Xiaomi announces a new HK$20 billion share buyback program—will the stock price get a boost?
Xiaomi Group Q1 2026 Earnings Live Broadcast (Simultaneous Interpretation)
[AI Key Points Summary]
Financial performance
In the first quarter of 2026, the Group reported total revenue of RMB 99.1 billion and adjusted net profit of RMB 6.1 billion.
Smartphone segment revenue reached RMB 44.3 billion, with global shipments of 33.79 million units and an all-time high ASP of RMB 1,310, up 8.2% year-over-year.
Smartphone gross margin was 10.1%, up approximately 2 percentage points sequentially.
R&D expenses amounted to RMB 8.95 billion, an increase of 33.4% year-over-year.
Business Progress
- Xiaomi delivered 80,856 SU7 series vehicles in Q1; the new-generation SU7 has accumulated over 80,000 firm orders within 48 days of launch
- Launched the Xiaomi MiLM-V2.5 series large models, ranking jointly first among open-source large models on a globally recognized benchmark
- Global MAU reached 750 million, up 3.8% year-over-year; MAU in Mainland China hit a record high of 196 million
- Ranked second globally in TWS earphones and third globally in wearable devices
Guidance for next quarter’s performance
- Full-year vehicle delivery target remains unchanged at 550,000 units, with Q2 deliveries expected to rise significantly quarter-over-quarter
- A larger new vehicle model, developed on a new platform, will be launched in the second half of the year
- AI investment budget may increase from the initial RMB 16.1 billion allocated at the beginning of the year, subject to ROI-based adjustments
- Overseas IoT business is expected to continue delivering double-digit growth
Opportunity
- Overseas IoT business accounts for nearly 40% of total IoT revenue, with market potential twice that of the China market and, logically, four times the room for growth
- 2026 is seen as the pivotal year and starting point for AI smartphones, with user interaction shifting toward OS Agents
- Launching the MiLM Orbit 100-trillion-token initiative, offering free distribution to global AI users
- Advancing premiumization of IoT products in the China market, with new product sales exceeding expectations
Risk
- Storage costs continue to rise sharply; the upward cycle is expected to be prolonged and substantial, impacting the smartphone industry as a whole
- Facing compounded challenges from overlapping cost, demand, and competitive cycles
- Vehicle delivery lead times previously reached ten months, causing some customers to switch to competing brands
- Memory prices have risen approximately fivefold, exerting pressure on product costs
[AI Conference Record]
Operator
Good day everyone, and welcome to Xiaomi Group’s investor conference call and webcast for the first quarter of 2026. This call will be recorded. If you have any issues, you may disconnect now. If you would like to ask a question during the Q&A session, please press *1 to register your request, and press *1 again to cancel. I now hand over the call to Mr. Xu Ran, General Manager of Investor Relations and Capital Markets at Xiaomi Group, who will moderate today’s conference call.
Xu Ran
Good evening everyone, and welcome to Xiaomi Group’s investor conference call and webcast for the first quarter of 2026. Before we begin, we would like to remind participants that this presentation may contain forward-looking statements, which are subject to various risks and uncertainties and may not be realized due to a range of factors. Additionally, information regarding overall market conditions cited during this call is sourced from channels outside Xiaomi Group.
This conference includes certain unaudited non-IFRS financial measures, which are provided as supplementary information and should not be considered as substitutes for financial metrics prepared in accordance with International Financial Reporting Standards (IFRS). Joining us today are Mr. Lu Weibing, Partner and President of Xiaomi Group, and Mr. Lin Shiwei, Vice President and Chief Financial Officer of Xiaomi Group. The meeting will begin with Mr. Lu sharing recent strategic updates and business developments. He will be followed by Mr. Alan Lin, who will review Xiaomi Group’s financial performance for the first quarter of 2026. We will then open the floor for the Q&A session. I now turn the call over to Mr. Lu.
Lu Weibing
Good evening everyone, and thank you for joining our earnings announcement call for the first quarter of 2026. The first quarter of 2026 marks the first full quarter following the launch of Xiaomi’s new five-year strategy. Over the past five years, we have completed the strategic closed loop of our ‘People, Vehicles, and Smart Home’ ecosystem and achieved comprehensive breakthroughs in automotive chips, AI, and major home appliances.
Looking ahead to the next five years, in the short term, we face the triple challenge of overlapping cost, demand, and competitive cycles. In the long term, we are entering a new phase where AI is reshaping the integrated ecosystem of people, vehicles, and homes. Tonight, I will primarily share with you three key points: first, a review of our key performance in the first quarter of 2026; second, responses to issues of widespread concern; and third, our strategic direction and operational priorities for the coming quarters.
In the first quarter of 2026, the Group reported total revenue of RMB 99.1 billion and adjusted net profit of RMB 6.1 billion. By business segment, our smartphone business proactively managed inventory levels for mid-to-low-end products in distribution channels during the quarter, resulting in a decline in shipment volume. However, our average selling price (ASP) reached a record high, and we maintained our position among the top three globally.
According to Omedia data, our market share in the first quarter of 2026 was 11.3%, securing our place among the global top three for 23 consecutive quarters. Specifically, our smartphone shipments ranked second in Latin America with a market share of 17.4%, up 2 percentage points quarter-over-quarter, and ranked third in Europe, Southeast Asia, the Middle East, and Africa, with respective market shares of 17.2%, 19.3%, and 9.2%.
In the first quarter of 2026, Xiaomi ranked among the top three smartphone vendors in 47 countries and regions globally, and within the top five in 65 countries and regions. On gross margin, a sharp short-term increase in memory costs has pushed the entire industry into a new normal. We will not simply pass on cost increases linearly to consumers. Instead, we will reposition user demand and strike a balance between scale and profitability through product portfolio upgrades, software optimization, and leveraging our operational capabilities.
In the first quarter of 2026, we still achieved a smartphone gross margin of 10.1%, demonstrating the resilience driven by our internal capabilities. For IoT business, Q1 2026 revenue reached RMB 24.7 billion, primarily impacted by a high base effect from China’s national subsidy program in the prior year, which led to a year-over-year decline domestically. However, thanks to overseas channel expansion and an increase in international product categories, overseas revenue hit a record high and grew double-digits year-over-year.
Among IoT products, our TWS earbuds ranked second globally, wearables ranked third globally, and tablets ranked fifth globally. The IoT segment delivered a gross margin of 25.2% this quarter, fully reflecting our ability to hedge against volatility in any single business through cross-segment synergy, resulting in a more robust overall profit structure. This enables the IoT business to generate additional profits to offset potential declines in smartphone gross margins.
For automotive business, Xiaomi delivered 80,856 vehicles in the first quarter of 2026, primarily the Xiaomi SU7 series. As of May 6, 2026, the newly launched SU7 had accumulated over 80,000 firm orders within just 48 days of launch, demonstrating that our automotive business has withstood severe challenges and emerged from its most difficult period. As of April 30, 2026, cumulative deliveries of the Xiaomi SU7 over the past ten months totaled 232,000 units.
On May 21, we officially launched the Xiaomi SU7 Ultra, a sports-car-grade SUV designed for long-distance travel, priced at RMB 389,900. The standard version of the Xiaomi SU7 is priced at RMB 233,500. To refine its ultimate handling capabilities, the SU7 Ultra underwent extensive tuning and validation on the Nürburgring Nordschleife, ultimately setting a new SUV lap record there and becoming the fastest SUV in Nürburgring history.
The second topic I’d like to address concerns AI and large-scale robotics—issues of significant interest to many of you. We have emphasized on multiple occasions that the deep integration of AI with the physical world represents the future of intelligent technology. With a hardware ecosystem offering entry points on the scale of billions of devices, Xiaomi has a tremendous opportunity to become a leader in the AI era. If Xiaomi does not develop its own foundation models, our understanding of such models will always remain superficial. Only by mastering core technologies end-to-end can we build products with differentiated competitiveness.
The industry as a whole is still exploring the ultimate form of AI–hardware integration, but only those possessing both core technologies and real-world application scenarios will be able to rapidly advance when the trend becomes clear. Regarding the latest progress on Xiaomi’s large models, in April 2026, we launched the Xiaomi MiLM-V2.5 series, including MiLM-V2.5, V2.5 Pro, V2.5 77S, and V2.5 Sr—marking a comprehensive leap from 'usable' to 'highly usable.'
On the globally recognized comprehensive large model benchmark Artificial Analysis, Xiaomi's MiLM-V2.5 Pro ranks jointly first among open-source large models in overall intelligence index and fifth globally across all large models. Additionally, its Agent index also ranks jointly first among open-source large models. At the end of March, MiLM-V2 Pro set a new record on the OpenRouter platform, achieving top rankings across daily, weekly, and monthly leaderboards.
On March 31, the day MiLM-V2 Pro launched, our average weekly retention rate for API calls reached 35%. On April 3, 2026, we officially launched Xiaomi’s MiLM Token Plan, offering four tiers—Light, Standard, Pro, and Max—designed to help users enhance productivity at reasonable prices. Since the Token Plan’s launch, the Pro and Max tiers have accounted for over 50% of total revenue.
As of May 12, OpenRouter data shows that MiLM ranked first in usage volume among Hermes Agent models, contributing a cumulative 1.45 trillion tokens over the past month. This indicates that the world’s fastest-growing open-source Agent product has chosen Xiaomi’s MiLM as its preferred inference engine for real-world, high-intensity tasks. In appreciation of global developers, Xiaomi has officially launched the MiLM Orbit 100-Trillion-Token Initiative—a token distribution program offering free token entitlements to global AI users, with a target of distributing 100 trillion tokens within 30 days.
As of the morning of May 12, nearly 80 trillion tokens have already been distributed. Going forward, we will rapidly iterate our large models. Now, regarding AI phones: I believe 2026 is the pivotal year—the inaugural year—for AI phones. This milestone will profoundly reshape the future of smartphones. An AI phone isn’t merely about adding AI features to existing smartphones or simply building an app-based Agent; rather, it represents a fundamental shift from app-centric interaction to OS-Agent-centric interaction.
XiaoAI Assistant is our initial step in this direction. Recently, the China Academy of Information and Communications Technology (CAICT) launched evaluations for on-device intelligent assistants, and Xiaomi’s XiaoAI became one of the first domestic smartphone assistants to pass this authoritative assessment. Additionally, XiaoAI’s public beta now extends across multiple devices—including tablets, PCs, Macs, and automotive systems—marking a significant upgrade in cross-device capabilities. In the future, smartphones will evolve from mere operational tools into personal executive assistants and digital avatars.
Accordingly, we will deeply evolve our operating system with Agent at its core, further strengthening the deep integration between our models and upper-layer frameworks. Now, turning to advanced driver assistance: in March this year, we officially unveiled the architecture of Xiaomi’s VLA cognitive large model, enabling Xiaomi’s intelligent driving assistance to move beyond perception and imitation toward genuine understanding and reasoning. In May 2026, building upon VLA, we officially launched and fully open-sourced DriveVLM, our autonomous driving model.
DriveVLM is a cutting-edge spatial-language-world reasoning framework that unifies three major technical approaches—VLA world modeling and advanced spatial reasoning—into a single cohesive architecture, enabling large-model inference that is both fast and accurate. Finally, let’s discuss large-scale robots. Large robots represent the ultimate integration platform for AI capabilities, chip performance, OS sophistication, and manufacturing prowess, with an exceptionally high entry barrier.
On April 27, 2026, we demonstrated new capabilities and officially launched the full post-training pipeline for Xiaomi CyberOne. On the same day, Xiaomi’s robot moved from the factory floor to the Xiaomi Investor Day event, greeting guests with a heart-shaped gesture and receiving widespread acclaim. This progress is just the beginning—robots will unlock entirely new industrial and market opportunities for Xiaomi.
Historically, Xiaomi has navigated two major challenging periods: one from 2014 to 2016, and another from 2022 to 2023. After each difficult cycle, Xiaomi emerged stronger and achieved even more robust growth. Thus, it is precisely during times of short-term uncertainty that a company’s strategic resolve, organizational capability, and execution efficiency are most rigorously tested.
Xiaomi remains firmly committed to becoming a global leader in next-generation hard-core technology, continuously intensifying innovation in AI, chips, operating systems, smart vehicles, and other domains, while simultaneously driving organizational innovation and upgrades to support the long-term, healthy growth of all our businesses. That concludes my sharing for today. Now, I’ll hand over the time to President Alan.
Lin Shiwei
Thank you, Mr. Lu, and good evening to all our investor friends. As Mr. Lu just shared with everyone, we believe 2026 will be a year marked by both short-term challenges and long-term opportunities. Regarding our performance, in the first quarter of 2026, our total revenue was RMB 99.1 billion, and our consolidated gross margin was 22%.
Breaking it down by segment, revenue from our Smartphones × AIoT segment amounted to RMB 79.3 billion, with a segment gross margin of 22.5%, up 2.5 percentage points sequentially. In the smartphone business, persistent and significant increases in memory costs have had a broad impact on the industry. This quarter, we focused on optimizing our sales mix and channel management.
This quarter, our smartphone revenue was RMB 44.3 billion, accounting for 44.7% of the Group’s total revenue. Our global smartphone shipments reached 33.79 million units. Thanks to our strategic adjustments, our smartphone average selling price (ASP) hit a record high this quarter at RMB 1,310, an 8.2% year-over-year increase.
According to third-party data, premium smartphones accounted for 23.5% of our total smartphone sales in mainland China during the first quarter. Per Omedia data, we ranked third globally in smartphone shipments in Q1 2026, with a market share of 11.3%, marking our 23rd consecutive quarter in the global top three.
Despite rising memory prices, by proactively managing shipments of mid-to-low-end smartphones and controlling channel inventory, we maintained a healthy smartphone gross margin of 10.1%. In IoT, our first-quarter revenue reached RMB 24.7 billion, with overseas revenue growing robustly—delivering double-digit year-over-year growth and reaching a new historical high.
We continue to prioritize profitability and refrain from engaging in industry-wide price competition, while simultaneously expanding our overseas channels and broadening our overseas product portfolio. IoT gross margin reached 25.2%, up 5.1 percentage points sequentially. From a product category perspective, this quarter we ranked third globally in wearable device shipments, second in TWS earphone shipments, and maintained steady operations in our tablet business, ranking among the top five globally.
In internet services, we have built a substantial global user base. By March 2026, our global monthly active users (MAUs) reached 750 million, up 3.8% year-over-year. Notably, MAUs in mainland China hit a record high of 196 million, representing an 8.1% year-over-year increase.
In the first quarter of 2026, our internet services revenue was RMB 9.5 billion, up 4.3% year-over-year. The gross margin for internet services this quarter was 76.1%. Advertising continued to drive growth in internet services, with ad revenue reaching RMB 7.1 billion this quarter, an increase of 7.8% year-over-year.
Turning to our smart electric vehicles and other innovative businesses powered by AI: this segment generated RMB 19.9 billion in revenue this quarter, up 6.9% year-over-year, accounting for 20% of the Group’s total revenue. However, due to lower deliveries of the SU7 series this quarter, we delivered 80,856 new vehicles in Q1.
Smart EV sales revenue was RMB 19.1 billion, while other related business revenue totaled RMB 0.9 billion. This quarter, our average post-tax selling price per vehicle was RMB 235,000. Impacted by reductions in purchase tax subsidies and rising costs of certain raw materials, the gross margin for the smart EV and AI innovation business segment was 20.1% this quarter.
In the first quarter of 2026, our Smart Electric Vehicles and AI Innovation segment reported an operating loss of RMB 3.1 billion. As of April 23, 2026, cumulative deliveries of the new-generation SU7 had exceeded 26,000 units. By April 30, 2026, cumulative deliveries of the Xiaomi SU7 over the past ten months surpassed 230,000 units.
Our R&D expenses in the first quarter of 2026 amounted to RMB 8.95 billion, representing a 33.4% year-over-year increase. Our capital expenditures reached RMB 3.27 billion, up 20% year-over-year, with the Smart Electric Vehicles and AI Innovation segment accounting for 45.6% of total capital expenditures.
In terms of net profit, the Group reported an adjusted net profit of RMB 6.1 billion for the first quarter of 2026. We also remain committed to enhancing shareholder value and have been actively repurchasing our shares in the open market. From the beginning of 2026 through today, we have repurchased HK$8.4 billion worth of shares, exceeding the full-year total from last year, reflecting our confidence in the company’s long-term future.
We are also actively advancing our sustainability initiatives. In April 2026, we released Xiaomi Group’s 2025 ESG Report—our eighth consecutive annual ESG report. The report comprehensively outlines our strategies and achievements in 2025 across areas including data privacy, responsible marketing and services, the circular economy, sustainable supply chains, talent development, and corporate governance.
Regarding ESG ratings, in March 2026, Xiaomi Group maintained its 'A' rating. Additionally, we were honored to rank among the top 1% of Chinese companies in the S&P Global CSA assessment. Our ESG efforts have received recognition from third-party institutions. Thank you all—this concludes the content we wanted to share today, and we will now move to the Q&A session.
Xu Ran
Thank you, President Allen. We will now proceed to the Q&A session. To allow more investors the opportunity to ask questions, please limit your inquiries to no more than two per participant. Thank you. The Q&A session is now open. If you would like to ask a question, please press *1. To cancel your request, please also press *1. Our first question comes from Andy Meng of Morgan Stanley.
Andy Meng
Good evening, President Lu and President Allen. First, congratulations on delivering better-than-expected results despite headwinds in the overall smartphone market. I have two questions—I’ll start with the first one. We observed that the AIoT business delivered exceptionally strong gross margins in the first quarter, and following last week’s new product launch, sales of our earclip-style headphones have also been very robust.
Could management elaborate on whether the AIoT business can achieve strong revenue growth this year alongside its high gross margins? Additionally, can we expect notable highlights in the coming quarters regarding new product launches and overseas market expansion? Thank you.
Lu Weibing
The AIoT business is of great interest to everyone. As early as the third quarter of last year, we anticipated a super-cycle of rising memory costs—one that would be both prolonged and substantial in magnitude. Under these circumstances, products with high memory usage—such as smartphones, tablets, and laptops—are significantly impacted, as memory constitutes a relatively large portion of their total cost structure.
We’ve also been thinking that the AIoT business is extremely important. From Xiaomi Group’s perspective, it helps balance or alleviate our overall operational pressure. That’s why in Q3 last year, we elevated the strategic importance of our AIoT business for 2026. After raising its priority, you can actually see that we’ve adopted somewhat different strategies overseas versus in China.
In the Chinese market, our key focus remains on premiumization. This year, you should have gradually noticed that we’ve launched a large number of products—firstly, these products are exceptionally well-designed, highly accepted by users, selling very well, and receiving excellent reviews. At this launch event, we introduced our earclip-style earbuds, which I strongly recommend everyone try out—the fit, design, sound quality, integration with XiaoAI Assistant, and compatibility with iPhones are all outstanding.
Beyond that, you’ll also notice our ‘Powerful Wind’ air conditioners. As temperatures have started rising recently, our Powerful Wind Pro series has been selling extremely well. Users really appreciate how quickly these units can cool a room and how comfortable the airflow feels. You can also see the significant improvements in our refrigerators—particularly in freshness preservation—and our newly launched air purifiers.
So, you can see that our progress in premiumization is moving very, very well. Overseas, our priority remains rapidly scaling up, because our current market share there is still quite small, yet demand is enormous. However, the IoT business faces numerous market access requirements overseas—it’s not as simple as taking a product we sell in China and selling it abroad. There are many regulatory and certification hurdles, so we need to tackle them country by country.
Last year, we largely resolved market access issues in most countries. Therefore, this year you’ll find that we’ve significantly expanded our product portfolio across many markets, and our new retail stores—Mi Home stores—are also expanding rapidly. Coupled with our deep collaborations with key distributors, we’ve essentially achieved double-digit growth with our core distribution partners.
Combining these two efforts, you can see that in Q1, our AIoT business performed quite well—both in terms of revenue growth and gross margin maintenance. I’d like to emphasize that the overseas market still holds tremendous potential. If you look at our position in the Chinese market—specifically in earbuds and smartwatches—you’ll see our share in the premium segment is still relatively low, meaning we have substantial room to grow.
Currently, our overseas business is only about half the size of our China market, but its total addressable market is roughly twice as large as China’s. Logically speaking, that implies at least four times the growth potential for us to capture. Overall, I believe the competitive environment overseas is somewhat more favorable than in China. All of this represents massive growth opportunities—that’s my answer regarding the IoT question.
Lin Shiwei
Andy, let me add a few numbers. While President Lu just outlined our strategic, category, and channel-level expansion of overseas IoT, we’ve also observed that—even though it’s not peak season—overseas revenue has already achieved double-digit year-over-year growth, reaching a historical high and accounting for nearly 40% of our total IoT business.
Andy Meng
Thank you, President Lu and President Allen. My second question is about Xiaomi Auto. Following the launches of the SU7 GT and Standard Edition, based on user demographic analysis of buyers, have you observed any differences compared to the initial SU7 or last year’s SU7? Furthermore, how will these two new models impact your existing vehicle lineup in terms of sales and profitability, and what effect will they have on Xiaomi Auto’s key metrics—such as total sales, revenue, and profit—in 2026? Thank you.
Lu Weibing
Regarding the SU7, as everyone can see, I think this time we launched two models—one high-end and one entry-level. On top of the original Standard Pro and Max versions, we added an Ultra version, and below that, we introduced a new Standard version, renaming the previous Standard version to Long Range. A key consideration behind adding these two versions is that, compared to the Model 3, our SU7 clearly holds a very strong advantage. However, compared to the Model Y—as Lei mentioned during the launch event—I believe there’s still a gap in terms of sales volume.
After analyzing this gap, we found that the standard or entry-level version of the Model Y accounts for a very high proportion—my impression is that it makes up roughly 60% to 70% of its total sales. When we competed using our Long Range version, however, we realized we faced some disadvantages in terms of cost or pricing. Moreover, for urban drivers’ daily commuting needs, a range of around 600 kilometers is generally sufficient; any additional range may unnecessarily increase vehicle weight and other costs.
Based on this line of thinking, we introduced a Standard version, and above the Max, we added an Ultra version. As many have noticed, this Ultra version immediately broke the SUV lap record at the Nürburgring Nordschleife—it essentially represents the performance ceiling for SUVs.
With these two new models, I believe our SU7 product lineup has been significantly enhanced and completed. At the top, we now have the Ultra version. We’ve observed that over half of the customers who purchase the Ultra opt for our full-package configuration priced at RMB 429,900—which includes all available options bundled at a discounted rate—and more than half of Ultra buyers choose this complete package.
During my recent nationwide market visits—just returning from them—I noticed that among SU7 buyers, some originally intended to purchase the SU7 Max but upgraded to the Ultra, while a significant portion came from owners of performance models from BMW, Mercedes-Benz, and even Porsche.
Currently, initial production capacity for the Ultra is limited—approximately 2,000 to just over 2,000 units per month. If you’re interested, I’d recommend placing your order promptly; otherwise, delivery lead times may be longer.
The Standard version effectively addresses the needs of urban commuters requiring roughly 500–600 kilometers of range, while also offering a more budget-friendly price point for those with cost considerations. During my field visits, I asked frontline staff—many of whom previously worked at Tesla before joining Xiaomi—and they told me that when customers test drive and compare these two vehicles, 70% to 80% ultimately choose our Standard version.
In terms of competitiveness, I believe its advantages are very clear across dimensions like intelligent features, comfort, and equipment configuration. Regarding the gross margin you just asked about—the lowest-tier model’s gross margin is certainly lower. As you know, it’s priced RMB 20,000 below the Long Range version, but the actual cost reduction is less than RMB 20,000, so gross margin does decline. However, the Ultra’s gross margin, I believe, will be better than before.
Since sales figures aren’t yet available, it’s difficult to quantify the exact impact on the overall series’ gross margin. But I’m confident that its contribution to total volume—and overall enhancement—is definitely positive.
Andy Meng
Great, that’s very clear. Thank you, Mr. Lu.
Xu Ran
Alright, thank you for your question. The next question comes from Goldman Sachs—please go ahead.
Goldman Sachs Analyst
Understood, Mr. Lu, and good evening, Mr. Allen. Thank you very much for taking my question. I have two questions for management. My first question relates to our AI initiatives. We’re genuinely pleased to see Xiaomi’s AI business delivering progress this year that has exceeded market expectations. Given the current momentum in our AI business and the increasing investment by other major AI players, could you share how Xiaomi is thinking about its AI investment budget of approximately RMB 16 billion for this year?
Since the launch of the MiLM 2.5 model, what additional operational metrics can you share? What will be the update cadence for future versions of the MiLM large models? Specifically regarding XiaoAI Assistant, what are your plans for upcoming updates, including the roadmap or timeline for the next-generation AIOS? That concludes my first question. Thank you.
Lu Weibing
Let me elaborate on our XiaoAI Assistant initiative. Regarding AI-related expenditures, I believe Mr. Allen can provide further details. As for XiaoAI Assistant, we are pushing forward at full speed. We firmly believe XiaoAI represents a transformative shift in user interaction. That’s why you’ve seen us rapidly roll out XiaoAI on smartphones. Currently, we’re working urgently to enable cross-device functionality—including tablets, smart TVs, wearables, and eventually vehicles—as well as smart speakers.
Looking ahead, another critical aspect will be whether we can achieve seamless integration of user data and memory across devices. Once realized, this will significantly enhance user experience—it will profoundly augment human capabilities. However, none of this would be possible without our large models, our underlying framework, and the transformation of HyperOS into AIOS. These elements are absolutely essential.
At present, we’re fully committed and making strong progress. Our entire team is working closely together toward this goal. Therefore, this year’s new OS launch event is being shaped with precisely this direction in mind.
Lin Shiwei
On the topic of AI Tokens, let me highlight a few points. As you’ve seen, we’ve already begun experimenting with Tokens and have introduced various Token packages to our users. The initial response has been very positive. As Mr. Lu mentioned earlier, over 50% of current usage involves our Pro and advanced-tier models, indicating strong user recognition—particularly for our premium models.
Regarding the first question on AI investment: our current view remains aligned with our initial annual guidance of RMB 16.1 billion. That said, we will adjust this amount as our AI business evolves, given the substantial opportunities we continue to identify in AI—especially if model iterations gain strong user acceptance and we increase inference capacity accordingly. In such a case, our budget may rise, but any incremental investment will strictly depend on achieving a clear return on investment (ROI). Thank you.
Goldman Sachs Analyst
Thank you, Mr. Lu and Mr. Allen. My second question is a follow-up regarding the automotive business. From a delivery perspective, we saw over 80,000 units delivered in Q1 and more than 30,000 in April, which appears somewhat behind the annual delivery target of 550,000 units. Could you share your plans for new vehicle models in the coming quarters?
Additionally, regarding the gross margin of our automotive and new businesses, we observed a sequential decline in Q1. Mr. Allen just mentioned that this was partly due to cost pressures, but also partly attributable to a one-time impact from purchase tax subsidies for cross-year orders. Could you help us quantify the specific impact of this purchase tax subsidy? And how should we assess the gross margin trend for automotive and new businesses in Q2 and for the full year? Thank you.
Lu Weibing
Alright, I’ll let Allen address the gross margin question, but first let me clarify something upfront. As you’ve seen, in Q1—specifically January and February—we were only delivering one model, the SU7. With this latest refresh and launch of the new SU7, we essentially paused sales of the older version for nearly two months while preparing for the new product. We did this because we noticed that when some competitors rapidly iterate their products, it often leads to significant backlash from existing users—a sense of being 'backstabbed.'
Therefore, based on that observation, we deliberately halted sales of the old model for roughly two months to ensure users were fully informed about our upcoming new SU7. We wanted to provide maximum transparency. In hindsight, this approach worked well—the user feedback on this refresh has been very positive, and we believe we made our best effort. However, this decision did have one consequence: during January and February, we only delivered the SU7.
As a result, total deliveries in Q1 were around 80,000 units. But by April, deliveries had already started ramping up significantly. Even though the delivery period for the SU7 in April was still relatively short, actual deliveries were notably higher than the 30,000 units we publicly disclosed. Therefore, I’m quite optimistic about Q2.
Moreover, we plan to launch a larger vehicle later this year. This model will be built on a completely new platform, and personally, I believe it’s an extremely innovative and highly competitive product. We have very high expectations for it. Overall, we remain confident in achieving our full-year delivery target of 550,000 units.
Lin Shiwei
Regarding the slight decline in vehicle gross margin in Q1, as mentioned earlier, the first factor was the purchase tax subsidy. Specifically, for orders placed by the end of last year that hadn’t yet been delivered, we provided users with a purchase tax subsidy of approximately RMB 10,000 to RMB 15,000 per vehicle. This subsidy is reflected in our average selling price (ASP) and thus impacted gross margin.
Secondly, we began selling display vehicles, as we had accumulated a certain inventory of such units. This also had some impact on our overall ASP and gross margin. Additionally, broader macro factors—including rising costs for batteries, memory, and chips—have increased our input costs.
However, as we’ve noted before, some of these cost increases are industry-wide. Given that our ASP remains relatively higher than peers, the impact on our gross margin hasn’t been as severe. Hence, despite the sequential decline, our gross margin still remains at a relatively high level within the industry.
Also, as Mr. Lu mentioned, we only delivered just over 80,000 vehicles this quarter, which affected the amortization of our fixed costs. Starting in Q2, however, we’ve begun deliveries of the new-generation SU7, including both the Standard and GT versions. Therefore, we expect a clear sequential increase in volume—that’s certain.
Secondly, deliveries of the GT variant will also boost our sequential ASP. At the same time, we’re actively implementing ongoing cost-reduction initiatives with our suppliers, which will partially offset these pressures. Taken together, these factors will shape our gross margin trajectory over the coming quarters.
Goldman Sachs Analyst
Great, thank you. Thank you, President Lu, and thank you, President Allen.
Xu Ran
Thank you for your question. Our next question comes from Wen Hanjing of CICC. Please go ahead.
Wen Hanjing
Hello, President Lu, good evening President Allen. Thank you for taking my question. I have two questions. My first question focuses on our smartphone business. We’ve observed that amid significant memory cost pressures this year, Xiaomi’s smartphone business demonstrated notable operational resilience in Q1, both in terms of shipment volume and gross margin.
I’d still like to ask—looking forward, and particularly toward 2027—how do you plan to balance key business metrics such as volume, pricing, and gross margin in a market environment marked by both crisis and challenges? Thank you.
Lu Weibing
OK, we’ve actually discussed this issue several times before. Overall, I believe the long-term upward trend in memory costs aligns closely with my original assessment—it’s a prolonged cycle with substantial price increases. Specifically, DDR5 prices began rising in Q2 last year, while LPDDR started increasing in Q3 last year. Because LPDDR started later, its price surge has been even more dramatic.
By Q2 of this year, based on market quotations—or rather, contract prices, since spot prices are currently very volatile—the increase has been roughly fivefold. This represents an enormous challenge to memory costs and, by extension, overall smartphone costs. Moreover, I don’t believe this is the end; the cycle still has a long way to run.
Looking further ahead to 2027 and 2028, my forecast starting around Q3 is that we’ll transition from sharp quarterly increases—which previously reached 60% to 80%, or even 100% month-over-month in some cases—to a phase of more gradual, sustained price growth.
Even with slower growth, the pressure remains immense because the base level is already so high. Even a modest percentage increase translates into a significant absolute cost impact. Under these conditions, I believe the effect on consumer electronics as a whole—not just smartphones—is substantial.
Take televisions, for example: although memory constitutes a smaller portion of their total cost, their memory price increases have actually far outpaced those of smartphones—precisely because the baseline cost is lower, making percentage increases appear larger. Therefore, I believe the overall impact on the consumer electronics sector is profound.
Under these circumstances, I believe we have still taken several actions. In the first quarter, as you can see, we made certain adjustments based on this strategic approach. The first point is that we must dynamically balance the relationship among selling price, cost, volume, and gross margin.
I don't think it's rational to sharply raise prices in the short term, slash prices at any cost, or aggressively undercut competitors with very low pricing to capture market share. Each company has its own strategy and timing for how it assesses and responds to its market.
Therefore, in Q1, you can see that our shipment volume declined somewhat, but our average selling price (ASP) increased by nearly 10%, and our gross margin improved by almost two percentage points compared to Q4—demonstrating, in my view, the effectiveness of this strategy.
Secondly, as you can observe, we continued launching new products this year, such as the Redmi K70 Max and the Redmi Note 13 Max, which we just introduced last week. Given that costs are already at this level, I believe we cannot simply pass all cost increases directly onto consumers.
Simply increasing prices on existing products to shift costs onto consumers would be extremely difficult for them to accept. Instead, we must redefine and redesign our product lineup—understanding what users truly need under this new cost structure—and deliver products that meet those needs at reasonable prices. That, I believe, is what we should be doing.
As you’ve seen, we’re rapidly streamlining our product portfolio. I believe our recent models have been performing quite well, especially considering the current macro environment of rising prices. Overall, I’d say our performance remains very solid.
Looking at the Chinese market over recent months—from the perspective of BCI activation data—our market share has remained stable at above 14%. This, in my view, reflects a critically important strategic achievement.
We’ve also committed to minimizing price increases on legacy products—delaying them as long as possible and keeping them as modest as feasible. As you may have noticed, we were among the last to implement price hikes, only beginning in mid-April, with increases of RMB 200 to 300.
Overall, I believe users have shown us considerable understanding. Therefore, I think we should adhere to this overarching strategy to navigate what could be another two years of rising costs.
Given the challenges ahead, I believe the situation will be quite difficult, and under these conditions, we absolutely must seize the major opportunity presented by AI—particularly the significant potential of AI smartphones.
Wen Hanjing
Great, thank you, Mr. Lu—that was very clear. My second question is about our overseas expansion in the automotive business. Earlier, both Mr. Lu and Mr. Allen shared some updates regarding our progress in this area. From an internationalization perspective, considering our upcoming vehicle launch schedule and the continued increase in the number of our overseas new retail stores, could you please share more details on this? Thank you.
Lu Weibing
Regarding our overseas expansion, from a disclosure standpoint, I’d say there hasn’t been much change since last quarter—around March. We’re largely sticking to what I outlined at the last investor day. First, our overall timeline targets overseas launches in Q3–Q4 of 2027. Our strategy generally follows this sequence: starting with developed markets before moving to developing ones; beginning with mid-to-high-end segments and then expanding into mid-range offerings; and initially focusing on right-hand-drive markets before entering left-hand-drive markets.
Currently, our entire team is working intensively on preparations for this overseas rollout. Automotive globalization is quite complex—it’s a heavily regulated product category, subject to numerous legal and regulatory requirements, extensive localization adaptations, and critically, the need to establish a robust channel network and physical retail presence.
Therefore, I believe our current progress remains on schedule.
Wen Hanjing
Great, thank you, Mr. Lu. I have no further questions—thank you very much.
Xu Ran
Thank you for your question. The next question comes from Yin Xingchi of CITIC Securities. Please go ahead.
Yin Xingchi
Good morning, Mr. Lu and Mr. Allen. I’m Yin Xingchi, an automotive analyst at CITIC Securities. Thank you very much for the opportunity to ask questions. I have two questions for the management team. My first question concerns our MiLM large language model. As mentioned earlier, our daily average token usage has already surpassed one trillion tokens.
I’d like to ask for more specifics: what are the current user conversion rate and retention rate under this Token Plan? Additionally, regarding our AI team’s commercialization roadmap, are there any internal KPIs or performance metrics you could share with us at this stage?
My second question is about the IoT business. Earlier, Mr. Allen mentioned that overseas IoT business saw very rapid growth in both revenue and proportion in Q1. Could management please provide some updated guidance on the expected growth and proportion of our overseas IoT business for the remaining three quarters, especially by year-end? That’s all for my two questions—thank you,各位 leaders.
Lu Weibing
I’ll start by addressing IoT, and then Mr. Allen will speak about the MiLM large model. Your question actually seems quite similar to the first one we received today. We still view the IoT business as a very important balancing factor over the next few years to mitigate the impact of rising memory costs on our smartphone business. Therefore, achieving high-speed growth in IoT remains a key strategic priority for us.
Regarding AIoT, we remain firmly committed to premiumization in the China market—I emphasize ‘firmly committed’ to premiumization. As you saw from this recent product launch, several new products were introduced, and I believe every single one met or exceeded our internal targets. In particular, our ear-clip headphones significantly surpassed our original target by nearly twofold.
So, as you can see, our product strength has improved substantially, and market reception has been very positive. We believe the overseas market still offers enormous potential. Historically, our relatively low share overseas was due to several key reasons: first, insufficient product adaptation for overseas markets; second, some products met functional requirements but failed to satisfy regulatory or market entry criteria; and third, many overseas markets have quite high entry barriers.
Over the past two years, we’ve invested significant effort into addressing these issues—particularly the third one, which is channel development. Starting last year, we increased our investment in distributors, and this year, we’re ramping up investment in Xiaomi Stores overseas.
We believe these two initiatives will greatly support the overall development of our overseas IoT business. As you’ve seen, our overseas IoT business grew rapidly in Q1, and we expect this momentum to be sustained over a relatively long cycle because the addressable market remains very large. I personally believe there is still 3x to 4x room for growth in our overseas IoT business—it all depends on how effectively we execute.
Lin Shiwei
Regarding your first question about MiLM usage volume, let me share a few observations following the launch of our Token Plan. First, the proportion of paying users remains very high—we see that over 30% of users making token calls are paying users, which is a notably high conversion rate.
Second, as I mentioned earlier, users of the Pro and Max versions account for over 50% of total token usage. Third, overseas usage is also quite substantial—overseas users represent roughly 50% or more of total usage. While we haven’t disclosed specific retention rates, I can say that retention remains at a very high level.
However, stepping back, our AI models are not yet at a stage of true business closure. As you can see, whether in China or overseas, many new models are being launched, and we are still in an active iteration cycle.
While users are calling tokens, we are continuously learning and iterating our models. Therefore, going forward, we may need to adjust our Token Plan according to industry developments—that’s the first point. Second, we are still in a growth phase focused on data collection and model iteration; we haven’t yet reached the stage where we can fully scale the Token Plan into a major monetized business. I hope you understand.
Yin Xingchi
Okay, thank you all leaders very much. That’s very clear—I have no further questions.
Xu Ran
Thank you for your question. The next question comes from Kyna Wong of Citi. Please go ahead.
Kyna Wong
Mr. Lu and Mr. Allen, thank you for giving me the opportunity to speak. First, I’d like to follow up on some of the earlier questions regarding large AI models. As you know, many companies in the market working on large models are expected to release updated versions around June or July—essentially iterating every three months or so.
I’d like to ask whether Xiaomi’s MiLM large model will follow this market-driven update cadence or if the company has its own strategic considerations guiding its release timing. Additionally, regarding the earlier discussion about API calls, we’ve recently seen price reductions and telecom operators launching their own large models to monetize computing power. How does Xiaomi position its large model in the market, and what’s the strategic approach here?
Secondly, how does Xiaomi expect its large model to improve overall corporate efficiency? Specifically, can we expect to see optimization reflected in operating expenses (OPEX)? I’d appreciate your insights on these AI-related questions.
My second question concerns the automotive segment. In Q1, the combined operating loss for EVs, AI, and new businesses was approximately RMB 3.1 billion. Given Mr. Lu’s earlier confidence and the full-year target, we should see quarterly deliveries exceed 100,000 units in the coming quarters. Based on this, it seems feasible to offset the Q1 operating loss—though this also depends on Xiaomi’s ongoing investment in AI. Could you share your outlook for the full-year direction of EVs, AI, and new businesses? Thank you.
Lu Weibing
I’ll address the AI part first, and I’ll ask Allen to respond to the second part. We don’t have a rigid internal timeline mandating model updates every three months. We believe new iterations should only be released when truly necessary—not just for the sake of releasing something. Xiaomi will maintain its own rhythm and judgment on this matter.
Regarding Xiaomi’s AI large models, I think it’s important to emphasize that MiLM is fundamentally designed to serve Xiaomi’s core businesses. Deep integration with our business operations is critical. Xiaomi operates across numerous data-rich scenarios, and once this data flywheel gains momentum, it significantly enhances our large models.
For large models, we must consider both model quality and cost efficiency. Without a closed-loop business ecosystem, it’s extremely difficult to properly evaluate these factors. Therefore, Xiaomi’s internal data flywheel provides meaningful value in strengthening and refining our large models.
Beyond this, I believe we will also sell some tokens to players within the industry and participate in market competition, which I think would also benefit Xiaomi. I feel that pursuing these two paths—using both legs to walk—is mutually reinforcing and supportive. This approach, in my view, is the key difference between Xiaomi and other independent third-party large model companies.
Lin Shiwei
Regarding pricing, let me add a few points here. As Mr. Lu mentioned earlier, we are not yet an independent large model company, right? Selling tokens is certainly not our core business, so we won’t release a large model just for the sake of launching one. We’ll wait until we feel it’s truly ready before making the release.
At the same time, internally, we’re actively encouraging our own employees to use our MiLM model more frequently. On one hand, this helps them improve their efficiency; on the other hand, as I mentioned earlier, it also generates more data to help us iterate and enhance our own model, making it stronger. So, this efficiency gain isn’t just about improving our internal productivity—it’s also about strengthening our model’s capabilities. I see these as two complementary aspects.
Regarding trends in EV and AI-related new businesses, I think it’s quite evident—you’ve probably noticed that the biggest difference between Q1 and Q4 or previous quarters is the decline in deliveries. As we explained earlier, this is mainly because we delivered almost no SU7 units in Q1; we were primarily focused on delivering the SU7 and preparing for the launch of the next-generation SU7.
So, this has certainly had some impact on revenue, right? Looking ahead, as we’ve previously indicated, EV deliveries in Q2 are expected to increase sequentially—that’s the first point. Of course, we will also continue to ramp up our investments in AI. Therefore, going forward, our operating profit or loss will ultimately reflect the combined effect of delivery volumes and our AI-related expenditures.
Kyna Wong
Thank you.
Xu Ran
Thank you for your question. Our next question comes from David Joe of JP Morgan. Please go ahead.
David Joe
Good evening, Mr. Lu and Mr. Allen. I’m David from JP Morgan. I have a question for both of you. First, SU7 shipments are projected to peak in the second half of 2025 and then start to decline. Could management explain the underlying reasons for this trend and how the company plans to reignite growth momentum for the SU7? Thank you.
Lu Weibing
OK, regarding SU7 shipment volumes—you mentioned a peak followed by a decline—I believe the primary reason is that our initial order backlog for the SU7 was very large. However, we later realized that the delivery lead time was still quite long; at its peak, the wait time reached ten months. I think this definitely created some friction for consumers during the waiting period.
So some users may have switched to other brands, as we’ve observed that certain competitors have offered users a RMB 5,000 discount on their deposit orders, among other incentives, which led to order conversions. I’ve noticed there is indeed some customer attrition toward these other brands.
Secondly, looking at our product lineup, we recognize there’s a gap. Everyone can see that the benchmark and reference model for this product is clearly the Model Y. After reviewing user needs, we found that in urban environments, a daily driving range of around 600 kilometers is sufficient for most customers—but we currently lack such a product.
We previously had a standard-range version planned, but ultimately didn’t launch it; instead, we released a long-range version with 800 kilometers of range. However, from the user perspective, this range is somewhat excessive—especially in mid-sized cities—where I believe a range of roughly 550 to 600 kilometers would fully meet the need to charge just once per week.
Therefore, I feel there’s a clear gap in our product portfolio. With this latest launch of the SU7, we’re addressing this by offering both a high-end and a low-end variant: on the lower end, we’ve added a standard version, and on the higher end, we’ve introduced the Ultra version to reignite the growth momentum for the SU7.
So far, I still believe that since the launch—from this past weekend through today’s test drives and deposit orders—the response has been very strong. I’m confident these measures will help revitalize the SU7.
David Joe
Understood. Thank you, thank you, Mr. Lu. We don’t have any further questions from our side.
Xu Ran
Thank you for your question. The next question comes from Jimmy Xu of UBS. Please go ahead.
Jimmy Xu
Great, thank you to the management team for the opportunity. My first question is—since you just mentioned the impact of smartphones on sales, and while we may implement some balancing measures going forward—we have indeed observed that gross margin performance in Q1 this year remained very robust, showing clear sequential improvement.
The question I’d like to ask is this: considering that part of the situation may be due to product mix optimization and another part may stem from inventory we built up previously, I’d like to understand—looking solely at smartphones or consumer electronics—what trend should we expect for profitability or gross margins in the coming period, especially given that memory prices continue to rise? That’s my first question. Thank you.
Lu Weibing
OK, I believe that under the current trend of rising costs, this pressure will persist—it’s an objective reality. To address this issue, we need to strike a proper balance among scale, profitability, and market share.
But more importantly, amid this broad cost-increase trend, we must answer a simple question: why would users be willing to pay more for a product than they did before? Therefore, our strategy is not simply to raise prices on existing products and pass them on to consumers. Instead, under the new cost structure, we must redefine our products so that users still feel they’re getting exceptional value—which I believe aligns with Xiaomi’s core values. I think this is the first—and very critical—point.
Secondly, after optimizing our product mix, we must effectively address our average selling price (ASP). The decline in volume we’re seeing is inevitable, but I believe we can largely offset the negative impact of lower volumes on revenue through a meaningful increase in ASP, while also stabilizing our gross margin. In this way, I believe our total gross profit amount will remain manageable.
So, these are, broadly speaking, our fundamental thoughts and strategies regarding our smartphone business.
Jimmy Xu
Thank you very much for your question.
Xu Ran
Due to time constraints, today’s meeting ends here. Thank you all for your time, and we hope you’ll continue to strongly 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 automated subtitles. The above content does not represent any position of Futu and does not constitute any investment advice. The 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 (369)
to post a comment
51
5