Record-high revenue and profits exceeding expectations! Can Xiaomi stage a rebound?
Xiaomi Group Full-Year 2025 Earnings Live Broadcast (Simultaneous Interpretation)
[AI Key Points Summary]
Financial Performance
- Total revenue reaches 457.3 billion yuan in 2025, breaking through the 400-billion-yuan mark for the first time, up 25% year-on-year, setting a new historical high
- Adjusted net profit hits 39.2 billion yuan, up 44% year-on-year, setting a new historical high
- Comprehensive gross margin at 22.3%, up 1.3 percentage points year-on-year, setting a new historical high
- Revenue from innovative businesses such as smart electric vehicles reaches 106.1 billion yuan, growing over 200% year-on-year, with annual operating profit turning positive for the first time at 900 million yuan
Business Progress
- Global smartphone shipments reached 165 million units, capturing a 13.3% market share, ranking among the top three globally for 22 consecutive quarters
- Xiaomi Auto delivered over 410,000 units in 2025, surpassing its initial target of 300,000 units, with cumulative deliveries exceeding 600,000 units
- Released Xiaomi MiMo Large Model V2 series, with over 1 trillion parameters, placing it among the world's leading open-source models
- Humanoid robots entered automotive factory internships, operating autonomously for three consecutive hours with a 90% success rate
Guidance for next quarter’s performance
- R&D investment is projected to exceed 40 billion yuan by 2026, with 16 billion yuan allocated to AI and humanoid intelligence sectors
- Xiaomi Auto aims to deliver 550,000 units by 2026
- AI investments will exceed 60 billion yuan over the next three years
- Cumulative R&D spending will surpass 200 billion yuan over the next five years
Opportunity
- AI technology has been deeply integrated into Xiaomi’s full ecosystem, including humans, vehicles, and homes, with the MiMo large model ranked among the world's leading open-source models
- IoT business shows significant growth potential in overseas markets, with opportunities six times greater than domestic markets; overseas expansion of new retail accelerates
Humanoid robots will enter Xiaomi factories in large quantities over the next five years
The high-end push of IoT business continues, with significant room for ASP growth
Risk
Memory prices have risen sharply, and the cycle may extend through 2027, significantly impacting the costs of products like smartphones
The Middle East conflict has an indirect impact on the global economy and industrial chain, leading to rising costs of bulk materials such as plastic particles
The start of 2026 faces external challenges including rising memory prices and subsidy reductions, putting short-term performance under pressure
[AI Conference Record]
Operator
Hello everyone, and welcome to Xiaomi Group's 2025 full-year earnings release investor conference call and online audio livestream. This meeting will be recorded, and if you have any issues, you may disconnect now. If you wish to ask a question during the Q&A session, please press star one-one to cancel your question press star one-one. Now let’s invite General Manager of Investor Relations and Capital Markets, Xu Ran, to host this conference call.
Xu Ran
Good evening, everyone, and welcome to Xiaomi Group's 2025 full-year earnings release investor conference call and online audio livestream. Before we begin, we would like to remind everyone that this presentation may contain forward-looking statements, which are subject to various risks and uncertainties and may not be realized in the future due to multiple factors. Additionally, overall market condition information is sourced from channels outside Xiaomi Company.
This presentation includes certain unaudited non-IFRS financial metrics which should be considered supplementary and cannot replace financial metrics prepared by the company according to IFRS. Present at today’s meeting are Mr. Lu Weibing, Xiaomi Group Partner and President, and Mr. Lin Shiwei, Xiaomi Group Vice President and Chief Financial Officer. To start, Mr. Lu will share recent strategic updates and business conditions, followed by Mr. Allen, who will review the company’s 2025 full-year financial performance. Finally, we will move into the Q&A session. I will now hand over the meeting to Mr. Lu.
Lu Weibing
Good evening everyone, thank you for joining our full-year 2025 earnings call. Tonight, I will share with you three key areas: first, a review and summary of the main achievements we made in 2025; second, updates on breakthroughs in our core technology fields, particularly in AI and mega-intelligence; and third, an outlook on our strategic development direction and operational focus for 2026.
First, let’s look at some of the accomplishments we achieved in 2025. In 2025, Xiaomi Group continued its rapid growth, with both annual revenue and net profit reaching new historical highs. The group's total revenue reached 457.3 billion yuan, surpassing the 400 billion yuan mark for the first time, representing a 25% year-on-year increase. Adjusted net profit reached 39.2 billion yuan, growing by 44% year-on-year.
Breaking it down by business segment, first, let’s examine our smartphone business. According to the data, in 2025, our global smartphone shipments ranked among the top three, with a market share of 13.3%, maintaining a steady position in the top three globally for 22 consecutive quarters. Our smartphone shipments ranked second in Latin America and Southeast Asia, and third in Europe and Africa. Additionally, according to third-party data, our smartphone activation volume in mainland China rose to second place in 2025.
In terms of premiumization, the proportion of our high-end smartphone sales in mainland China within our overall smartphone sales reached 27.1% in 2025, increasing by 3.8 percentage points year-on-year. In the price range of 6,000 to 10,000 yuan, our market share in smartphone sales increased by 2.3 percentage points year-on-year. Domestically, we have solidified our high-end foundation while continuing to break through the ultra-high-end market priced between 6,000 and 8,000 yuan.
In February 2026, we launched our first flagship smartphone priced at 1,999 euros for the global market, marking a milestone in Xiaomi's overseas premium strategy. We will continue to compete at the forefront in mature markets worldwide, driving overseas high-end smartphone sales to new heights.
The second segment is IoT business. Last year, our IoT business experienced rapid growth, with annual revenue exceeding 120 billion yuan for the first time, a year-on-year increase of 18.3%, reaching 123.2 billion yuan. Both overseas and domestic revenues hit record highs. Among them, large home appliances achieved a year-on-year revenue growth of 23%, with shipments setting new historical records. Wearable products ranked first globally, TWS earphones ranked second globally, and tablet shipments grew by 25.2% year-on-year, ranking fifth globally.
In June last year, we launched our AI glasses, which ranked third globally and first in China in terms of shipment volumes. In terms of premiumization, we have continued to push forward the premiumization of all product categories domestically and internationally, with overseas premium products also performing exceptionally well.
The third segment is Xiaomi Auto. In 2025, Xiaomi Auto delivered over 410,000 vehicles for the entire year, far exceeding the initial target of 300,000 units. By February 13, 2026, cumulative vehicle deliveries exceeded 600,000 units. In 2026, we will fully strive to achieve the delivery target of 550,000 units.
On March 19 this year, we officially launched the new generation of Xiaomi SU7. The new SU7 has undergone significant changes from the inside out. We have comprehensively upgraded the three-electric system, chassis system, and electronic architecture, and enhanced various aspects, such as safety, where we have taken another major step forward in safety redundancy. Within just 34 minutes of the new SU7's release, pre-orders exceeded 15,000 units, and within three days, pre-orders surpassed 30,000 units.
Additionally, at MWC, we showcased our satellite machine concept car. It is not only a concept car but also represents Xiaomi Auto's latest exploration into innovative design based on hardcore technology. We were the first invited Chinese automotive brand to participate, and we believe that this concept car represents the design and innovation of China’s auto industry, capable of competing alongside the world's leading brands to a certain extent.
Lu Weibing
The second important aspect is our breakthroughs and investments in the field of hardcore technology. In 2025, Xiaomi Group's R&D investment exceeded 33 billion yuan, and by 2026, we expect it to surpass 40 billion yuan. We anticipate investing 16 billion yuan in areas such as large-scale AI models. These R&D investments have built a solid product foundation and competitive barriers for us, as our breakthroughs in hardcore technology have established this groundwork.
First, let’s talk about AI. It’s fair to say that in 2025, we’ve made groundbreaking progress in AI. First, AI is undergoing a historic leap from being functional to being highly usable, representing a fundamental shift in application paradigms. It is moving from handling single tasks to managing complex ones, from passive consumption to active planning, and from public attributes to ecosystem attributes. For a company like Xiaomi with a rich variety of end-user products such as smartphones, cars, home appliances, and IoT, the value of AI far exceeds that of companies focused on a single product category.
Secondly, Xiaomi’s MiMo large model has entered the top tier of global open-source models. In March this year, we released three models at once: Xiaomi MiMo V2 Pro, Xiaomi MiMo V2 Omni, and V2 77S. These three models form Xiaomi’s complete technical foundation in the AI era. Moreover, the total parameters of V2 Pro exceed 1 trillion, supporting a context length of up to one million tokens.
In the Artificial Analyst large model intelligence index global rankings, we ranked eighth overall. By brand ranking, we are fifth globally. During both the anonymous testing period and after official launch, our models achieved first place on the weekly leaderboards multiple times on the Open Router platform, with daily usage reaching more than double that of the second-place model. On the benchmark test list, our average task completion rate ranks third globally. We will continue to iterate and upgrade the MiMo models, advancing towards true general-purpose intelligence.
Thirdly, Xiaomi is poised to be the first to push AI into the physical world. The deep integration of AI with the physical world represents the next phase of smart technology. Xiaomi has an ecosystem entry point with hardware ranging from tens of thousands to hundreds of millions of units. In terms of personal devices, in March this year, we launched the AI agent Xiaomi Cloud on our smartphones, initiating a small-scale closed beta test. We were also the first terminal manufacturer to attempt deploying Lobster on mobile phones, exploring ways to provide users with a true AI smartphone and ecosystem experience.
In automobiles, the new generation Xiaomi SU7 is equipped with the Xiaomi XLA cognitive large model, achieving improvements in three major capabilities: understanding and reasoning in complex scenarios such as mall-level parking and voice-control assistance. Additionally, the Xiaomi SU7 features the Super Xiao AI smart cockpit and advanced AI vehicle control with smart home integration experiences.
In the realm of smart homes, we released a whole-house solution in October last year, giving full-home intelligence the ability to see, think, and act. This marks the first real-world application exploration of Xiaomi MiMo, outlining a new paradigm for the next generation of smart homes.
Fourth, AI has become the core driving force behind technological innovation within Xiaomi. Among the winning projects of the 2025 annual technology awards at the end of last year, over two-thirds utilized AI technology, reimagining existing processes using AI across materials and structures, AI chips, OS, autonomous driving, and tech evaluation fields.
Therefore, we believe that backed by China’s strong AI industry foundation, the next decade of AI will see Chinese consumers having access to top-tier large models globally, along with abundant long-term data and scale advantages across diverse ecosystems. Thus, in the next three years, we will continue to invest heavily in AI, committing over 60 billion yuan.
Lu Weibing
Finally, let's discuss giant-body intelligence. In 2025, we will continue to explore the boundaries of physical intelligence and achieve phased results in the field of giant-body intelligence. The giant-body intelligent robot is the ultimate integration platform for AI capabilities, chip capabilities, OS capabilities, and manufacturing capabilities, so the threshold is very high. Additionally, we have achieved phased results.
In February this year, we first released a tactile-driven fine-grasping micro-model, which is related to dexterous hand technology for humanoid robots. Following that, we open-sourced the Xiaomi Robotics VLA model, Xiaomi Robotics-0, successfully breaking multiple records. In March 2026, Xiaomi's giant robot entered an automotive factory for an internship and operated autonomously for three consecutive hours at the self-tapping nut workstation. The bilateral simultaneous installation success rate was 90%, while also meeting the production line’s fastest cycle time requirement of 76 seconds.
Therefore, I believe the current progress marks just the beginning. Over the next five years, we believe a large number of giant robots will enter Xiaomi factories to perform tasks.
In 2025, Xiaomi reached the significant milestone of its 15th anniversary, also marking the start of a new five-year journey. However, we face many challenges such as rising memory prices and the reduction of subsidies, among other external factors. Thus, the beginning doesn’t seem very easy, and there may be some pressure on short-term performance.
However, Xiaomi's goal of becoming a global leader in hardcore technology remains unchanged. We will continue to increase investment in technological innovation, with a focus on breakthroughs in areas like AI, chips, operating systems, and military intelligence. At the same time, we will expand Xiaomi's 'human, vehicle, home' full ecosystem model globally, promoting the globalization of Chinese technology. Moreover, we will certainly seize the historic opportunities brought by AI. That concludes my sharing today, and now I'll hand it over to Allen.
Lin Shiwei
Friends, good evening. As Lu just shared with everyone, 2025 is a year of historic leap for us. In 2025, our total revenue reached 457.3 billion yuan, a new high for a single fiscal year, representing a 25% year-over-year growth. In the fourth quarter of 2025, our revenue also hit a new quarterly record, reaching 116.9 billion yuan.
This year, our overall gross margin was 22.3%, an increase of 1.3 percentage points year-over-year, also a new historical high. Although the second half of 2025 was more challenging than the first half, for the full year, our divisional revenue from smartphones multiplied by AIoT reached 351.2 billion yuan, up 5.4% year-over-year, another annual record. The divisional gross margin for smartphones multiplied by AIoT reached 21.7%, up 0.5 percentage points year-over-year.
In terms of smartphones, this year's revenue was 186.4 billion yuan, accounting for 40.8% of total revenue. In 2025, our global smartphone shipments totaled 165 million units. Our high-end strategy has also shown clear results, with continuous improvement in product capabilities. According to third-party data, in 2025, our share of high-end smartphone sales in mainland China reached 27.1% of total sales, up 3.8 percentage points year-over-year.
In the price range of RMB 4,000 to RMB 6,000 in mainland China, our smartphone sales market share reached 17.3%, representing a year-over-year increase of 0.5%. In the RMB 6,000 to RMB 10,000 price range, our smartphone sales market share has reached 4.5%, with the market share nearly doubling year over year.
In 2025, our global smartphone shipment volume ranked among the top three, with a market share of 13.3%, maintaining a stable position in the top three globally for five consecutive years. Our smartphone shipments ranked among the top three in 58 countries and regions, and among the top five in 70 countries and regions worldwide. In 2025, despite the rise in memory prices in the second half of the year, our mobile phone gross margin remained relatively healthy, with an annual gross margin of 10.9%.
Lin Shih-wei
In terms of IoT, this year we performed exceptionally well both in revenue and gross margin. In 2025, our IoT revenue grew rapidly by 18.3% year-over-year, reaching RMB 123.2 billion, setting a new historical high. Both our domestic and overseas revenue hit record highs. Benefiting from product mix optimization and enhanced product capabilities, our IoT gross margin reached 23.1% this year, setting a new historical high, increasing by 2.8 percentage points year-over-year.
By category, our performance in smart home appliances was particularly outstanding this year, with revenue from smart home appliances growing by 23.1% year-over-year, hitting a new historical high. Our wearable devices also maintained rapid growth and continued to hold a leading position in the industry. Our wearable devices ranked first globally in terms of shipment volume, our TWS earphones ranked second globally, and our tablets also maintained rapid growth. In 2025, our shipments increased by 25.2% year-over-year, ranking fifth globally.
In terms of internet services, we continue to expand our user base. In December 2025, our global MAU reached 750 million, representing a year-over-year growth of 7.4%, with MAUs in mainland China reaching 190 million, up 10.1% year-over-year.
In 2025, our internet services business revenue reached RMB 37.4 billion, setting a new historical high with a year-over-year growth of 9.7%. In the fourth quarter alone, our internet services revenue reached RMB 9.9 billion. The gross margin of our internet business remained relatively robust throughout 2025, with a full-year gross margin of 76.5%. Advertising will continue to drive the growth of our internet business, with revenue this year reaching RMB 28.5 billion, setting a new historical high. Overseas internet services revenue grew by 15.2% year-over-year, accounting for 33.8% of total internet services revenue.
Let’s talk about our smart electric vehicles. Revenue from our smart electric vehicles and AI-related innovative businesses reached RMB 106.1 billion in 2025, breaking through the RMB 100 billion mark in less than two years, with a year-over-year growth of over 200%. This revenue accounted for 23.2% of the group’s total revenue, with RMB 103.3 billion coming from smart electric vehicle sales and RMB 2.8 billion from other related businesses.
The gross margin of our smart electric vehicles and AI-related innovative business division reached 24.3% this year, improving by 5.8 percentage points year-over-year. In 2025, our smart electric vehicles and AI-related innovative business division achieved positive annual operating income for the first time, with operating income reaching RMB 900 million.
In 2025, we delivered a total of 411,082 new vehicles, with 145,115 new vehicles delivered in the fourth quarter of 2025 alone, setting a new quarterly high. Our average post-tax selling price this year was RMB 251,171, representing a 7% year-over-year increase.
Lin Shiwei
Our cumulative R&D expenditure over the past five years has reached 105.5 billion yuan. In 2025, our R&D spending reached 33.1 billion yuan, increasing by 37.8 percentage points year-on-year. We estimate that starting from 2026, our accumulated R&D expenditure over the next five years will exceed 200 billion yuan.
In terms of net profit, the group’s adjusted net profit in 2025 was 39.2 billion yuan, a new historical high, representing a year-on-year increase of 43.8%. Our capital expenditure in 2025 reached 18.2 billion yuan, growing by 73% year-on-year, with smart electric vehicles and AI-related innovative businesses accounting for more than 66%.
We also continuously strive to enhance shareholder value and repurchase shares on the open market. In 2025, our share repurchase amounted to 6.3 billion Hong Kong dollars, and since early 2026, our share repurchase has reached 4.7 billion Hong Kong dollars. In January 2026, we also announced an automatic share repurchase plan of up to 2.5 billion Hong Kong dollars, demonstrating the company's confidence in long-term future development.
We actively practice the concept of sustainable development. In 2025, Xiaomi Group purchased more than 40 million kilowatt-hours of green electricity, with transaction volumes increasing more than tenfold year-on-year. In 2025, the photovoltaic power consumption at Xiaomi’s automobile factory exceeded 13 million kilowatt-hours, reducing carbon emissions by nearly 10,000 tons annually.
In terms of ESG ratings, Xiaomi Group achieved a management-level B rating in both the 2025 CDP Climate Change Questionnaire and Water Security Questionnaire. Additionally, in March 2026, we received the EcoVadis gold medal for the third consecutive year, achieving a record-high score of 81 points. Our ESG efforts continue to gain recognition from third parties.
In 2026, we will continue to adhere to a prudent yet aggressive operational strategy. We look forward to achieving another remarkable performance this year. Thank you all. That concludes my sharing, and we can now move on to the Q&A session.
Xu Ran
Thank you, Mr. Ai Lun. We will now proceed to the Q&A session. To allow more investors the opportunity to ask questions, please limit your inquiries to two per person. Thank you. The Q&A session begins now. As a reminder, if you wish to ask a question, press star one-one; to cancel, press star one-one.
Operator
The first online question comes from Andy at Morgan Stanley.
Andy
Mr. Lu and Mr. Allen, good day to both of you. First of all, congratulations on the company achieving record-high revenue and profits for the 2025 fiscal year. I have two questions. The first question is about storage components, particularly memory prices. The sharp increase in memory prices has been a major risk factor that investors have been most concerned about regarding the smartphone business since the fourth quarter of last year. Recently, we’ve noticed that competitors have started raising the selling prices of some smartphones to pass on cost pressures, while Xiaomi's smartphone prices remain stable.
Does this comparison indicate that Xiaomi has a clear advantage over competitors in terms of inventory and supply chain management for memory, allowing us to perform better than competitors amid industry challenges this year? Could Mr. Lu share your latest thoughts on memory cost trends and Xiaomi’s strategy to address this issue this year? That’s my first question. Thank you.
Lu Weibing
Alright, because memory has been a topic of interest starting from the past three quarters, I think everyone will continue to pay attention to it every quarter. I've shared some views on this matter on multiple occasions. There are a couple of key points: First, I believe this is a long-term cycle, which will likely last until 2027. Secondly, the extent of price increases could be substantial. Based on these two assessments, we've taken some measures in the past.
But now, looking back, I personally feel that the cycle may last even longer than our current projections. Due to the surge in server demand driven by AI, we are seeing that the rise in memory costs is actually much higher than our previous forecasts. My earlier predictions were already considered quite aggressive within the industry, but in reality, things are even more intense than I expected.
This is the reality we face—a long-term cycle with a steep rise in prices—and it is having a significant impact on all consumer electronics, not just smartphones. While smartphones receive more attention, there are other product categories where memory requirements were originally smaller, and now those products are facing shortages or even production halts. This is the actual situation of the industry at present.
To assess its impact, we have a straightforward method: calculating the proportion of memory BOM (Bill of Materials) relative to the total device BOM. If the BOM cost of memory represents a high percentage of the overall device BOM, it indicates that the category is more significantly affected. Conversely, if the proportion is lower, the impact will be smaller.
Among our existing product categories, such as smartphones, tablets, and laptops, the proportion of memory BOM relative to the total device BOM is relatively high. Of course, from another perspective, high-end devices, including premium smartphones, tend to have a lower proportion of memory costs, whereas low-end smartphones are relatively more affected by memory costs.
So, if a company has a high proportion of low-end smartphones in its portfolio, it will certainly be more impacted. On the other hand, if the company focuses more on high-end smartphones, the impact will be relatively smaller. This is a simple way to evaluate the impact on each company.
Lu Weibing
Indeed, over the past two weeks, I believe that due to the increase in memory costs, there has been pressure on mobile phone manufacturers to raise the end prices of their devices. I fully understand these actions. I think for any mobile phone manufacturer, unless it reaches an unbearable point, no one would be willing to raise the prices of existing products. However, I believe that the timing or inevitability of these price increases is unavoidable.
Today, I think for Xiaomi, our pressure is also extremely high. But I also believe that we should try to bear more for consumers and hold out a bit longer. If we can't bear it anymore, we will definitely raise prices. When that time comes, I hope consumers will have some understanding of our situation—just that we held on longer than others. But I believe this trend of price increases is inevitable.
To address this issue, I think Xiaomi has several relative advantages. First, I believe in the diversity of our business operations. For example, home appliances are less affected by memory costs compared to mobile phones. As for cars, while memory capacity requirements are higher and losses could be significant, considering the overall Bill of Materials (BOM) for an automobile, I think the proportion is still smaller compared to mobile phones.
I think the diversification of Xiaomi’s product categories, with various product lines developing simultaneously, allows us to better tackle this problem. Second, I believe Xiaomi's demand for memory is quite substantial. We have smartphones, tablets, laptops, TVs, and automobiles. Overall, I think we are among the top few companies globally in terms of memory demand.
Over the past few years, I believe we have established strong trust-based relationships with global memory suppliers, along with long-term supply agreements that demonstrate our commitment. Therefore, for Xiaomi, I don’t think there is much risk of shortages currently, and relatively speaking, we might even have some pricing advantages. That’s my second point.
Third, because I was somewhat pessimistic about the rise in memory costs earlier, we stocked up more aggressively. So I think our inventory levels are relatively sufficient. However, overall, I think the impact on the cost of our terminal devices, especially mobile phones, is still very significant. Therefore, I think this short-term pressure is undeniable.
Andy
Thank you, Mr. Lu, for the clear explanation. My second question is about our new energy vehicles. We noticed that the new generation Xiaomi SU7 was successfully launched last week on the 19th. After its release, during our communication with investors, some noted that Xiaomi did not disclose the pre-order figures but only the locked order data. Some interpreted this as a sign that pre-orders did not meet expectations, which is a relatively negative interpretation.
On the other hand, some investors viewed this as Xiaomi adopting a more rational and composed sales strategy, suggesting that the steady-state sales volume of the new Xiaomi SU7 would be better. This is a more positive interpretation. Could management share their thoughts on these two investor perspectives? Thank you.
Lin Shiwei
Alright, Andy, let me answer this. Following the launch of Xiaomi SU7 last week, we observed a few phenomena from the user side. First, within the first three days, more than 50,000 users took test drives. We also released two figures: 15,000 locked orders within 34 minutes, and after three days of sales, over 30,000 locked orders. We fulfilled our promise by starting deliveries from the fourth day onward, as we had pre-produced a significant number of vehicles to ensure we resolved issues faced by users of the previous generation—such as long waiting times for delivery. So, I believe these measures reflect an iteration based on our experience with the first generation.
You just asked why we only release the number of locked orders and not the number of large deposits. I think for us, locked orders are a fairer way to look at this matter rather than considering order reservations, small deposits, or large deposits. Since locked orders mean that users have definitely decided to take delivery of the car, after locking the order, we also enter our production cycle. Therefore, we believe that locked orders may be a more reasonable figure to assess our current actual situation. So, we made a change from last time, and I think in the industry, everyone might have different practices. Anyway, we will stick to this approach. I think that's the first point.
Then, let’s take this opportunity to report to everyone about some situations related to users based on locked orders. Because we also noticed some differences compared to the previous period, perhaps it's better. First, many investors have asked whether the new locked order users come from our previous generation SU7 or are new users. In fact, we can see that the majority are new users because most users with locked orders may not change their cars within two years. So, what we are seeing is continuous expansion beyond existing circles. I think that's the first point.
Secondly, the percentage of female users is still higher than the previous generation. We observed that iPhone users were around 50% for the first generation, but now nearly 60% of the current generation are iPhone users. Additionally, the progress of our locked orders is faster than the data from the first generation. The two data points I just mentioned show that compared to the previous generation, both the speed and overall progress surpass the earlier figures.
We also see that everyone is very supportive of our optional configurations. Nearly 60% of people chose a paid color option. So, looking at several factors—whether it's the penetration rate compared to the previous generation, the speed of locking orders, the proportion of female users, the share of iPhone users, or the selection of various colors—we find that the penetration rate is higher than the previous generation.
Andy
Alright, thank you very much, Mr. Alan, your insights were very helpful. Thank you.
Xu Ran
Thank you for your question. The next question comes from Goldman Sachs. Please go ahead with your question.
Goldman Sachs Analyst
Alright, Mr. Lu and Mr. Alan, thank you for taking my questions. I have two questions here, both related to AI. The first one is that we have seen Xiaomi complete the foundational model and multimodal capability layout over the past year or two. I would like to ask about the significance of our multimodal AI capabilities for the overall ecosystem. Regarding the recently launched MiCloud, how does MiCloud differ from tools developed by other companies, especially internet platforms developing similar tools like Xiaolongxia? How do we view the competition between MiCloud and internet platforms? And what special significance does MiCloud have for the future development of AIOS? That’s the first question.
The second question is slightly more numerical. I’d like to ask how we consider AI’s service model or commercialization plans for Xiaomi internally, for Xiaomi’s ecosystem users, as well as third-party users or developers. Regarding the application of large models, what KPIs are considered for the team? Additionally, I remember Mr. Lei previously mentioned a plan for 60 billion yuan in expenditures over the next three years. How should this 60 billion be distributed between our OPEX and CAPEX? Thank you.
Lu Weibing
OK, actually, I think over the past two to three years, we've spent a lot of time thinking about how to position our AI strategy for Xiaomi's future. Ultimately, we believe it should be approached in several layers or stages. The first layer is certainly the infrastructure of AI, which includes computing power, such as servers and so on. But after the infrastructure is built, the focus will naturally shift to the explosion of applications, which must serve real-world use cases.
Therefore, last year, we predicted that 2026 would be the breakout year for AI applications. In the end, we still believe that AI will transition from the virtual world to the physical world, which was our original big-picture judgment. Based on this overarching view, we began constructing our relevant AI systems. Starting in 2024, we invested heavily in large models, and by 2025, people will see our progress in the MiMo large model across various stages, including our language models, voice models, and multi-modal models.
Last year, we became very clear that this year, 2024, would certainly be the year of an AI application boom. Thus, last year, we discussed at length what the future AI agents could achieve, particularly in personal smart devices. I think they can play a significant role in accomplishing tasks that were previously impossible. After OpenAI emerged, it gave us tremendous inspiration, allowing us to quickly deploy this technology on mobile phones—Xiaomi was the first company globally, and likely the first Chinese company, to deploy such advanced capabilities on smartphones.
So far, feedback has been extremely positive. However, we believe that a much larger market lies ahead—AI transitioning from the virtual world to the physical world. In my view, autonomous driving and humanoid robots will be the two most important sectors, and this is where we're focusing our efforts. Xiaomi has made strategic moves in these areas.
Ultimately, I believe AI must serve Xiaomi’s broader ecosystem, spanning people, vehicles, and homes. This is our fundamental approach. Guided by this vision, we have made corresponding arrangements, and you’re seeing some of the results being delivered now.
Regarding your question about the differences between MiCloud and other vendors' agents, I believe the key advantage of MiCloud lies in its system-level permissions. Since we are a smartphone company, once we integrate our own agent into the system, it gains high-level permissions. This requires not only strong model capabilities but also sufficient permissions, robust data handling, and seamless integration of all these aspects.
For Xiaomi, we have the models, system-level permissions, access to data, and users. Additionally, we handle the integration ourselves, which I believe is our biggest advantage. Third parties would find it very difficult to secure such comprehensive system permissions and data. Moreover, we manage cloud-side and device-side data separately based on specific needs. At the same time, I believe our security measures are relatively stronger.
Therefore, we believe that MiCloud represents the prototype of a future AI operating system, which is our major judgment. However, I think it’s still very early days. Currently, we don’t have detailed commercialization plans or even specific KPIs for the team because KPIs typically come into play when the business matures. Given the rapid changes happening now, it’s premature to talk about setting KPIs. Regarding the six-billion-dollar AI OPEX and CAPEX, let me invite Allen to address that.
Lin Shiwei
Well, the issue here now is that if you look at what we've said about this year's RMB 16.1 billion and the future three-year RMB 60 billion scope, it actually includes R&D investment as well as our CAPEX investment. Of course, within the R&D investment, there's also the amortization of previous CAPEX. You can imagine that this R&D expense includes our current R&D costs, plus some amortization of previous CAPEX, and our CAPEX for the current year, which can be divided into three parts.
If you look at the RMB 16.1 billion in 2026, you could say that the majority is still R&D expenses. The current R&D expenses may account for over 70% of this RMB 16.1 billion, with CAPEX and previous amortization making up the rest. Over the next three years, as each year's CAPEX will also have some amortization of previous CAPEX carried over to the future, the proportion of R&D expenses might be relatively lower than 70% in the next two years, while other expenses and amortization may take a larger share. I'm not sure if this answers your question.
Goldman Sachs analyst
Very clear, thank you, Mr. Allen, thank you, Mr. Lu.
Xu Ran
Thank you for your question, please hold on. The next question is from Wen Hanjing of CICC, please ask your question.
Wen Hanjing
Alright, good evening, Mr. Lu and Mr. Allen. I have two questions to ask. The first question is about our IoT business because we see that in 2025, the overall performance of the IoT business has been excellent, whether in terms of revenue or the number of IoT device connections. I would like to ask about two aspects that everyone is concerned about this year. For instance, one positive aspect is that our smart home appliance revenue continues to reach new highs year-over-year. On the other hand, some investors are concerned about the impact of the domestic subsidy reduction on our domestic IoT market this year?
On the other hand, considering our new retail strategy, I would like to ask about the progress and planning of our IoT going overseas. That's my first question. The second question focuses on the automotive business. As we see that by 2025, the company's overall automotive business has become profitable. Considering the recent launch of the new SU7 model and subsequent plans, what are our targets for profitability and profit for the automotive business this year? Thank you.
Lu Weibing
First, I will talk about IoT, and then Alan will answer the second question. Regarding the IoT business, because we have a particularly wide range of product categories, I will discuss China and overseas markets separately. In the Chinese market, I think a very important opportunity for us is the premiumization of the IoT business. Overall, although our IoT business has been large in scale, I believe our ASP is still relatively low. Last year, I think we made significant progress, but I feel there is still quite a gap from our target.
For example, products like our smartwatches and hair dryers—I am just giving these examples—you will find that the average selling price (ASP) of products such as rice cookers is still far from the industry leaders. However, with the investments we have made over the past few years in R&D and manufacturing, I think you can see that this year, we should make considerable progress in high-end products. This is what I see as our biggest opportunity.
Of course, regarding large home appliances, I believe our current market share is still relatively low. For instance, our refrigerators and washing machines hold around four to five percent of the market, while air conditioners perform slightly better at about ten percent. From both a scale and ASP perspective, I think we still have a lot of room to grow. Additionally, with the development of our new retail strategy, it provides strong support to our entire IoT business. By the end of last year, we had nearly 18,000 stores, which significantly boosts our IoT business growth.
As for overseas markets, I believe there is enormous potential for IoT expansion. Excluding North America, the overall overseas market capacity could be three times that of the domestic market. Currently, our IoT business overseas accounts for roughly one-sixth of our domestic operations. Even if we cannot achieve the same level of growth abroad as we have domestically, there is still substantial room for improvement.
Therefore, I think overseas, especially in terms of market coverage, we are still in a phase where we can tap into many untapped markets. Relatively speaking, by sending teams, establishing channels, and following up with suitable products, we can leverage the groundwork we've laid in overseas localization and certification over the past few years. I think these efforts are mostly complete, and the potential is immense.
Additionally, similar to the domestic market, I think there is still much room for growth in our premium or higher ASP products. Alongside our new retail presence overseas, last year we had 450 new retail stores, and by the end of this year, we will exceed 1,000 stores, primarily focusing on mid-to-high-end products.
When I visited the Xiaomi Home store in London recently, I was quite impressed. The store mainly features high-end products. In categories like cleaning appliances, we still have a lot of gaps to fill. People say that once they become aware of these products, they are expected to sell extremely well—even items like rice cookers would perform exceptionally well in the UK. Therefore, I think, whether domestically or overseas, IoT still has tremendous growth potential.
Lin Shih-wei
Han Jinghao, let me address the automotive question. Last year, in 2025, we delivered over 410,000 vehicles, surpassing our initial target of 300,000 units. Hence, at the beginning of this year, Lei Jun also set a delivery goal of 550,000 vehicles for this year. Therefore, we expect growth in our automotive business in 2026 compared to 2025. Although there might be some pressure on the broader market this year, we remain confident in meeting our expectations and achieving new milestones.
Secondly, regarding the profit target you mentioned, I would like to take this opportunity to clarify that this segment is called 'Automotive and AI and Other New Businesses.' Therefore, it not only includes automotive operations but also covers our investments in AI and other emerging fields. Thus, one should not evaluate this segment based solely on automotive performance—it encompasses various new initiatives as well.
However, at this current stage, some new businesses are still in the investment phase. As we mentioned earlier, in terms of AI, we will indeed increase our investments this year, including in our robotics division. Therefore, I think the overall situation should be viewed in two parts: one is our automotive sector, and the other is our investment in AI and new businesses. So, while this segment achieved excellent results last year, I believe that this year's outcome will be a result of automotive growth combined with our investments in other areas.
Wen Hanjing
Alright, very clear. Thank you, President Lu, and thank you, President Allen. I have no further questions.
Xu Ran
Alright, thank you. We appreciate your question. Please hold on for a moment. The next question comes from Kina Wang of Citi. Please go ahead with your question.
Kina Wang
Management team, can you hear me clearly?
Xu Ran
Yes, it is possible.
Kina Wang
Alright, thank you. I have two questions. The first one is to understand the recent situation in the Middle East. I wonder if there is any impact on overseas business, including sales of smartphones and AIoT, or if there are any materials or logistics costs that might see an increase due to this situation?
The second question is about the overall trend in gross margin across the company's three main businesses. Specifically, regarding mobile phones this year, is there a guiding principle for maintaining a certain level of profitability or making price adjustments? Is there a principle determining this year's gross margin for mobile phones?
For automobiles, I remember it was mentioned before that this year the sector will face some pressure. But I wonder if we have any principles to outperform our competitors or ensure a certain profit margin. As for AIoT, we've heard that there is room for improvement in the future. However, given that high-end positioning is a strategy this year, will this support an improvement in the gross margin of AIoT? Thank you, management.
Lu Weibing
OK, the first issue is the Middle East. I think this conflict is something we all don't want to see. I hope this conflict can be resolved soon with a solution because its impact on the global economy and industries is definitely significant.
Regarding Xiaomi’s business, the revenue from the Middle East is not very high as a proportion of our total company revenue, only in the single digits. Therefore, even if there are impacts in the Middle Eastern market, relatively speaking, its impact on our overall business is small. So I think it is quite manageable, very manageable.
However, we have indeed seen that due to changes related to oil in the Middle East, there has been an increase in the costs of some bulk materials like plastic particles. These are indirect impacts, but overall, I think it is still controllable, still controllable. That's probably your first question.
As for your second question about gross margins and pricing principles for products like mobile phones, cars, and IoT, I think visibility into memory pricing is still not very clear. In other words, based on past high increases, our forecast for the next quarter also predicts high growth. But in reality, prices have risen even more, making precise quantification difficult now because even a small increase translates into a large absolute number when the base price is already high.
This makes setting product prices challenging, as it’s harder to anticipate far in advance compared to before. However, I still believe that for mobile phones, market share and scale remain extremely important. So if you ask me about our principle, I would say that we aim to make trade-offs while trying to maintain our market share. This is a fundamental consideration — preserving our relative position in the market remains a key priority.
As for mobile phones, since there are few listed companies in this field, there aren’t many reference points. It is difficult for us to obtain accurate data from competitors for comparison. But after years of competition, I think we have a basic understanding of where we stand.
In terms of automobiles, although the absolute value of memory used might be high, its proportion in the overall Bill of Materials (BOM) for vehicles is relatively low. Hence, while it is affected, I don’t think the impact is as significant as it is for mobile phones. When it comes to IoT, the impact is even smaller because these devices use very minimal amounts of memory.
Of course, I think what's good about us is that in many aspects of memory, our overall procurement cycle was quite long in advance. Therefore, I think this part everyone can basically assume won't be too much affected. Overall, I think the impact will mostly be on mobile phones, tablets, and laptops. As for cars, I think there will be an impact, but not as significant. Regarding IoT, I believe the impact will be very minimal. This is basically our overall view.
Kina Wang
Okay, thank you.
Xu Ran
Thank you for your question, please hold on. Next, we have a question from Yin Xingchi of CICC, please go ahead.
Yin Xingchi
Okay, thank you, Mr. Lu, and thank you, Mr. Alan, for giving me the chance to ask a question. Today I have two questions. The first one is about the AI business. We've seen the launch of MiCloud, and I was fortunate enough to participate in the internal testing. I think the performance is excellent. Could the company provide more guidance on the AI front? Specifically, under what preconditions might we consider monetizing related initiatives, whether it's large-scale models or the subsequent commercialization of MiCloud, and start reflecting that in revenue? That's my first question.
The second question is about the company’s progress in smart driving. In particular, could you update us on any new developments in smart driving capabilities and self-developed chips this year? Those are my two questions. Thank you, leaders.
Lu Weibing
I think I've already talked quite a bit about MiCloud just now. As an internal tester, I believe you must have felt that this product has given us quite a few surprises. However, since the product has only been out for a short time, I think there are still many areas that need improvement. I'm using it myself, and I've provided a lot of feedback. For instance, the user interface isn't very friendly yet, and there are many features I want that it currently can't achieve.
But I believe the iteration speed is very fast. You could say it's almost a new version now. I think people can see many improvements, and I have very high expectations for it.
I still think that for Xiaomi, overall, our AI will ultimately serve our entire ecosystem of people, vehicles, and homes. Within this ecosystem, I believe there are many user experiences, interactions, and services. Therefore, purely in terms of AI commercialization, it still seems a bit premature at this stage.
Although our large model has shown fairly high efficiency, and we have also engaged in some token-based commercialization efforts, in absolute terms, the scale is still relatively small. So, discussing commercialization at this point feels somewhat premature for us.
Regarding autonomous driving, I believe everyone has already seen how our XLA model has enhanced our autonomous driving capabilities. In this product launch, we dedicated a portion to introduce XLA. The full OTA update should be available around May, but based on our internal testing results so far, the performance has been very, very impressive.
Since I alternate between driving the two SU7 models, I consider myself a very heavy user of assisted driving with an extremely high usage rate. Every time I drive, I encounter issues that I believe need addressing, and I report them to our team. There are many such cases, and I can genuinely feel that our autonomous driving system is making steady progress.
Moreover, you can see that our autonomous driving solution, our model, and even the chips you mentioned, are all part of our strategic layout. The end-to-end integration will undoubtedly provide users with a significantly enhanced experience. So, I look forward to your feedback. If you haven't driven our car yet, I suggest upgrading to our latest version to experience these improvements in autonomous driving.
Yin Xingchi
Okay, thank you.
Xu Ran
Due to time constraints, we will take the last question from Zoe Xu of UBS. Please proceed.
Zoe Xu
Thank you, management, for taking my questions. Many topics have been discussed by others before, so I’ll ask two questions: one regarding investment in new businesses and the other about AI in IoT. First, with recent accelerated iterations of the MiMo large model, management reiterated our commitment to investing in AI. Could you elaborate on whether this impacts the priority of our self-developed chip initiatives? And can we expect new devices featuring the Pengpai chip in the near future?
The second is IoT. As Mr. Lu mentioned earlier, tablets and PCs are also affected by storage costs. In this regard, will we adopt a pricing strategy similar to that of smartphones, focusing on market share and consumer experience? Those are the two small questions. Thank you.
Lu Weibing
This year, we have indeed significantly increased our R&D expenditures. However, I believe that chips represent an extremely long-term strategic capability for us. I've mentioned multiple times that chips are essentially a platform capability. With this platform, we can empower a wide range of products. Especially with the advent of AI and large models, the reliance on chips has become even higher.
Therefore, although we are increasing our investment in the AI direction, we have not reduced our original investment in chips. On the contrary, I believe many chips need to rapidly transition towards AI. Chips are fundamentally an integral part of our broader AI strategy. Hence, I think there's no need to worry about it at all.
As for PCs and tablets, I think we might adopt a strategy similar to that of smartphones. However, I would like to ask you to pay attention to our latest notebook release. We just launched this notebook, and it’s the first Xiaomi notebook released after four years. The demand far exceeds supply, and the shortage is quite severe, surpassing our expectations.
When we released this notebook, we had already taken into account the rise in memory costs. Despite that, the response post-launch has been very positive. This shows that product strength is fundamental. If the product is strong, I believe users will accept a slightly higher price.
Thus, I think in this year's environment, what’s being tested more than ever is each company’s ability to innovate and their technical capabilities. The rise in memory costs is one factor, but I believe we have other ways to offset some of the increase. I believe that through our efforts in 2026, our management team will be able to deliver what we consider a satisfactory result.
Zoe Xu
Okay, thank you.
Xu Ran
Thank you. Today's meeting is now concluded. Thank you all for your time, and I hope you continue to support Xiaomi Group. Goodbye.
More details:XIAOMI-W IR
Note: The above content is generated by an AI language model based on publicly available data and third-party automatic subtitles. The content does not represent any position of Futu, nor does it constitute any investment advice. Futu Group makes no express or implied guarantees or statements regarding the accuracy, timeliness, or completeness of the above content.
Risk Disclaimer: The above content only represents the author's view. It does not represent any position or investment advice of Futu. Futu makes no representation or warranty.Read more
Comments (672)
to post a comment
282
11