From the perspective of revenue structure, Xiaomi’s smart home business has firmly become the group’s second largest profit pillar.
In 2024, Xiaomi’s IoT business revenue will exceed 100 billion yuan for the first time, reaching 104.1 billion yuan, a year-on-year increase of 30%; it will further increase to 123.2 billion yuan in 2025, a year-on-year increase of 18.3%.
Although Q1 of 2026 was affected by the high base of the decline in national subsidies for domestic home appliances, single-quarter IoT revenue of 24.7 billion yuan fell 23.7% year-on-year, but business quality has improved significantly, gross profit margin has reached a historical high of 25.2%, and gross profit contribution has risen to 28%, exceeding the 20% gross profit ratio of the mobile phone business, which has hedged against the pressure on mobile phone profits caused by the increase in memory chip prices.
The growth in user scale is also solid: as of Q1 2026, the number of connected devices to Xiaomi’s AIoT platform has reached 1.119 billion, a year-on-year increase of 18.5%; the number of users with 5 or more smart devices is 23.6 million, a year-on-year growth of 22.3%, which is higher than the growth rate of total devices, indicating that the depth of users’ ecological use is continuing to increase.
Mijia APP’s monthly active users reached 117 million, and Xiaoai’s monthly active users reached 169 million, both hitting record highs.
Overseas markets have become the largest growth engine at present: overseas IoT revenue in Q1 2026 will be close to 10 billion yuan, accounting for nearly 40% of total IoT revenue, achieving double-digit year-on-year growth.
Among them, the IoT market share in Europe, Southeast Asia, and the Middle East and Africa ranks among the top three in the industry, and the growth rate in Latin America exceeds 40%. The management clearly disclosed that the potential space in the current overseas market is twice that of the domestic market. However, Xiaomi’s current overseas IoT revenue is only half of the domestic one. In theory, there is still room for four times growth.
From the perspective of technology research and development, Xiaomi Smart Home has shifted from the cost-effective route of stacking hardware in the early years to the stage of AI-driven active intelligence.
Xiaomi plans to invest no less than 16 billion yuan in AI-related research and development in 2026, and the cumulative investment in AI in the next three years will exceed 60 billion yuan.
The newly released MiMo-V2.5 Pro large model is on the global open source evaluation platform Artificial
Analysis ranks first in the Agent Index and has been implemented in all smart home scenarios: the AI intelligent agent miclaw equipped with a large model can automatically disassemble complex instructions to complete cross-device linkage. For example, if a user says “Help me prepare a rest environment when I get home from get off work,” the system can automatically start the sweeper for cleaning in advance, link the air conditioner to adjust to the appropriate temperature, and adjust the bedroom light brightness according to the sleep data of the bracelet, without the need for the user to manually operate each one.
Patent and data barriers have also been initially formed: Xiaomi currently holds 5,000 related patents covering smart home interaction, device-side reasoning, gesture control and other fields. The massive real scene data accumulated by 1.1 billion terminal devices has shortened the iteration cycle of new products by 50% compared with the industry average. The device-side lightweight large model deployment capability has also solved the past pain points of smart home AI functions relying on the network and high latency and poor privacy.
The newly released Miloco whole-house intelligent solution relies on the bottom layer of ThePaper OS to enable full-device collaboration, and is also connected to third-party devices from international brands such as Bosch, Siemens, and LG. Users can customize automation scenarios through a low-threshold graphical interface. It is currently the whole-house intelligent solution with the widest coverage of DIY users.
From the perspective of comparing the industry competition landscape, Xiaomi, Midea, Huawei and other leading players have taken a completely differentiated route, with equally prominent advantages and shortcomings.
The current paths of mainstream players are clearly differentiated: Midea relies on its own closed loop of all categories of home appliances, focusing on in-depth proactive intelligent prediction of four major life scenes; Huawei relies on PLC wired connection technology, focusing on systematic whole-house installation solutions in the pre-installation market; while Xiaomi relies on 1.1 billion devices
Its total equipment volume, cost-effective product matrix and open ecosystem cover entry-level smart home users and DIY groups who like to customize their gameplay. Its overall device scale exceeds Midea’s 780 million units and Huawei’s 450 million units, making it the world’s largest consumer-grade IoT platform.
However, Xiaomi’s disadvantages are also very obvious: the current depth of active intelligence still lags behind its leading competitors, most scenarios are still in the rule-driven automation stage, and the predictive active service capability needs to be strengthened; the layout of the pre-installation market lags behind, and the depth of cooperation with real estate developers and decoration companies is far less than that of Huawei; at the same time, it has been subject to patent lawsuits by traditional home appliance manufacturers such as Oaks and Bull in the past two years. Traditional players are using patent defense to block Xiaomi’s expansion path in the core category of home appliances.
From the perspective of investment market evaluation, industry organizations have obviously opposing views on the future prospects of Xiaomi smart home.
Optimistic institutions are represented by HSBC and China Renaissance, both of which have given a buy rating. They believe that Xiaomi’s billion-level hardware ecological entrance is an unreplicable core advantage in the AI era. The synergy of the entire ecosystem of people, cars, and homes will significantly increase user ARPU value. It is predicted that Xiaomi’s IoT business revenue can exceed 200 billion yuan in 2028.
Institutions that hold a cautious view, represented by BoCom International, maintain a neutral rating, believing that the current commercialization of Xiaomi AI is still in a very early stage, and the AI Token value-added service has just begun to contribute a small amount of revenue, and has not yet formed a stable second growth curve. At the same time, this round of memory chip price increases has a long cycle, component cost fluctuations may continue to erode the profits of the IoT business, and there is uncertainty in short-term profitability.
From an integrated perspective, Xiaomi Smart Home has now surpassed the stage of extensive growth that relies solely on low prices for scale. The scale resilience and global growth potential of the basic market have been verified by data. The 28% gross profit contribution is enough to make it a ballast stone to hedge against fluctuations in a single business cycle.
Next, there are actually only two core variables that determine the ceiling of Xiaomi’s smart home: First, whether Miloco’s whole-house smart solution can gradually make up for the shortcomings of active intelligence in scene depth, equal the experience gap between Midea and Huawei, and enter the more profitable mid-to-high-end whole-house smart market; second, whether the unique cross-ecological collaborative experience of “People, Cars, and Homes” can be implemented and create a differentiated monetization path that is completely different from traditional home appliance brands.
Judging from the current investment pace and user base, it has every opportunity to grow into a rule-maker in the global smart home market. However, whether it can ultimately transform the traffic advantage of 1.1 billion devices into sustained high gross profit revenue depends on the actual progress of AI implementation in the next 2-3 years.
