our technology equipped on their mainstream overseas models. It can be said that in the current process of Chinese automobiles going global, Horizon has already become one of the core forces enabling their overseas expansion.

Secondly, regarding staying, how can joint venture brands develop better in the fiercely competitive domestic market? Currently, we have cumulatively obtained nominations from 9 joint venture brands for more than 35 models in the domestic market. As everyone knows, we have deep cooperation with Volkswagen Group's software company Cariad and have established a joint venture. So it is not surprising that we have secured many Volkswagen models.

But it is worth emphasizing that among the joint venture brand models we mentioned, over 35 nominations, the proportion of models outside of Volkswagen Group has already exceeded 60%, reflecting Horizon's strong trust among many joint venture brands.

Finally, it is about making overseas models more intelligent. In the first half of 2025, we disclosed to everyone that we had obtained nominations from 2 international automakers in markets outside China. In the second half, we achieved another success and obtained nominations from a third international automaker. The overseas vehicle models we have obtained have reached a total life cycle shipment of 10 million units.

This not only means that our technological capabilities have been recognized by the most demanding customers, it also signifies that we have taken a very critical step in helping overseas traditional automakers with their intelligent transformation.

Thanks to the positive business progress mentioned earlier, we achieved a revenue of CN¥3.76 billion during the reporting period, a year on year increase of 57.7% with a comprehensive gross margin of 64.5%, of which automotive business revenue accounted for nearly 95%, continuing the high gross margin level at 67.2%.

On the profit side, during the reporting period, our adjusted operating loss was CN¥2.37 billion, mainly because we further increased our R&D investment in 2025. I'd like to specifically explain the strategic considerations behind this. We firmly believe that HSD is not only Horizon's core strategic product for winning urban NOA, but also the technological base for future L4 and L5 level autonomous driving.

Its underlying AI foundation model will further become the enabling technological base for industries such as robotics. For this reason, we proactively and firmly increased related R&D investment, especially for cloud service related training expenses. We are not afraid of high R&D investment. On the contrary, we believe that through continuous R&D investment, we can keep improving Horizon's AI foundation models, build a deep moat, and ultimately bring L4 L5 level autonomous driving into thousands of households.

Next, I'd like to specifically introduce Horizon's revenue situation in automotive solutions and IP licensing and services. As mentioned earlier, driven strongly by increases in both volume and price, our automotive products and solutions revenue reached CN¥1.62 billion, which is 2.4 times that of the same period in 2024.

The revenue of this business segment has risen to 43% of the company's total revenue, a significant increase compared to 28% in 2024, showing the optimization of revenue structure after Horizon's advanced intelligent assisted driving solutions entered the large scale mass production cycle. Among all shipments, the proportion of SoC solutions supporting NOA functions is 45%, but they contributed more than 80% of the revenue.

Looking ahead, our future order pipeline is also very strong. In the reporting period, we added over 110 nominated models comprehensively covering mainstream Chinese domestic brands and joint venture automakers. Among them, our cutting edge HSD solution has been nominated for more than 20 models. Also, I want to say that this is only the data snapshot at the end of last year. This year, we will see an explosion in the number of nominations for HSD solutions.

Urban NOA is becoming the core competitiveness of the next generation mainstream models, and we are at the center of this technological transformation.

Regarding IP licensing and services, with world class software algorithm capabilities, outstanding ecosystem influence in China's automotive grade SoC field, a flexible ARM plus Android business model, and highly praised HSD mass production performance, Horizon's software algorithms and tools have already become the trusted foundation technology and model empowering many automakers and ecosystem partners to license and adopt them.

During the reporting period, we recognized revenue of CN¥1.94 billion from licensing and service business, a year on year increase of 17.4%. In addition to our well known joint venture with Volkswagen Cariad, Japan's largest auto parts group has also started licensing cooperation on algorithms and software and became one of our top five customers in 2025.

Looking ahead to 2026, I have many things to share with everyone. First, looking at the automotive layout, last year I still remember our outlook for revenue growth over the next few years was around 50%. Standing at today's point, because we see a strong product cycle ahead, a full pipeline for nominations, and the leading next generation hardware and software technologies, we are confident of continuing the growth momentum of volume and price increases this year and driving the average revenue growth rate over the next few years to 60%, which means a steeper growth curve than our previous expectations.

In terms of SoC solution shipments, the industry's first end to end urban NOA solution based on a single Journey 6M SoC has been officially delivered and installed in vehicles. This system, through extremely deep coordination between software and hardware, greatly reduces system complexity and overall cost. This year, we will work with many ecosystem partners to jointly promote the urban NOA function from the CN¥150,000 mainstream market further down into the CN¥100,000 national car market, bringing additional incremental volumes to the Journey 6M SoC solutions.

As for the hardware and software combination of Journey 6P plus HSD, this year we will release HSD 2.0, which will become the industry's technical benchmark, achieving absolutely leading experience and realizing ten times growth in MPI takeover mileage. More importantly, we will turn leading technical capability into leading market share.

This year, we will launch China's first full vehicle AI agentic SoC for cockpit driving fusion, that is agentic car SoC, and a full vehicle AI agentic OS, namely agentic car OS at the right time. This agentic SoC will provide abundant ARM CPU computing power, Horizon BPU AI computing power, and GPU rendering computing power, as well as rich and powerful multimedia functions.

It is the industry's only all in one intelligent agentic SoC that can efficiently run agent software like OpenAI and HSD urban intelligent assisted driving NOA at the same time. The agentic OS will help cars become super intelligent and caring travel assistants. We can see that ARM CPU computing power plus OpenAI is becoming the new AIPC standard, and Horizon's agentic car SoC and agentic car OS are becoming the new standard for intelligent cars.

At present, in the cockpit domain controller SoC field in the Chinese market, Qualcomm firmly holds a monopoly position with over 70% market share. We have no intention to engage in close combat in the mature cockpit SoC market. In fact, many Chinese companies in the past have tried to directly shake Qualcomm's market position, but none succeeded based on the results.

On the contrary, Horizon's consistent strategy has never been about simple homogeneous competition, but about elevating the dimension of problem solving for customers through underlying technological innovation and system level optimization. Therefore, we are fully betting on the cockpit driving fusion intelligent agentic solution. By breaking the underlying SoC architecture of intelligent assisted driving and cockpit, and connecting the upper level software ecosystem, we will create a blue ocean market and provide end customers with ultimate intelligent experiences.

Here is a vivid example. Today is Thursday, after everyone listens to our earnings call, you get into the car and say to the car system "I want to go watch a recent new movie." The automotive intelligent system will sequentially recommend movie lists, lock in your desired film, decide on suitable cinema locations, automatically complete analysis of the number of people going to watch the movie, select seats, book tickets, and plan driving routes, actively avoid congested roads, adjust the driving mode to after work relaxation mode, safely and timely take you to the cinema, complete automatic parking and payments, and might even order your favorite popcorn.

This revolutionary framework experience connects a simple "after work I want to watch a recent new movie" with a complete and coherent set of scenario based services including efficient commuting, extended needs, and relaxing entertainment.

Beyond upgrading to provide customer value, the cockpit driving fusion intelligent agentic solution is also highly cost competitive. In the process of evolving from 2 separate domain controllers to 1 central computing architecture, we will significantly reduce system level software and hardware costs for our automaker customers, including wiring harnesses, cooling, PCB boards, domain controller shells and other peripheral hardware, especially greatly saving memory chips whose costs have soared this year, while simultaneously improving the product strength and price competitiveness of vehicle models.

For more value highlights and more fun applications, please follow our annual product launch event in April.

Looking back over the past few years, Horizon has launched major new products every year, from Journey 5 SoC to Journey 6 family SoCs, to the HSD solution for urban intelligent assisted driving, and now to this year's cockpit driving fusion intelligent agentic solution, continuously enhancing our product competitiveness. In this continuous upgrading process, we have practiced Horizon's strategic thinking of competing where there's no competition, enabling us to remain in a leading position that is difficult to replace in what seems to be an intensively competitive intelligent EV market.

Looking ahead to our 2027 product plan, based on the newly released next generation computing architecture Riemann architecture, we are continuously promoting the R&D of the Journey 7 series SoCs, whose performance indicators will be aligned with Tesla's next generation SoCs. The progress is smooth and we expect to officially release them next year.

Finally, I would like to report to everyone on Horizon's progress in the L4 frontier, namely the robotaxi related business that everyone cares about. We plan to cooperate with ecosystem partners in the second half of this year, leveraging the foundation model behind Horizon's HSD solution to assist ecosystem partners in conducting pilot robotaxi operations in specific cities in China.

We hope that through this collaborative cooperation model, we can accumulate L4 level experience on real open city roads. Although this year will only be pilot operations, it will prepare us for the next stage of technical evolution and commercialization of the foundation model, ultimately making us the mass production technology provider behind various L4 level passenger cars and robotaxi operators.

Thank you everyone. Now I would like to invite our CFO Wang Lei to briefly summarize the more detailed financial numbers for the full year.

Lei Wang

Thank you, Doctor Yu. I will now present our key financial results for the year ended December 31st, 2025. To be mindful of the time of this call, I will focus on the main items in my discussion today. I encourage participants to refer to our earnings release, which was posted on our investor relations website for additional details.

Our total revenue for 2025 amounted to RMB 3.8 billion, representing an increase of 57.7% compared to the same period last year. This growth was primarily driven by our automotive solutions, which achieved robust growth of 53.9%. Our revenue from automotive solutions is comprised of revenue from product solutions plus licensing and services.

Revenue generated from product solutions surged by 144% to RMB 1.6 billion. This exceptional growth stemmed from 38% higher delivery volumes and 75% higher dollar content per vehicle due to a higher mix of advanced SoCs that support NOA features. Revenue generated from licensing and services increased by 17.4% year on year to RMB 1.9 billion.

Beyond our partnership with Cariad, we have successfully expanded our client base. As Doctor Yu just introduced, a couple of major industry players including Japan's largest OEM and auto parts group have commenced licensing our algorithms and software.

In the reporting period, our overall gross margin was 64.5% for the full year, maintaining a very strong position compared with other automotive solution providers. Moving on to expenses, our R&D expenses reached RMB 5.2 billion, representing a 63.3% increase year on year. We have strategically increased our R&D investment to support development of Horizon Super Drive, primarily on cloud related services. We have also incurred tape out costs for mass production of the Journey 6 series SoCs.

We firmly believe that Horizon has built a solid foundation for automotive intelligence through high performance SoCs and end to end foundation models. Our strong commitment to R&D will continue to deepen our technology moat, laying the groundwork for empowering industries beyond automotive in the future.

Administrative expenses increased by 13.9% year on year to RMB 726 million. Sales and marketing expenses increased by 54.2% to RMB 632 million. Both SG&A expenses grew at slower rates than our revenue growth, reflecting the enhanced efficiency of our daily operations.

This resulted in our operating loss of RMB 3.3 billion. We define adjusted operating loss as operating loss for the period adjusted by adding back share based compensation, which are non cash in nature, and a small portion of capital raising expenses which mainly related to our top up placement that happened last year. Our adjusted operating loss was RMB 2.4 billion as compared to RMB 1.5 billion in 2024.

Last but not least, moving to the bottom line, accounting for equity method investment losses and losses from fair value changes of preferred shares, we reported a net loss of RMB 10.5 billion versus a net profit of RMB 2.3 billion for the same period last year. Please note, this was primarily driven by increase in the fair value of Cariad convertible bonds, reflecting our rising market value. I would call this a good problem to have.

This concludes management's presentation. I will now turn the call over to the operator to facilitate the Q&A session.

Operator

Thank you, management. Now we will start the Q&A session. If you want to raise a question, please press * followed by 1. If you want to cancel the question, please press * followed by 2. Please feel free to raise any questions. To participate in our written Q&A, just type your question into the Ask a Question text area, then click the submit button.

For the benefit of all participants on today's call, please limit yourself to two to three questions, and if you have additional questions, you can reenter the queue. If you are a Mandarin speaker, please ask your question in Chinese first, then follow with English translation.

Our first question today comes from Yukun Ding from HSBC. Please proceed with your question.

Yukun Ding

Hello management, I'm Yukun from HSBC. Congratulations on the strong results. I have two questions. First is on strong revenue growth. The company guided 60% revenue growth going forward. Could you help us break down the drivers behind it and how this momentum could sustain to support this high CAGR in coming years? And also, it would be great to break down chips and software pricing and volume trends.

The second question is on our crown jewel HSD. What's the shipment target for 2026? Thank you.

Doctor Kai Yu

Thank you, Mr. Ding. This is a big question. I'd like to answer it from several aspects. First, the company's automotive revenue will grow more steeply. The automotive segment revenue growth was 54% in 2025. In 2026, as we just introduced, we are confident it will further accelerate to around 60% growth.

The reasons behind this can be shared again from the perspective of shipment volume and ASP. Let me start with sales volume. 2025 is the first year of democratization of intelligent assisted driving, but the pace of advancement varies slightly. Most automakers only caught up with the trend of intelligent upgrades in the second half or even the fourth quarter, starting to equip a small number of models with highway NOA and above intelligent assisted driving functions.

In this environment, last year we shipped 4 million intelligent assisted driving SoC solutions, nearly half of which were advanced J6M and J6P models. In 2025, we also secured over 100 new car nominations, most of which are mid to high end models that will start mass production in 2026, forming a strong pipeline for the coming years.

This year, based on last year's 4 million units, sales volume will further grow strongly with a growth rate of around 35%. More importantly is the structure. Last year, 45% of our shipment volume was AD SoCs and 55% was ADAS SoCs. This year, the proportion of AD SoCs in shipments will further increase to over 55%.

We just mentioned that overall SoC shipments will grow strongly with the AD proportion still increasing, which means that if we look only at AD related product shipments, the growth rate will be remarkable. The unit price of AD products is far higher than ADAS products. The high growth rate of AD SoCs will likewise bring us positive effects on pricing.

If we take our SoC related revenue from last year and divide it by last year's 4 million units of shipments, everyone can actually calculate that although our ASP grew by more than 75% over the past year, it has only reached a level of less than $60, which is still a relatively low absolute amount.

At this ASP level, there is still 50% room for improvement compared with the current average selling price of our highway NOA products and 10 times room for improvement compared with the current ASP of our urban NOA products. The increase in product unit price will contribute more to our revenue growth in the next few years than shipment volume.

So we expect sales volume to have dozens of percent growth in 2026. ASP will have several dozen percent growth. Our SoC related revenue will definitely exceed double digit growth this year. In the automotive industry, it is actually very rare for core business revenue to exceed double digit growth every year.

Regarding IP licensing related business, we expect to remain stable this year. As for HSD's performance this year, as of December 31st last year, we had already secured over 20 HSD production nominations, which include China's highest volume automakers. We are currently actively cooperating with these top tier Chinese automakers on HSD and we can see very good progress.

The number of HSD solution nominations is expected to reach new highs. This year we estimate HSD shipments will reach about 400,000 units.

Operator

Our next question is from Ming Chang from Citigroup. Please proceed with your question.

Ming Chang

Thank you management for giving me the opportunity. I'm Jeff from Citi. I have two questions. The first question is could you share the outlook on gross margin performance over the next several years? Do you expect it to remain at healthy levels such as about 60% plus?

My second question is about memory price hikes. We recognize that automakers are increasingly focused on cost and price competitiveness, and we've seen significant DRAM price inflation this year. So could you walk us through the key levers Horizon is using to maintain its gross profit margin profile amid these headwinds? Thank you.

Doctor Kai Yu

Thank you, Jeff, for your question. Regarding gross margin, based on the average revenue growth rate of 60% in the next few years, we are confident in maintaining a high gross margin level above 60%. In our view, there are mainly 2 peers we face: NVIDIA and Huawei. Both can be referred to as the Hermès of the technology field, belonging to the world's top tier players.

Therefore, the track we are in is naturally not a low margin competitive field. From Horizon's mainstream business model perspective, it is SoC plus software licensing, where SoCs have a gross margin of 40 to 50% and software licensing is close to 100%. The entire business model is a high margin business model.

In the first year of HSD mass production, in order to achieve the fastest mass production and vehicle installation and reduce unnecessary multi-party collaboration and communication, our first few vehicle models were indeed delivered with Horizon's domain controllers. But this time will not be long. This year we will hand over domain controller hardware to Tier 1 partners, so Horizon will continue to focus on SoCs plus software algorithms.

Regarding this year's memory price impact, it is true that the impact on automakers is very significant. As one of the largest intelligent assisted driving solution shipment companies in the industry, Horizon has already locked in memory supply prices in 2025. Therefore, this year's memory price fluctuations will not affect our gross margin.

I also want to emphasize again that Horizon's basic business model is SoCs plus software. The products we deliver actually do not include other hardware, so memory price fluctuations do not directly affect Horizon's product delivery.

Just as we introduced earlier, Horizon has always liked to respond to industry competition through dimensional elevation and innovation. When our peers are competing over SoCs, we provide more powerful high computing power SoCs and more powerful software HSD. When the overall industry profit is under pressure, we launch cockpit driving fusion intelligent agentic SoCs.

Through one SoC all in one solution to complete autonomous driving and intelligent cockpit computing, this underlying technological innovation achieves system level cost efficiency optimization. A very important point is solving cost problems for customers through dimensional elevation.

Specifically regarding the memory price increase issue you mentioned, our cockpit driving fusion full vehicle intelligent agentic SoC launched this year simplifies two sets of memory into one set of memory. Just on memory alone, it is expected to save our automaker customers thousands of yuan in costs. If we further consider wiring harnesses, cooling, PCB area, hardware boxes, etc., it will further help our customers reduce costs.

So it will actually further highlight Horizon's overall market competitiveness. But our competitiveness is not highlighted through price wars. On the contrary, it is through technological innovation. Whether it is Horizon's gross margin level or our revenue, they will actually further grow as a result.

So overall outlook, considering the environment we face and comprehensively considering the various factors mentioned above, we are very confident that we can maintain high gross margin levels above 60% in the future.

Operator

Our next question comes from Tien Hsiao from Morgan Stanley. Your line is now open. Please go ahead.

Tien Hsiao

Hello management, thank you for taking my questions. I'm Tien from Morgan Stanley. I have two questions. The first question is about the integrated solution, because over the past few months we do notice that automakers are accelerating integrating smart driving and smart cockpit functions to further achieve AI agentic functionality.

Horizon Robotics apparently had a head start in launching the agentic car SoC product. So I just want to know if management can share a bit more about the strategy regarding this product, how to truly differentiate Horizon from global peers like Qualcomm or NVIDIA to gain market share?

My second question is about Horizon Super Drive, because this is one of the key drivers into 2026 from the perspectives of commercialization and technical progress. How fast could HSD empower automakers in China to smoothly upgrade their AD capability from current level 2 plus to level 4 or robotaxi like functions? What are the key milestones or indicators investors should focus on? That's my second question. Thank you.

Doctor Kai Yu

Thank you, Tien, for your questions. Judging from future trends, we believe that cockpit driving fusion or integration is the inevitable and natural result of technological evolution, capable of providing a more seamless user experience. Under the wave of rapid development of AI agents, as you know that a lot of people are playing with ChatGPT right now, we also hope to rely on greater computing power and richer hardware types to turn cars into AI agents that surpass mobile phones.

In this process, we can indeed enter the cockpit field through this higher dimension of integration. Horizon will enjoy higher ASP and higher gross margins, and car manufacturers will get more cost effective system costs and brand new intelligent selling points. Consumers will gain an in-vehicle AI assistant with hundreds of teraflops of computing power, rich in-vehicle and external cameras and imaging devices, high quality displays and interactions, as well as more diverse gameplay and customer situations. Please look forward to our April product launch event.

In addition, as I just mentioned the industry cost pressure, I believe that cockpit driving integration will become an emerging trend this year. As for competitors in the industry after cockpit driving fusion, I actually think there are fewer of them. Horizon and NVIDIA's current core advantage areas are intelligent driving.

Horizon is cutting into the cockpit field from intelligent driving, while we haven't seen NVIDIA have a clear strategy to cut into the cockpit field. Qualcomm, which dominates the cockpit field, is now also advancing from the cockpit field into the intelligent driving field.

The competitive difference with Qualcomm lies in software ecosystem. Everyone can see that Qualcomm's market share in the domestic AD autonomous driving market is very low. Some cars equipped with Qualcomm intelligent driving SoCs actually haven't launched any influential intelligent driving systems. This is not just a hardware issue, it's also a software ecosystem issue.

Compared with NVIDIA, it's more interesting. Actually, NVIDIA has been trying for some time with an overseas luxury automaker's cockpit integration project, but it has never reached mass production. To a large extent, we believe NVIDIA currently doesn't have a very good cockpit integration solution or third-party solution partners. Relying solely on NVIDIA to do a complete set of solutions is not their main battlefield.

Horizon can connect everything from SoCs to intelligent driving software to agentic OS based on our own integrated software and hardware solutions. This is an area where Qualcomm's capabilities cannot reach. Transitioning from intelligent driving as a core base to cockpit, we generally feel it is more of a dimensional reduction strike compared to transitioning from cockpit as a core base to intelligent driving.

After all, competitive intelligent driving software ecosystems are scarce now, while cockpit Android ecosystems are actually widespread and within reach.

Regarding the second question, previously we disclosed on one occasion that this year's MPI target is to increase tenfold, roughly from tens of kilometers last year to hundreds of kilometers this year and to thousands of kilometers in 2027. If MPI in certain regions can reach the level of over 10,000 kilometers per takeover, combined with remote human intervention for getting out of difficulties, it would mean the capability to operate L4 robotaxis in those regions has been achieved.

As far as I know, Tesla's current FSD in the United States has already achieved a comprehensive MPI of over 1,000 kilometers and has begun testing robotaxis in some areas, which means FSD's MPI in certain regions may have reached over 10,000 kilometers with almost zero safety takeovers. Horizon's HSD will follow a similar path, within two to three years as a technology provider supporting ecosystem partners to achieve commercial operation of robotaxis in certain regions in China.

To advance from L2 plus plus to robotaxi, besides observing MPI, we also care about the proportion of intelligent driving mileage. During the Chinese New Year, 41% proportion of intelligent driving mileage made us feel that consumers truly like using HSD intelligent assisted driving solutions.

If this proportion of intelligent driving mileage breaks through 50%, it would mean that machine driving mileage exceeds human driving. We judge that users will be willing to use HSD via subscription, and Horizon will usher in restructuring of the business model.

Operator

As we are reaching the end of our conference call now, I'd like to turn the call back over to the company for closing remarks.

Zhuan Jiang

Due to time constraints, we will now conclude today's call. Thank you again for joining us today. If you have further questions, please feel free to contact our IR team. Thank you and bye.

Operator

That concludes today's call. You may disconnect your line.

Details at HORIZONROBOT-W IR

Tips: The content presented above were generated by AI language model with publicly available information and auto-generated subtitles from third-party. The above material does not represent the position of Futu and shall not constitute any investment advice. Futu makes no express or implied warranties or representations regarding the accuracy, timeliness, or completeness of the information shown in the above content.

129K Views