Good evening, ladies and gentlemen. Welcome to the investor conference call and audio webcast hosted by Xiaomi Corporation regarding the company's 2026 Q1 results.

Before we start the call, we would like to remind you that this call may include forward-looking statements, which are underlined by a number of risk 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 quarter 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 Q1 2026 results announcement conference call.

The first quarter of 2026 marks the first full quarter of our new five-year plan. Over the past 5 years, we have completed the strategic closed loop of our entire ecosystem encompassing human, vehicles or cars, and homes, and achieved comprehensive breakthrough in areas such as large appliances, EVs, chips, and AI.

Looking ahead to the next 5 years, in the short term, we face the challenge of a triple cycle of cost, demand, and competition. In the long term, we're entering a new phase where AI is reshaping our entire ecosystem of human, cars, and homes.

Tonight, I will mainly share three aspects with you. First, we will review our key performance in the first quarter 2026.

Second, we'll respond to issues that everyone is concerned about. Third, we'll look ahead to our strategic development direction and business focus for the next few quarters.

First, in Q1 2026, our group's total revenue was RMB99.1 billion; adjusted net profit, RMB6.1 billion. Looking at business segments: first, in Q1, we proactively controlled shipments of mid-to-low-end products and channel inventory.

While our smartphone shipment declined, our ASP reached record high. We also maintained our position among the top three globally.

According to Omdia data, our market share in Q1 2026 was 11.3%, maintaining our position among the top three globally for 23 consecutive quarters. Our smartphone shipment ranked second in Latin America with market share of 17.4%, up 2 percentage points compared to the previous period.

We ranked third in Europe, Southeast Asia, and Middle East and Africa with market shares of 17.2%, 16.9%, 13%, and 9.2%, respectively. In Q1 2026, we ranked among top three in smartphone shipment in 47 countries and regions worldwide, and among top five in smartphone shipments in 65 countries and regions worldwide.

For GP margin, the sharp rise in memory costs in the short term has ushered in a new normal for the entire industry. We'll not simply pass on the increased cost linearly to consumers.

Instead, we'll reposition user needs and achieve a balance between scale and profit through product matrix upgrades, software optimization, and leveraging our operational capabilities. In Q1 2026, we achieved gross profit margin of 10.1% for smartphone, which reflects the operational resilience brought about by our own capabilities.

IoT business: in Q1 2026, IoT business achieved revenue of RMB24.7 billion, mainly due to a year-on-year decline in domestic revenue caused by the high base of national subsidy last year. However, thanks to expansion of overseas channels and the increase in overseas product categories, overseas revenue reached record high and achieved double-digit year-on-year growth.

Among them, TWS earbuds ranked second globally, wearable products ranked third globally, and tablets ranked fifth globally. For gross profit margin of our IoT business this quarter, it was 25.2%, which fully demonstrates our ability to hedge against fluctuations in a single industry through multi-business synergy.

This quarter, the business can provide more profit to mitigate the risk of declining gross margins on smartphone business. For Xiaomi EVs, we delivered 80,856 cars in the first quarter 2026, mainly the Xiaomi U7 series.

As of May 6th, 2026, our new-generation SU7 has accumulated over 80,000 pre-orders in 48 days since launch. This demonstrates that our EV business has withstood severe challenges and has emerged from most difficult period.

As of 30th April 2026, Xiaomi U7 has achieved cumulative delivery volume of 232,000 units in 10 months. On 21st May, we officially released Xiaomi U7 GT, positioned as a sports-car-level SUV suitable for long-distance travel, priced at RMB389,900.

After extensive tuning and testing, Xiaomi U7 GT finally broke the SUV lap record at the Nürburgring, becoming the fastest SUV in Nürburgring history. Xiaomi U7 standard edition is priced at RMB233,500.

To refine its extreme handling capabilities, Xiaomi U7 GT underwent testing at the Nürburgring Nordschleife. The issues of AI and androids that everyone is concerned about: we have emphasized on multiple occasions that deep integration of AI with physical world is the next stage of intelligent technology.

Xiaomi, with its hardware ecosystem reaching over 1 billion users, has a tremendous opportunity to become a leader in the AI era. If we don't create our own large-scale model, our understanding of the base model will always be like looking at flowers through the fog.

Only by mastering core technologies end-to-end can we truly create products with differentiated competitiveness. The industry is still exploring the ultimate form of the combination of AI and hardware; however, only those with core technologies and application scenarios can quickly keep up when the trend becomes clear.

Latest progress: in April 2026, we released Xiaomi MiLM 2.5 series. Xiaomi MiLM 2.5 series includes MiLM 2.5 Pro, V2.5 TTS series, and V2.5 ASR.

This represents a comprehensive leap from usable to easy-to-use. On Artificial Analysis global authoritative comprehensive intelligence ranking list for large open-source models, Xiaomi MiLM 2.5 Pro ranks first in the Comprehensive Intelligence Index among global open-source large models and is among the top five in the overall global ranking of large models.

At the same time, the Agent Index ranks first among global open-source large models. At end of March, MiLM-V2-Pro set another record on OpenRouter platform, achieving first place in the daily, weekly, and monthly rankings.

In the week of March 31st, MiLM-V2-Pro's weekly token consumption surpassed 4 trillion. After free trial period ended, our average weekly token retention rate reached 35%.

On April 3rd, 2026, we launched Xiaomi MiLM Token Plan, offering four tiers: Lite, Standard, Pro, and Max, aiming to help users improve productivity at reasonable prices. Since launch of our token plan, the Pro and Max tiers have accounted for over 50% of revenue.

According to OpenRouter data released on May 12th, MiLM ranked first in terms of Hermes Agent model call volume, contributing an cumulative 1.45 trillion token calls in the past month. This means that the world's fastest-growing open-source agent products has chosen Xiaomi MiLM as its preferred inference engine in real-world, high-intensity task scenarios.

To reward global developers, Xiaomi officially launched MiLM Orbit 100T Token Program, distributing token benefits to AI users worldwide for free, with a plan to distribute a total of 100 trillion tokens within 30 days. As of morning of 12th May, nearly 80 trillion tokens have been distributed.

We'll continue to iterate on our base model. The moment of AI mobile phone has arrived.

Perhaps 2026 will be a pivotal year for AI smartphones, and arrival of such a moment will profoundly influence future changes in mobile phones. AI phones are not simply adding AI features to existing smartphones, nor are they simply creating an agent app.

Instead, it shifts from an app-centric interaction approach to an OS agent-centric interaction approach. Mi Claw is our initial attempt.

Recently, CAICT launched an evaluation of mobile smart assistants, and Xiaomi Mi Claw became one of the first mobile smart agents in China to pass this authoritative evaluation. At the same time, the packaging and testing categories of our Mi Claw have expanded to multiple terminals, such as tablets, PCs, Macs, and speakers with screens, marking significant upgrade in cross-platform capabilities.

In the future, mobile phones will be upgraded from operation tools to executive assistants and digital clones for people. We will deeply evolve our Surge OS, taking the agent as the core, and further increase the deep coupling between model and upper-level framework.

We continuously enhance product experience of assisted driving. In March of this year, we officially released Xiaomi XLA cognitive big model architecture, enabling Xiaomi's assisted driving to move from perception and imitation to understanding and reasoning.

In May 2026, building upon XLA, we took it a step further and officially released and fully open-sourced the Xiaomi 1VL autonomous driving model, a 1-step latent-space language-visual reasoning framework. It unifies the three major technical approaches: VLA, world model, and latent-space reasoning into a unified framework.

Then embodied robots: we continue to explore the boundaries of physical intelligence. Embodied robots are the ultimate integrated platform for AI capabilities, chip and OS capabilities, and manufacturing capabilities, and have extremely high barriers to entry.

On April 27th, we'll bring a new capability demonstration and officially release the full process of training on a real device with Xiaomi Robotics Studio. On the same day, Xiaomi robot, which had turned from tightening screws in a factory, came to the Xiaomi Group Investor Day event and made a heart shape with guests, winning praise from many friends.

The current progress is just the beginning. Robots will open up entirely new industrial and market opportunities for Xiaomi.

We believe that we are going to achieve a lot of good results. We're going to be a leader in hardcore technology, will continue to do more in innovation and many other smart areas, will continue to innovate and upgrade and support our healthy development of various segments.

That's all in my part, I'll pass the floor to Mr. Alain.

Thank you, Mr. Lu.

Hello, everyone. Good evening.

As Mr. Lu shared with everyone, 2026 will be a year of both short-term challenges and long-term opportunities.

Let me talk about our results. In Q1 2026, our total revenue was RMB99.1 billion.

Overall gross margin was 22%. Looking at various segments, our Smartphone × AIoT segment revenue was RMB79.3 billion.

Smartphone × AIoT segment gross margin was 22.5%, up 2.5 percentage points quarter-on-quarter. For smartphones, the continued significant increase in memory costs has had an overall impact on the smartphone industry.

We focused on optimizing our sales structure and channel management. Revenue in this quarter was RMB44.3 billion, accounting for 44.7% of our group's total revenue.

Our global smartphone shipments reached 33.79 million units. Our average selling price for smartphones, thanks to our strategic adjustments, reached a record high of RMB1,310, representing a year-on-year increase of 8.2%.

According to third-party data, in first quarter 2026, our high-end or premium smartphone sales in mainland China accounted for 23.5% of our overall smartphone sales. According to Omdia data, in the first quarter 2026, we ranked among the top three globally in smartphone shipments with market share of 11.3%, maintaining our top three ranking for 23 consecutive quarters.

Despite rising memory prices, through proactive control of mid-to-low-end phone shipments and channel inventory, our smartphone gross margin remained relatively healthy at 10.1%. Regarding IoT, in Q1, our revenue of IoT business reached RMB24.7 billion, with overseas revenue maintaining steady growth, achieving double-digit year-on-year growth and setting a new record.

We prioritized profitability, refrained from participating in industry price competition, and simultaneously expanded our overseas channels and increased our overseas product categories. This quarter, gross profit margin for IoT reached 25.2%, up 5.1 percentage points quarter-on-quarter.

From product category perspective, this quarter, we ranked third globally in wearable band shipments and second globally in TWS earbuds shipment. Our tablet business performed steadily, ranking among the top five globally this quarter.

Regarding internet services, we have accumulated a large user base globally. In March 2026, our global MAUs reached 750 million, up 3.8% year-on-year.

Among them, MAUs in mainland China reached a record high of 196 million, up 8.1% year-on-year. In the first quarter 2026, our internet services revenue was RMB9.5 billion, up 4.3% year-on-year.

In this quarter, gross margin for internet services was 76.1%. Advertising business continued to drive internet business growth.

Advertising revenue of this quarter reached RMB7.1 billion, up 7.8% year-on-year. Let me talk about intelligent or smart EVs, AI, and other new initiatives segment.

Revenue reached RMB19.9 billion this quarter, up 6.9% year-on-year, accounting for 20% of the group's total revenue. Specifically, due to a decrease in quarterly deliveries of the SU7 series, we delivered a total of 80,856 new cars in Q1 2026.

For smart EV sales revenue, it reached RMB19 billion; for other related business revenue, it was RMB900 million. Our average after-tax selling price this quarter was RMB235,000.

Affected by purchase tax subsidies and rising raw material costs, gross profit margin of smart EVs, AI, and other new initiatives segment was 20.1% this quarter. In Q1 2026, our smart EVs, AI, and other new initiatives segment incurred an operating loss of RMB3.1 billion.

As of 23rd April 2026, cumulative deliveries of the new-generation SU7 exceeded 26,000 units. As of 30th April 2026, cumulative delivery of Xiaomi U7 in the past 10 months reached 232,000 units.

In the first quarter 2026, our R&D expenses was RMB8.95 billion, up 33.4% year-on-year. CapEx reached RMB3.27 billion, up 20% year-on-year.

Of this, smart EVs, AI, and other new initiatives business accounted for 45.6% of total CapEx. For net profit, in the first quarter of 2026, adjusted net profit was RMB6.1 billion.

At the same time, we focused on enhancing shareholder value and actively repurchased shares in the open market. Since the beginning of 2026, our share buyback amount has reached about HK$8.4 billion, exceeding the total amount for the entire previous year, demonstrating the company's confidence in our long-term development.

We actively practice the concepts of sustainable development. In April 2026, we released the Xiaomi Group 2025 ESG Report, marking our eighth consecutive year publishing ESG reports.

The report comprehensively showcases our strategies and achievements in 2025 regarding data privacy governance, responsible marketing and service, circular economy, sustainable supply chain, talent development, and corporate governance. In terms of ESG ranking, in March 2026, we maintained our A rating in MSCI ESG rating.

Besides, we received the top 1% of Chinese Companies honor in the S&P Global CSA score and were included in the S&P Global Sustainability Yearbook China Edition 2026, demonstrating continued recognition of our ESG initiatives. Thank you, everyone.

That concludes what I wanted to share with you today. Now we can begin the Q&A session.

Thank you, Alain. Now we will move on to Q&A.

In order to allow more investors to ask questions, please limit your questions to a maximum of two. Thank you.

Mr. Lu, Alain, good evening.

Congratulations. Given the pressure in smartphone industry, you still performed and exceeded expectation.

I have two questions. First question: for AIoT business, in Q1, gross profit was very good.

Last week, after the launch, your earbuds also sold well. This year for AIoT business, while achieving high profit, high gross profit, in terms of revenue scale, can you achieve good performance in the coming two quarters?

When it comes to new product launch and overseas market developments, what are some significant highlights? Thank you.

AIoT business, I'm sure you all are interested in that. Last year in Q3, we already expected that there would be memory cost increase, and it is going to be a super long cycle, and the extent of cost increase is also big.

For products relying on big memory, like tablets and so on, that would be bigger impact. Memory cost accounted for quite a big share of some of these products.

AIoT business to our group is important. We want to achieve a balance and also alleviate the pressure from memory cost.

In Q3 last year, for AIoT business, we enhanced its importance. In mainland China and overseas, we adopted different strategies.

For China market, we focused on premiumization. This year, I think you gradually see many of our products.

Our products are doing well. They are well received by users.

We achieved good word of mouth. In recent launch, we introduced very good product with good external appearance, and also sound quality, and also the connection with super AI assistant, and so on.

At the same time, you also see our air conditioner. It became very hot, and it also sells very well.

Users like our air conditioner because it cools very fast, and also the fan effect is very good. Then for refrigerator, it keeps the freshness of food in a very good way.

Then it is well received, especially when people need to redecorate their homes. Progress has been very good within mainland market.

For overseas, we are expanding our scale. Overseas, our share in overseas market is still very small.

Then for our AIoT business, there are many markets that we need to enter. Given this situation, last year, basically, we have solved the issues in many countries.

Now we're in many countries, and we have expanded our product mix. Besides, we collaborated with many other e-commerce players, and so we're able to double our growth.

Here, if you look at our overall AIoT business growth, we're able to achieve a good gross profit margin. In the future, I think potential is still huge.

If you look at the China market, our earbuds, our watch, well, our share is still on the low side. There is still a long journey for us for overseas.

Overseas market is two times the potential of the Chinese market, and the competitive environment of overseas is different from the China market. I think for IoT business, that's my answer.

Let me supplement with some numbers. Just now Mr.

Lu talked a lot about strategies, and categories, and channels, and also about our overseas expansion of IoT. Now, this is not a peak season or busy season, but for AIoT revenue overseas in Q1, it reached double-digit growth year-on-year.

It already reached a record high in history, and it is around 40% of the overall AIoT total.

Okay. Thank you, Mr.

Lu and Alain. My second question is about Xiaomi EVs.

After the launch of U7 and also looking at the buyer profile regarding new-generation SU7 and last year U7, what are the differences in user profile? With the launch of the two new models, in terms of sales volume and profits, comparing with the situation without these two new models regarding sales volume, revenue, and profits, and important metrics, what will be the impact of these two new models?

For U7, you can see that we have introduced two new products: so the U7 standard, and also we have introduced a GT model. For the two models or versions, one important idea is for SU7, comparing with Model 3, I think they may be rather similar.

Then for U7, I think there are still differences. After doing some analysis, for Model Y, standard or entry-level version, it accounted for 70% of the sales, and we realized that we may have some disadvantages in terms of cost.

Then for user scenario, 600 kilometers range, I think their needs can be met. When it comes to even longer range, then there may be cost issue.

Given all these considerations, we have introduced a standard version and also a GT version. Regarding the GT version, we have already broken the SUV speed in the Nürburgring ranking.

It is a ceiling of SUV functionality. For these two products, in the future, I'm sure they can actually complement to our overall product line.

After users have bought the cars, more than half of the users choose our RMB429,900 option. That is the full configuration.

More than half of them have chosen this version. I have done a market research, and I realized from the research that there are some U7 users who are buying this, and also many of the users are owners of Benz, Mercedes-Benz, and BMW cars.

Right now for GT, at the initial launch, every month the capacity is around 2,000-odd units. I think if you need it, you can continue to place order.

The cycle will be shorter. For the standard version, I think that it can very well solve the daily commuting needs.

A range of 600 kilometers roughly. I think the price can also suit your budget in that case.

I think in my market research, I asked the frontline colleagues. Many of them joined Xiaomi from Tesla.

They said that as long as users can give it a trial of our product, then they will choose our model. When it comes to car size, degree of comfort, functionality, and so on, our products are very satisfactory.

For gross profit margin, the price difference is RMB20,000. Then for GT, I think gross margin will even be better.

Now sales volume is not out yet, I cannot really give concrete comment. Then on the total volume, I think GT is going to do well.

Thank you.

Thank you, Mr. Lu.

Thank you.

Okay. Mr.

Lu, Alain, good evening. Thank you for taking my questions.

I have two questions. First, AI.

We are happy to see that this year you have exceeded market expectation in your AI progress. When it comes to AI business development and also a bigger investment by many AI companies, well, this year you're talking about RMB16 billion of AI investment budget.

From MiLM 2.5 launch all the way till now, do you have other operation data to share with us? For MiLM large model rhythm, how will it be?

Then for Mi Claw, what will be the future plan, and also next-generation AI OS roadmap and timetable? How do they look like?

That's my first question. Thank you.

Okay. Let me comment on our Mi Claw.

For AI expenses, I'll ask Alain to supplement. For Mi Claw, now we are promoting it with full force.

With Claw's emergence, I think we are very quickly launching Mi Claw. Then we are trying to make sure that it can be used with our tablet, our wearables, and also our future EVs, our soundbox, and so on.

In the future, what is important is whether we can connect user data and then whether we can improve user experience. I think it is going to enhance experience a lot.

Of course, all these cannot depart from large model and our framework and our AI OS, HyperOS transformation. I think all these are important.

Now we are proceeding with full force, and our team is collaborating. This year for the OS launch, we are moving towards that direction.

This is important.

Okay. Another question is about AI token and so on.

Let me make a few points. As you can see, right now we have already started some token attempts.

We have introduced different token packages for our users. Far, feedback has been quite good.

As Mr. Lu said, 50% of the users are using our new versions, especially for the high-end models.

The use and deployment is very high. As a result, there are revenue related to AI token.

You asked about our AI investment and also our judgment. At the beginning of the year, we mentioned 16 billion.

When our AI business develops, we will also continue to adjust this number because we realize that there are huge opportunities about AI. Regarding model iteration, if we get users' recognition, then we will also release more volume for inference.

In the future, there may be a higher budget. Then this has to be based on our clear ROI calculation when making new investment.

Thank you.

Okay. Thank you, Mr.

Lu and Alain. Now my next question is about EV business.

Looking at delivery in Q1, more than 80,000 units; in April, more than 30,000 units, comparing with the whole year 550,000-unit target. Seems that the run rate shows a difference.

What is your plan about new model? Then for EV business and new initiatives gross margin in Q1, on a quarter-on-quarter basis there is decline.

Alain just said that there is cost increase pressure. At the same time, there is one-off impact from cross-year order and also tax subsidies.

Can you quantify the impact? Then for overall gross margin, how should we look at it in terms of new vehicles?

Yes. Let me first comment.

In Q1, in January and February, there is only delivery of U7. For SU7, we are talking about this iteration.

We have spent 2 months of not selling the old products. We have also looked at the fast iteration of our other peers, and so we have spent 2 months to work on our old products so that users know that we are launching the new product, new SU7.

After doing that, users' word of mouth is very good. That's because of our huge effort.

Then in January and February, in Q1, only 80,000 units were delivered. Then in April, our delivery increased in April.

For SU7 delivery, the delivery lead time was rather short. Actual delivery was a lot more than 30,000 units.

We do look forward to bigger volume. Then there is a large model that would be launched in the second half of the year.

We believe that that model is going to be like a new platform. That model is very innovative.

It is highly competitive. We have a lot of confidence in that model.

For the whole year target, we have confidence that we can attain the full-year target. Regarding EVs, in this quarter, you can see that the gross margin of our EV business had declined.

As you said, first of all, there is the purchase tax subsidy issue, that's the main reason. Regarding the end of last year, there were undelivered orders, and then there was the VAT subsidy.

So 10,000 to 15,000 yuan was the amount. ASP cost impact on gross margin, that's the first point.

Secondly, we started to sell some display car, and so that would also affect ASP and gross margin. At the same time, for the macro environment, there are issues regarding batteries, and memory, and also commodities, which would increase our cost.

However, there are some factors that affect the whole industry. Relatively speaking, our ASP is higher than our peers.

There is cost increase, but then it did not affect our gross margin that much. That's one point.

Even though gross margin has declined, our overall gross margin within our industry is still on a high side. As Mr.

Lu just said, we delivered 80-odd thousand units in this quarter, which would affect the apportionment of our fixed cost. Now we're going to deliver the new-generation SU7 and also the U7 standard, U7 GT.

On a quarter-on-quarter basis, I think volume will increase. Secondly, GT delivery will drive our quarter-on-quarter ASP.

Besides, regarding suppliers, we will continue to put in place some cost reduction measures, there'll be some hedging. These are the few factors that will affect our gross margin in the coming few quarters.

Thank you, Mr. Lu, and thank you, Alain.

Good evening. Thank you for the opportunity.

I have two questions. First, about the smartphone business.

This year, memory cost pressure was big. In Q1 for your smartphone business, shipment and gross profit had still shown some resilience.

In the future, given such crisis and challenges, how can you balance various metrics, including quantity, price, and gross profit?

Right. Thank you.

We have discussed that many times. This trend about the super cycle of memory cost increase did not differ a lot from our judgment.

The cycle is very long. At the same time, the increase is of a big extent.

Since Q2 last year, the cost started to increase, and then in Q3 as well. Then in that quarter, the cost increase was big.

In Q2 this year, based on market quotation and also contract price, if we look at contract price, I think the increase is still big. When it comes to cost increase and also cost for smartphone, this is a very big challenge.

Besides, it is not going to end yet. It is a very long cycle.

I think we have to look towards 2027 and 2028. In Q3, well, in the past the increase is 60% to 80% or even 100% on a quarter-on-quarter basis.

I believe that in Q3 this year, the increase will slow down, but still the pressure is going to be huge because the base is very high. Even though there is just a small increase, the absolute value is very big already.

The impact on the overall industry is big, not only for smartphone. For many products, memory cost account for a small share.

I believe that there would be huge impact on the overall consumer electronic industry. Given that, we have done a few things.

In Q1, you can see that we have made some adjustments. First, we have to dynamically balance selling price, cost, volume, and gross profit margin, and the relationship among these factors.

Short-term price increase or price reduction may not be very wise. We have to come to adjustment about the market.

In Q1, for volume, there is some decline; but for ASP, it increased by about 10%, and our gross margin rose 2 percentage points. This shows that our strategy is effective.

Besides, this year, we introduced some new products, Max and some other new products. In terms of cost, I think we cannot just simply pass on the cost increase to the consumers.

If we just increase the price linearly, it is difficult for consumers to accept. We have to look at the users' needs as well.

We have to make products that can meet users' needs, and then sell the products to users at a reasonable price. That should be something we need to do.

We have to rationalize our product lines. Given the overall environment of price increase, I think we are still doing well in the China market.

If you look at the VCA activation, we have 14% market share. This is because of our very important strategy.

Besides, we tried our best not to increase price for all products. Even if we increase price, it is a small one only.

We are the last to increase price. We started to increase price in mid-April.

Overall speaking, users still understand our stance. We have put in place this overall strategy to face up to the coming increase in this super cycle.

Given this environment, we have to seize the AI phone opportunities and AI-related opportunities.

Okay. Thank you, Mr.

Lu, for the clear answer. My next question is about your going abroad plan.

Just now you shared with us your progress in EVs. Regarding going abroad and also your rhythm of new EV launch and also the increase in number of overseas sales centers, can you do more sharing with us?

For going abroad, I think comparing with the last quarter, we did not differ a lot. In Q3, Q4 2027, we have to go abroad in terms of store opening.

We'll first go to advanced countries, and then developing countries. First high-end, and then medium-to-high-end.

That's our strategy. Far, for our team, we are doing preparation for going abroad because going abroad is complicated.

There are a lot to do. There are a lot of legal and compliance requirements, and there are also localization needs for product match.

Then we have to also make plans about network and sales centers. We are actually doing all these preparation works.

Thank you.

Thank you, Mr. Lu.

I have no further questions.

Mr. Lu, Alain, greetings.

I am from CITIC Securities. Thank you for this opportunity.

I have two questions. First, about MiLM large model.

You just said that the daily average deployment already exceeded 1 trillion tokens. I want to ask right now for token plan, what is the user conversion rate and retention rate?

For your AI team, in terms of future commercialization, do you also have internal appraisal targets or metrics? My second question is about AIoT business.

Just now Alain said that for overseas AIoT business in Q1, the number and share rose fast. In the coming three quarters, especially towards the end of this year, for overseas AIoT business, how much will be the growth?

Would that be new guidance for growth and share?

Thank you. Let me first comment on AIoT.

Alain can talk about MiLM large model. I think the first question is this: we still think that AIoT business is, in the coming few years, the way that we can balance out the increase in memory cost and the impact on our business.

This is going to be a very important balancer for our business. For AIoT business, we want it to grow fast.

Regarding AIoT business in the China market, we insist on premiumization. Premiumization is the key.

In the recent launch, you can see that we have launched a few products. All those launched products have reached our internal targets.

For earbuds, it had exceeded our target two times. For our products, I think our product capabilities have been enhanced a lot, and market acceptance is also very good.

In overseas market, the room is very big. Our share is just quite low still.

We have to do more product development overseas. Some products can satisfy overseas needs, but then very often there are thresholds and barriers that are still high in overseas.

In the past 2 years, I spent a lot of time to work on that. The third issue is channels.

For channels, starting last year, we put in investment in e-commerce. I think that for overseas AIoT development will still meet a lot of resistance.

AIoT business overseas growth can be fast, and I believe this is going to be a long cycle. The room is huge.

Overseas AIoT business will see three times, four times growth potential.

All right. Let me get back to your first question about deployment of MiLM.

Let me make a few points. We have introduced token plan.

After that, we saw a few interesting phenomena that I can share with you. First, fee-paying volume is high.

Token deployment exceeded 30% of fee paying. Fee-paying ratio is very high.

Besides, for token Max version, token deployment exceeded 50%. Third point: overseas number is high.

The share overseas exceeded 50%. Retention rate: we did not disclose this number, but I think retention rate is also high.

However, we are now working on AI models. Far, we have not reached the business closed-loop stage.

If you look at many analyses about China and overseas, there are many new models being launched. We are still in the iteration cycle.

We hope that when users deploy tokens, we continue to learn and iterate our model. In the future, we still need to follow the market to adjust our token plans.

In the growth stage, we will collect data and iterate our models. That's the stage that we're in.

We have still not reached the ultimate token plan fee model being a large business. I hope you can understand this.

Okay. Thank you, management.

I have no further questions.

Thank you, Mr. Lu and Alain, for the opportunity.

I want to follow up on AI large models. Okay.

Right now in the market, there are many large models companies that will launch some new versions in June and July. Once every 2 to 3 months, there will be an iteration.

Regarding MiLM large models, what is the renewal time point? Are you going to follow the market rhythm, or are you going to have your own plan?

You have your own strategic consideration. Just now you talked about deployment.

We have seen some price decrease. It seems that operators are launching large models, and they also have some sale plan.

What is your positioning in the market? What are your strategies?

Regarding your company's efficiency enhancement, how can it be seen? Will it be realized in OPEX?

That's my question about AI. Second question is about EVs.

For Q1, EV, AI, and new initiatives business segment operation loss is 3.1 billion. In the future, as Mr.

Lu said, you have much confidence in reaching the target. In the coming few quarters, we'll see more than 100,000 breakthrough in shipment.

For operating loss, I think in Q1 that will be the chance to cover the operating loss. You still need to look at the AI investment of your company.

EV, AI, and new initiatives segment for this year, how will be its profitability? What will be its direction?

Thank you.

Okay. Let me talk about AI, and then Alain can comment on the next part.

We do not have an internal target of iteration once every 3 months. We do not have a specific timeline.

Regarding model iteration, we do not have the need to announce it. We should not announce for the sake of announcing.

We have our own rhythm and our grasps for AI large models. Xiaomi MiLM large models should serve our business.

It must be coupled with our business. This is very important.

Now we have many business centered around scenarios with the data and also the AI large models. We also have to look at its cost.

Without business closed loop, it is difficult to assess. We need to strengthen our large models.

At the same time, we will also sell tokens to the market. Of course, that depends on market competition that may be helpful to Xiaomi.

We will take this two-pronged approach. I think it will be a balance between medium and long term, and also horizontal and vertical.

This will be a more favorable approach. I think this is the biggest difference between Xiaomi and other companies.

Right. Yes.

Let me supplement on a few points. Just now Mr.

Lu said that we are not a system focusing on large models. Our main business is not to sell tokens.

We will not announce or launch new models, large models for the sake of launching. We will launch when we are ready.

Regarding internal enhancement, we encourage our employees to use our MiLM models more. This is to help them enhance efficiency, and they can also generate more data to help us to iterate our own models so that our models will be better.

This is part of efficiency enhancement. It is not only to enhance our own efficiency, but also the capability of our models.

These are two parts. Your second question is about our EVs.

EV, AI, and other new initiatives future direction: I think you can see quite clearly that in Q1, comparing with the previous quarters, the biggest difference is deliveries have come down. Just now we have explained that in this quarter, we did not deliver SU7.

We focused on U7, and we prepare for the new-generation SU7 launch. There was some impact on revenue.

In the future, in Q2, quarter-on-quarter delivery for EV will rise, for sure. That's the first point.

Of course, we will increase our investment into AI. In the future, I think these are the two main results of our operating profit or loss.

It will be the result of these two factors together: volume, gross margin, and AI investment.

Thank you.

Mr. Lu, Alain, good evening.

I'm from JPMorgan. I have a question.

My question is, in 2025 second quarter, delivery reached the peak, and then after that, it declined. What are the reasons behind?

How can you get back the expected growth momentum? Thank you.

Okay. For the expected delivery, you said that there was a decline from the peak.

The main reason is, in Q1, first of all, when we made the expectation, at first the volume was big, but later on delivery cycle was rather long. When consumers waited, there was also some disruption or disturbance.

Some users switched to other brands. We saw some peer brands which had locked in some users' orders.

Some users had switched to other brands. Looking at our product line layout, basically, we are benchmarked against Model Y.

Then within the cities, for daily commuting, range is around 600 kilometers that can satisfy users' needs. We have also got another model regarding 800 kilometers.

For 600 kilometers, I think that range can satisfy people's commuting need already. For the overall car model launch, we have already added one more model.

We have now U7 standard and U7 GT. After the launch in the past weekend, we saw quite good response from users test driving.

I think gradually we will get closer to our expected volume. Thank you.

Thank you, Mr. Lu.

I have no further questions.

Thank you, management, for the opportunity. My first question is about smartphone and impact on sales volume.

Then you will try your best to achieve some balance. In Q1 this year, gross margin was rather steady on a quarter-on-quarter basis.

My question is, your products have been enhanced, that is part of it, and besides, there is global inventory factor. If you look at smartphones and consumer electronics in the future, what is the outlook on gross margin?

What would be the trend, taking into consideration cost increase for memory?

I think given the cost increase, pressure will continue. That is an objective fact.

We need to solve this problem. We need to balance scale, profits, and also gross margin.

We need to achieve a balance. More importantly, when cost increases, we have to ask a simple question.

Now users may need to spend more time. Are they still willing to do that?

We cannot simply pass the cost increase to consumers. Given the new cost structure, we need to make sure that users are happy with the value for money.

We have to satisfy our own values as well, that is important. Besides, we need to adjust our product structure and also quality.

Scale decline is for sure, but I think ASP increase can offset part of the impact from scale decline. We need to work on our GP margin.

Then I think gross profit will be controllable. That is our overall thoughts and strategies about smartphone business.

Thank you for your question. Because of time, we will conclude the call here.

Thank you for your time. I hope that you will continue to support Xiaomi Group.

Thank you, and goodbye. Okay.

Thank you.