XIACY (XIACY) Q2 2026 Earnings Report, Transcript and Summary

Ladies and gentlemen, welcome to Xiaomi's 2026 Interim Results Announcement Investor Conference Call and audio webcast. Today's conference is being recorded. If you object, you may disconnect at this time. [Operator Instructions] Now I'd like to hand the conference over to your host today, Ms. Isabella Gao from Group Investor Relations and Capital Markets Department. Please go ahead.

Good evening, ladies and gentlemen. Welcome to the investor conference call and audio webcast hosted by Xiaomi Corporation regarding the company's 2026 interim results. Before we start the call, we'd like to remind you that this call may include forward-looking statements, which are underlined by a number of risks and uncertainties and may not be realized in the future for various reasons. Information about general market conditions comes from a variety of sources outside of Xiaomi. This presentation also contains some unaudited non-IFRS financial measures that should be considered in addition to, but not as a substitute for the company's financials prepared in accordance with IFRS. Joining us on the conference today are Mr. William Lu, Partner and President of Xiaomi Corporation; and Mr. Alain Lam, Vice President and CFO of Xiaomi Corporation. To start, Mr. Lu will share recent strategic and business updates of the company. Thereafter, Mr. Lam will review the company's financial performance in the first half of 2026. Following that, we'll move on to the Q&A session. I will now turn the call over to Mr. Lu.

Good evening, everyone. Thank you for joining our Q2 2026 results announcement call. In Q2 of 2026, external environment remains challenging with the cost of core components such as memory remaining high, consumer market demand still recovering slowly and industry competition becoming more intense. Tonight, I will take this opportunity to share with you about 3 aspects. First, a review of our key performance results for Q2 2026. Second, we will share our latest advancements in the fields of AI and embodied robots. Third, we will look at our strategic development direction and business focus for the next few quarters. First, in Q2 2026, the group's total revenue was RMB 108.9 billion. Adjusted net profit was RMB 6.2 billion. Looking at different business segments, first, smartphone. In Q2, we proactively optimized our product structure and increased selling prices, resulting in a record high ASP for smartphones. Although smartphone shipment declined, according to Omdia data, we maintained our position among the top 3 globally in Q2 of 2026, marking the 24th consecutive quarter that we have ranked among the top 3 globally in market share. In Q2 2026, we ranked among top 3 in smartphone shipments in 53 countries and regions worldwide and among top 5 in smartphone shipment in 67 countries and regions worldwide. Our smartphone shipment ranked second in Southeast Asia, Latin America and Middle East with market share of 19.3%, 16.2% and 13.5%, respectively. We ranked third in Europe and Africa with market share of 16.5% and 10.9% respectively. For gross profit margin, memory costs remained at historically high levels in Q2. We have consistently achieved a good balance between scale and profit through product mix upgrades, software optimization and improved operational capabilities. In Q2 '26, our smartphone gross margin was 8.5%. Gross margin for the first half was 9.3%. Second, IoT business. In Q2 2026, IoT business achieved revenue of RMB 31.3 billion. Due to the high base effect of national subsidy in Chinese market last year, revenue declined year-on-year. However, thanks to the expansion of overseas channels and the increase in overseas product categories, overseas revenue increased significantly year-on-year this quarter. As of June 30, we have opened more than 640 new retail stores overseas. In second quarter 2026, Xiaomi tablets rose to the fourth place globally. TWS Earbuds ranked second globally, wearable products ranked second globally. This September, we'll also attend the IFA Exhibition in Berlin, Germany to showcase our comprehensive ecosystem of products for Human × Car × Home to the world. Our large home appliances will begin to fully enter the European market. Xiaomi EVs. Xiaomi delivered 104,199 vehicles in Q2 2026, marking the sixth consecutive quarter of year-on-year growth in deliveries. In the first half this year, Xiaomi SU7 series ranked first in sales among pure electric sedans priced above RMB 200,000 in Mainland China. As of 17th August 2026, cumulative delivery volume of Xiaomi SU7 Series exceeded 500,000 units. In July 2026, we held the Xiaomi Automotive Technology Launch Conference, officially releasing our first extended range vehicle architecture, the Xiaomi Kunlun Technical Architecture. Based on this, we launched Xiaomi SkyNomad series of extended range SUVs positioned as intelligent variable space SUVs, including the Xiaomi SkyNomad N90 Max and N70 Max with presale prices of RMB 299,900 and RMB 259,900, respectively. Preorders have been enthusiastic. Xiaomi SkyNomad will be officially launched in September, and we hope to receive everyone's continued support. We continuously refine Xiaomi EV technology using the Nurburgring Nordschleife track. In June 2026, Xiaomi U7 GT with track package set a new Nurburgring Lab record with autonomous driving with a time of 10 minutes and 29.483 seconds. This achievement is the result of the joint effort of Xiaomi EV European R&D center and our Mainland assisted driving team. We hone our smart driving technology on the track to improve driving safety in extreme scenarios for mass-produced vehicles. Second, the issues of AI and embodied intelligence that everyone is concerned about. Latest progress on Xiaomi's foundational large model. In April 2026, we released Xiaomi MiMo-V2.5 series. On 2nd August 2026, CCTV reported on the latest weekly global large-scale model call volume ranking of OpenRouter. With Xiaomi MiMo-V2.5 topping the list, its call volume increased more than sixfold from RMB 1.5 trillion to RMB 10.5 trillion in 2 months. Developers choose MiMo primarily because of its outstanding full model perception, intelligent agent and code capabilities, highly competitive pricing and the fact that the entire series is open source. Ultimately, the value of technology lies in the breadth of its use. We continue to explore how to extend MiMo from model capabilities to the agent ecosystem. In June this year, we officially released an open source MiMo code, an exploratory AI programming assistant. It's not just a useful AI coding tool, but also an AI teammate who lives in your computer and gets better at understanding you the more you use it. In the same month, we released official version of Xiaomi MiLM Cloud and partnered with Kingsoft Office ecosystem to launch a full chain document office efficiency improvement solution. At the same time, we officially released Miloco 2.0, an open source whole house smart AI solution centered on the MiMo Big Data model. This represents a further step in exploring the next generation of smart homes, enabling home devices to go beyond simply following commands and begin to think, thus exploring a new form of smart home in greater depth. Benefiting from the continued growth in call volume, our API calls and token plan began contributing revenue this quarter. We'll continue to iterate on our foundational model, striving to place it among the top tier of models with the same parameter levels. It is gratifying to see that Chinese models have entered the forefront of the world. Next AI mobile phones. We have always emphasized that AI phones must integrate AI capabilities into OS. The Xiaomi HyperOS 4 that we just released is an important start. We have upgraded AI from simple question-and-answer function in individual software to a user-friendly assistant that can remember user habits, understand user thoughts and truly help users get things done. Based on Xiaomi self-developed MiMo model and the AI-driven transformation of the OS has been fully upgraded to Hyper XiaoAi 2.0, bringing a brand-new interactive experience with the Island and Inspiration ball features while also offering 2 modes, quick and expert. The powerful expert mode with its comprehensive integration with operating system applications can seamlessly complete complex tasks across applications and devices. For expert mode, we have launched Hyper XiaoAi 2.0, expert mode points-based membership subscription plan. Each user will receive 1,000 points per month for free. Heavy AI users can flexibly choose their subscription plan with monthly special offers starting at just RMB 19. Latest advance in embodied robots. In June 2026, -- our robotics team won 2 championships at CVPR 2026 and ICRA 2026 WBC. In July this year, we released several more important updates. At Xiaomi's EV factory after quarter of effort, Xiaomi's embodied robots have improved the success rate of dual-side task at the self-tapping threaded insert loading station to 98%. At the same time, our embodied robots have begun to explore new workstations in the logistics area of the final assembly workshop, sorting the center console site cover returnable box folding with success rate of 90%. We have also officially released Xiaomi-Robotics-U0 and Xiaomi-Robotics-1. Robotics-U0 is the first unified generative model in embodied intelligence field that can handle all 4 types of tasks, driving solutions to data challenges faced by the embodied intelligence industry. Robotics-1 was pretrained on 100,000 hours of real-world data and then trained using cross-ontology data, ranking first in multiple simulation evaluations. On August 5, we officially open sourced this model. Finally, I would like to say that in July this year, we were included in the Fortune Global 500 list for the eighth consecutive year, ranking 232nd, an improvement of 65 places from last year, marking a new high since we first appeared on the list in 2019. In 2026, Kantar Brands top 50 global brands in China list, Xiaomi ranked second. Our expanding global brand influence will help us balance our regional impact and help Xiaomi navigate cycle economic cycles. Looking ahead to Q3, several factors that will put pressure on short-term operations remain. Storage costs will remain high, consumer demand recovery will take time and competition will remain fierce. However, as we have mentioned many times before, Xiaomi has always ushered in a new growth cycle after facing each difficulty by improving its capabilities. After proactive adjustments over the past 2 decades, past 2 quarters, we have a clearer understanding of the external environment and our capabilities are steadily improving. Short-term pressure will not change our long-term strategy. We remain committed to our goal of becoming a global leader in next-generation hard-core technologies and continue to increase our investment in hardcore technologies such as AI chips, operating system and embodied intelligence to build strength for the next stage of growth. That concludes my part. Now over to our CFO, Alain.

Thank you, Mr. Lu. In Q2 2026, our total revenue was RMB 108.9 billion. Our gross margin was 19.8% -- our Mobile times AIoT segment revenue was RMB 84 billion. Smartphone times AIoT segment gross margin reached 20%. In the smartphone sector, the continued significant increase in memory cost has had an overall impact on the smartphone industry. We proactively optimized our product mix and increased pricing, resulting in a record high ASP for our smartphones. Revenue for this quarter was RMB 42.1 billion, accounting for 38.7% of total revenue. Our global smartphone shipments reached 31.2 million units. According to third-party data, in Q2 2026, our high-end or premium smartphone sales in Mainland China accounted for 32.1% of total smartphone sales in China, a record high. Despite the significant increase in memory costs, thanks to our continued product mix upgrades and strategies to balance scale and profit through our operational capabilities, our smartphone gross margin for this quarter was 8.5%. In the IoT area, our revenue reached RMB 31.3 billion in Q2 2026. Overseas revenue saw a significant year-on-year increase driven by expansion into overseas channels and a wider product range. Gross margin for IoT segment was 20.1% this quarter. Although some product categories were affected by memory costs and the reduction in national subsidies, from product category perspective, we ranked second globally in wearable wristband device shipments and second globally in TWS earbud shipments. Our tablet business rose to fourth place globally this quarter. For Internet services, we have accumulated a large user base globally. In June 2026, our global MAUs reached 770 million, up 4.8% year-on-year. Among them, MAUs in Mainland China reached a record high of 198 million, up 7% year-on-year. In Q2 this year, our Internet services revenue was RMB 9 billion. Gross margin for Internet services this quarter was 76.8%, up 1.4 percentage points year-on-year. Now let me turn to Smart EV, AI and other new initiatives. For this segment, revenue was RMB 24.9 billion this quarter, up 17.1% year-on-year, accounting for 23% of the group's total revenue. In this quarter, we delivered a total of 104,199 new vehicles with smart EV sales revenue reaching RMB 23.9 billion. Other related business revenue was RMB 1 billion, including revenue from AI big models or large models. Gross margin of smart EVs, AI and other new initiatives segment was 19.2% this quarter. In Q2 2026, we continued to increase investment in AI and other new initiatives. Results resulting in an operating loss of RMB 2.6 billion for this segment. In this quarter, our R&D expenses was RMB 9.2 billion, up 18.9% year-on-year. In first half 2026, our R&D expenses was RMB 18.2 billion, up 25.6% year-on-year with AI-related investments accounting for nearly 30%. In terms of net profit, the group's adjusted net profit for Q2 2026 was RMB 6.2 billion. In Q2 2026, our CapEx reached RMB 3.6 billion, of which innovative businesses such as Smart EVs and AI accounted for 65.8%. We remain committed to enhancing shareholder value and actively buy back shares in the open market. Since 2026, our share buyback amount has reached about HKD 11.7 billion, exceeding the total amount bought back for the entire past year or previous year. Regarding ESG, in the research and development of low-carbon materials, we continue to increase our independent R&D investment in core materials. Xiaomi's self-developed Xiaomi Titan Alloy 2.0 released in July reduces carbon emissions by approximately 93% compared to traditional primary aluminum. It has passed international environmental production international environmental product declaration certification and completed registration and public announcement, while maintaining strength, its toughness is further improved and all indicators meet the stringent standards for automotive structural components. It has already been mass produced and applied to the new generation SU7 Series and Xiaomi U7 series. This breakthrough in material technology further strengthens our independent technological capabilities and high-end product competitiveness in the smart EV business and helps enhance the low-carbon competitiveness of Xiaomi products in the international market. In terms of disaster relief, in July 2026, Guangxi, Hubei, Gansu, Liaoning, Jilin, Hubei and other places suffered from floods, hail, storm and other disasters. The Xiaomi Foundation donated RMB 10 million in cash to support emergency relief, transitional resettlements and post-disaster reconstruction in the affected areas. In terms of technological innovation, as of 30th June 2026, Xiaomi Innovation Joint Fund had funded a total of RMB 274 million in research grants, supporting 182 teams in Mainland China.

Thank you, everyone. That concludes what we wanted to share with you today. Now we can begin the Q&A session. [Operator Instructions] Thank you. We will now proceed to the Q&A session. First question is from Morgan Stanley, Andy.

I have 2 questions. First, about smartphone. In Q2, we saw that memory costs continued to rise significantly. That's the overall background in the industry. In Q2 for your results, your smartphone ASP rose to record high, even though shipment came down. However, the 2 factors offset one another. So the scale exceeded 40 billion gross margin, 8.5%. After this test in the first half, can you say that Xiaomi smartphone business is such that price shipment and gross margin can be dynamically balanced. In the coming quarters, Xiaomi can continue to raise price and control volume and improve product mix in order to maintain the stability of your smartphone business. Can you do that? That's my first question.

There are a few points that will affect our smartphone results and performance at the same time. First, as you said, memory costs increase and the extent of increase and the rhythm of increase. Last year, we already predicted that memory cost would increase over a long cycle. And to be honest, regarding last year and Q4 and also Q1, Q2 this year, we believe that the increase exceeded our expectation. The cost increase was indeed alarming. And memory cost increased. There is no way for us to just pass on the whole cost increase to our consumers. We have to rationalize our product lines and adjust our product mix. So from product launch planning, there is the need for a bigger cycle. So we have to work on that. Secondly, we have to adjust our sales strategies. For low-end phones, the entry-level phones, they are subject to the biggest blow or threat. So for the memory version, looking at Q2 cost, Well, smartphone retail price of over RMB 1,500 is being subject to the impact of memory costs, not to mention other costs. So that means that for entry-level product, it has to be sold at RMB 2,000 -- if we talk about normal memory capacity, not super memory capacity. So these factors added together mean that we need consensus on different fronts, not only unilateral consensus. After this round of adjustments, for Q2, we achieved this report card. Our shipments declined on a year-on-year basis. However, we are still #3 globally. Our ASP was at record high. It rose by almost RMB 300 And then our gross margin was at 8.5%. So I think last time you are worried about our gross margin. However, we still kept it at 8.5%. I think this exceed your expectation, so overall speaking, I would like to say that when we move into Q3, memory cost increase, we are there in Q4, well, I think it would be a slowing process of cost increase. For smartphone business, I think we are now in an appropriate and controllable situation. In the past, when the increase was of big extent, well, of course, that was a big blow to us. Now I think we are in a controllable situation. In Q3, I think you saw already -- so in the Redmi flagship product, well, in the past, we launched in October. Now we advanced to August. So Pro Max and [indiscernible] Pro are launched together. So I think they are doing well. For K90 Pro, -- so I think more than 62% was 16 Plus. And then the ASP is 5,000-odd -- and today, sales have not been completed yet. But then comparing with the last generation, I think the growth is 60%. So I think you can see that given this product, well, first of all, it sells well because of its product capability. And then for future trend, memory increase trend and price trend, I think we have formed an expectation. Final conclusion is that we are in a controllable condition for this business.

My next question is about EV. In July, you introduced the SkyNomad series, which had a market attention. And there are many very unique innovations for this model. In July, you announced the price of the 2 models and you started to accept orders. So Mr. Lu, based on the current situation for SkyNomad, what are some unique characteristics and changes comparing with other competitors SUV? What are some points that will attract car owners favor SkyNomad.

I think we said clearly that there must be difference in terms of user targets for SU7 and U7. So right now l I think we have very well achieved this target for user base. The duplication with SU7 is very small, so looking at the duplication situation I think we are talking about the drivers, we are touch importance to drivers experience for SU7, U7 but then for SkyNomad we are focusing on the special experience for the space, so it is very different, it's we want to create a very good space and there are alot of pictures in terms of users scenarios. So after launch we have received many small orders, and we have done some analysis first, I think family -- how a family buyer, a household buyers and also thicker number of passengers they will be the main buyers for the SkyNomad. And age is -- their age is few years older then the SU7 buyers, this is clear unique point, besides for our very flexible space, I think there has been alot of attention pay to it , user have not really experienced that yet, but then they have thought a lot of imaginations, when we designed this car we have not thought of so many special features, but now I think it can already satisfy many scenarios, and so I think in terms of order expectations we are optimistic.

Timothy Zhao from Goldman Sachs.

My first question is about AI progress. We can see that in the China AI industry in the past few years, innovations came fast and there were many good progress. So can you systematically share with us the pathway of AI monetization? What is your consideration in the short term, medium term and long term? What are your directions? And what is the progress so far? At the beginning of the year, for the full year, you mentioned RMB 16 billion of AI investment budget. So this year in the first half and also in Q2, how much is the overall AI investment? What do you think of the whole year AI investment plan? That's my first question.

Let me take this question. as you can see, and I think you pay close attention to our MiMo development. In the first half, we introduced MiMo 2.5, after that -- well, it is well received by global users. In terms of token call volume at the end of July in 1 week, we came first globally and monthly call volume also reached #1 globally. We think that this is -- this means that global developers had really voted for our model. And then when it comes to model capability and efficiency and also cost control capabilities, there is a systematic advantage. Now let me give you a preview. In the future, well, recently, we will introduce some new applications so that people can use our MiMo. We will launch the first desktop application so that on your PC, you can use our MiMo. Secondly, I think you will pay close attention to our new model launch. It is being trained now. Very soon, it will face the world. So there are many MiMo developments. Now just now you asked a question about progress of our commercialization. -- for AI, we are now in a large-scale investment phase. So we are not too anxious to pursue monetization. We'll continue to promote AI development in a steady and practical way. So we will put our focus on human car home ecosystem. When we launch our OS 4.0, MiMo large model and the HyperOS are already integrated. Hyper XiaoAi 2.0 is also a very critical step. In the future, there will be more progress. So we are making use of our MiMo large model and our smartphone, our OS so that they are deeply integrated. And then in terms of foundational models, we have opened up some API and token plan. Earlier, I mentioned our token plan and API, which have already contributed some revenue. However, this business is just a start. We do not treat monetization as our primary goal. So we are now iterating on our model capabilities. We will launch some new models. At the same time, the models will be used in our OS scenarios. There is some API revenue. However, we will not treat revenue as our main monetization goal at this stage. So I hope I have answered your questions.

Okay. My second question is about EV and new initiative segment gross margin. In Q2, there is volatility on a year-on-year and quarter-on-quarter basis. So can you give us a breakdown on the gross margin? What are the reasons behind the volatility in Q2? What will be the impact on AI gross margin? Can you give us an analysis?

For N70, N90 Max presale price after its launch, feedback has been positive. In Q3, Q4, what are some factors that will affect gross margin of EV and other new initiative gross margin? Regarding gross margin, you can see that our gross margin had fluctuated on a year-on-year and quarter-on-quarter basis. On a year-on-year basis, last year in Q2, that was the delivery stage of SU7 ultra. Ultra gross margin is better relatively speaking. So gross margin in Q2 last year was relatively higher because of Ultra contribution. In this quarter for Ultra, it was worsened a bit. And then if you compare Q2 and Q1, we delivered more new generation SU7. It was launched in March. And you can see SU7 and U7 share or breakdown are different from Q1. And in the launch, we said that SU7 cost increased quite a lot besides -- well, that's why for SU7 delivery, when its share is bigger in our deliveries, gross margin came down on a quarter-on-quarter basis. The third reason was already explained when our large model revenue started, there is negative impact to gross margin of the whole segment. So these are the three main factors. Now regarding your second question, just now we said that for SkyNomad N70 and N90 in September, they will be launched to the market. Well, looking initially, response has been quite good, but we have not set the final price yet. So GP margin will rely on final price. So that would affect Q3 and Q4 gross margin of SkyNomad.

Next question is from Kyna Wong of Citi.

My first question is about smartphone. So regarding inventory and raw materials in the first half, there was a big increase. Some is related to backup or inventory for future use. And then for your strategic plan, especially for memory and also your reserve the margin or within controllable range, how much is from the inventory to support your profit and gross margin. For finished goods, it also came down fast. Originally, from finished goods inventory to support and now, we can anticipate big gross margin pressure. So in the coming quarter, will this happen? What is your strategic plan or preparation? So should I ask my second question or leave it later?

Okay, regarding our inventory, there are 2 reasons: first, raw materials inventory rose on a quarter-on-quarter basis. Last year or last quarter, it's almost RMB 30 billion in raw material inventory. Now in Q2, it is about RMB 40-odd billion of inventory. There are 2 reasons here. The first point is we need to prepare some inventory for the future, especially about memory. Mr. Lu has shared with you his evaluation. So that's one point. At the same time, for memory unit price, it is rising. So it's not only about volume, also about selling price, it is rising. So the value will increase. That is about strategic inventory preparation and also selling price impact. Second, finished goods inventory came down. This is normal, because at 618 we prepared some inventory, and we actually released some inventory for sale. And so I think the -- it is easy to explain the reasons behind the decline. Okay. So looking at inventory reserve, Q3 pressure -- oh, sorry, sorry, another factor I just forgot. As you know, for new generation SU7 delivery started in early February. So in -- at the end of Q1, we also advanced some production of new generation SU7. And in Q2, overall inventory will come down. So there is also some impact from EVs.

Okay. Understood. Then I have a question about overseas business, IoT overseas revenue increased significantly in Q2. I for retail stores, overseas, you have 600-odd stores. This year, are you going to follow your original plan to establish stores overseas. Are you on track in relation to your plan? For 2027, there will be even more products and also EV and you are going to have an overseas plan. So are you going to do some preparation. For example, are you going to find land in overseas? And also, are you going to set up plans. And then for large area stores, do you have a plan about expansion. So for IoT and EV overseas business, how is your preparation and what would be your future outlook?

Regarding IoT and also room for development overseas, I think that there will be big room for growth. Now we are on track in implementation. So we have 630-odd Xiaomi home stores in overseas. And next year, I will continue our store expansion for Xiaomi Homes. I think it is going to be for IoT and high end or premium phones. That would account for 30% of our overseas premium phones. This is going to lay a very big foundation for overseas business. And secondly, IoT especially for large appliances, we have been blank overseas in the past. But I think in the second half of the year, we would accelerate our development next year. We will see a very big growth in this area. And then in different countries, I think there are a lot of barriers and thresholds in market access or entry. So there are a lot of compliance-related work to do, and that takes time. So we will accelerate our work, and then we will go into more and more countries. That's about IoT. And then for EV going global, it is on track. In the second half of 2027, we'll go overseas. And this year, we have been to many countries. And for overseas EV working partners and dealers, well, they pay close attention to us. They want to carry our products. They want to be our dealers. I think they are very passionate and behind that, there are several judgments. First, China EV is going global, is going to be mainstream in the future. This won't change. Secondly, when there are so many Chinese brands going global, which will they choose to work with. So this is a very important standard. I think they find Xiaomi very different and outstanding, we can be regarded as unique. And we are very few in China who can work on premium products, our premium brands. So we are very rare. And third, they see Xiaomi as a tech company, producing EVs, not traditional vehicles. So in our experience, communicating with dealers in different countries, they are mostly top dealers, top 10 dealers. And there are at least 7 to 8 who approach us on their own initiative. And then we also visit many countries and many customers.

Wen Hanjing from CICC.

I have 2 questions. First, Internet business. This year, in Q1 and smartphone shipments was under pressure because of memory can Internet service revenue still grew steadily. In the future, well, you have to consider pressure from memory -- that might be alleviated or that might stabilize. So for Internet business, what will be future growth trend? And what will be its gross margin?

Thank you. For Internet service business, as you know, there are 2 parts. First is related with shipments. Second, is related to our existing users. On these 2 fronts for our existing users, MAUs reached historical high already, 770 million globally. In Mainland China, almost 200 users. So for the existing users, they bring stability to our Internet business revenue. In the past few years, we expand the premium phones and we achieved some progress. So regarding MAUs, premium users account for a bigger share. Premium users made bigger contribution to our Internet ARPU more than low-end users. So when our existing users increase and also with more premium users and share, then this is very important to our Internet business revenue stability. In the past few quarters, advertising revenue saw healthy developments. In the past 1 or 2 quarters, I think you saw a so-called war in terms of food delivery, and there was a lot of advertising and this helped our smartphone advertising business. So I think these two factors have brought stability to our Internet business and also some strengthening.

Okay. My next question is about AI. So what are your thoughts apart from foundational models, you also mentioned some new applications. You also launched MiLM Cloud and you said that you are looking into smart appliance. So for the overall layout in the future, at what time point are we going to see your AI and the overall human car home hardware integration.

Well, for AI plan, first of all, foundational large models. Our investment is the heaviest in this area. In the future, they will service our human car home full ecosystem. That's the first line. So it is at the foundation level. And then on top of it, there are a few lines. The first one is smartphone. For smartphone, we have to complete the traditional an Android OS. So we will use AI to reconfigure it. And then for this process, it will take some time. But now with OS 4, I think we have made a step forward. On the smartphone, we can make use of the agents to do a lot of execution. I think you can see it. For Home, I think one important factor is Miloco. Recently, there are many scenarios in which Miloco is being realized. And there is cloud as well. Time precision is very good, but cost is still quite high. But when we optimize it, -- and in the future, when cost starts to come down, I think we can see better results. The third line is related with EV business. So smart driving, I think you are also interested in smart driving. And then the fourth line is robots. For our robotics business, we have introduced a number of versions, so these are the few lines. I think all these lines are deeply integrated with MiMo, and then they would develop independently, and they will also be integrated. And for our full ecosystem support, it will be stronger. And at the end of the day, we will see -- we will form a moat in relation to the human CarHome ecosystem.

Next question, UBS, Jimmy Yu.

My first question is related to AI. Looking at revenue, we can see some growth related to AI. So can you give more details. So what would be future trend? -- in relation to AI. That's my first question.

Jimmy, regarding AI revenue, we put it under smart EV, AI and other new initiative segments. Well, now MiMo just started to be monetized and the revenue is not big. In the future, at appropriate times, we may disclose it separately. But at this stage, we will still put it within this RMB 1 billion. Okay. Well, looking at the overall operating environment, well, there is still big pressure from memory cost. So this year and next year, when it comes to R&D expenses, is the trend based on what you shared in the past for R&D expenses? We have not made adjustments. Overall speaking, I think we are quite firm on that.

Next question. JP Morgan, David Chou.

My first question is about the EV market. So what is the share of sale of Xiaomi, Smart EV. So what's the question? You mean the share for the whole year. So what is the estimate of the second half of the year or this year?

Well, it's difficult for us to disclose a specific percentage or share. But our future goal is about these 2 series. So I think the SkyNomad is based on the Kunlun architecture. And our goals are in relation to the smart EV technology or driving.

Okay. My next question is about EV. EV ASP. Will it continue to go up? And what about your profit situation?

For ASP volatility, it is based on our product mix. For Q1, there are more deliveries of U7. And then in Q2, more deliveries are about SU7. SU7 price is lower than U7 as a result, ASP fluctuated. Last year in Q2, as I said earlier, our ultra accounted for a bigger share than this year. And Ultra ASP is a lot higher than the average. So that's why I think ASP is a result to us. It's not a goal. ASP relatively speaking, is not related to GP margin. It doesn't mean that when ASP is low gross margin will also be low. I think you need to understand this relationship.

[indiscernible]

I have 2 questions. Looking at Q2, smartphone ASP, it rose 21.9%. It is at historical high those RMB 3,000 or above accounted for a bigger share and for premium products, Well, we saw some decline. So if memory costs continue to rise, then global smartphone overall situation, how will it change? In the future, when it comes to this product mix, what would be your new strategies in the second half of this year and next year, what would be your product mix trend, please?

Okay. Now let me share with you our judgments. First, memory, is it true that the price will stay high over a long period of time, I have doubts. But is it much higher than the past low cost level for memory, I think so. Right now, it is 5x that of the Q2 last year level, but it's difficult to stay at this 5x level for a long time. But how many times eventually, it's difficult to say, perhaps in the middle. So when it comes to smartphone cost and price, then we can turn to the overall situation. I think we will see a global equilibrium. But in the short run, it is difficult to give a very precise estimate because this involves different tug of war, so to speak. For example, this year, Apple adopted a low price increase strategy. And some time ago, they said that they would also raise price for the coming new flagship phone and some other vendors said that they would also increase price, some that they were lower price and so on. So in the long run, I think people will consider memory costs increase or change in adjusting the selling price. I think there would be an equilibrium reached after dynamic adjustment. For the smartphone category, I think it is almost a necessity. So I believe that I am -- I tend to be more optimistic.

Second question is about robots. Just now you said you talked about the open source models. And the smart robotics factory had been commissioned already. So for Xiaomi robots. What is its positioning internally? Is it a tool to lower cost? Or in the future, would it become a new source of revenue? And do you have any quantifiable appraisal targets for this business?

Well, I think for robots, for final large-scale maturity and development, well, I think it will take a longer time. It's difficult to share some short-term goals. And secondly, we'll continue to invest. We are positive about future direction, and it is going to integrate with our existing business. So that's why we are going to invest in it. So up to now, we haven't thought of working in the food delivery area with our robotics. We have our own factories, and then there will be many scenarios in which applications will come out in the future, many of our capability realization will be integrated with robotics. For example, large models. They will also be synergized. And then chips and also systems I think robot development will also feed back to our large model capabilities. So I think we expect a lot of synergy with our other businesses. But at this stage, I think we will not set too many concrete or specific goals will be more pragmatic.

Thank you for your questions. We'll conclude the session here. Thank you for your time. I hope you will continue to give strong support to Xiaomi Group. Okay. Thank you all. [Statements in English on this transcript were spoken by an interpreter present on the live call.]