Automotive revenue now makes up 20% of Xiaomi's total, with MiMo, boasting 10.5 trillion Tokens, also starting to generate revenue.

On August 18, Xiaomi unveiled its latest financial results, revealing a revenue of 108.92 billion yuan for the second quarter of 2026, marking a 6.1% year-on-year decrease. Gross profit fell by 17.2% to 21.61 billion yuan, while adjusted net profit dropped by 42.6% to 6.22 billion yuan. The primary pressure on performance stemmed from Xiaomi's traditional revenue pillars: smartphones, IoT, and internet services. The 'Smartphone × AIoT' segment contributed 84 billion yuan in revenue, down 11.3% year-on-year. Meanwhile, revenue from smart electric vehicles, AI, and other innovative businesses surged by 17.1% to 24.9 billion yuan, accounting for roughly 23% of the group's total revenue.

Image Source: Xiaomi Official Website

Currently, innovative businesses' revenue is predominantly driven by automotive sales, with automotive revenue reaching 23.9 billion yuan in the second quarter. Notably, the MiMo-related AI business was explicitly listed as a revenue source for the first time this quarter. Among the approximately 1 billion yuan in revenue from other related businesses, Xiaomi included AI business tied to the Xiaomi MiMo large model series as a growth driver. During the earnings call, Lin Shiwei, Xiaomi's Vice President and CFO, further disclosed that MiMo's API calls and Token Plan have commenced generating revenue.

Investment in this area continues to rise. The second-quarter financial report indicated that Xiaomi's R&D expenses hit 9.2 billion yuan, up 18.9% year-on-year, with the first-half total reaching 18.2 billion yuan. Lin Shiwei noted during the earnings call that nearly 30% of this investment was AI-related. Previously, Xiaomi announced its expectation to invest over 60 billion yuan in AI over the next three years.

Image Source: Xiaomi Official Website

On one hand, revenue from 'Smartphone × AIoT' is declining; on the other, automotive sales are soaring, and MiMo is beginning to contribute revenue. Xiaomi's new growth structure is also starting to extend from automotive further into AI.

01 Core Business Under Pressure, Automotive Drives Incremental Growth

Smartphones remain Xiaomi's largest single revenue stream.

The financial report revealed that in the second quarter, Xiaomi's smartphone revenue was 42.1 billion yuan, down 7.5% year-on-year, with global shipments at 31.2 million units, a 26.5% decrease. After adjusting the product mix, the average selling price of smartphones rose by 25.9% year-on-year to 1,351 yuan, a record high.

This adjustment is not unique to Xiaomi. IDC data shows that smartphone shipments in China decreased by 4.3% year-on-year, marking the fifth consecutive quarter of decline, with most Android manufacturers raising prices or reducing low-end model supply. Omdia data indicates that Xiaomi's shipments dropped by 26% in the second quarter of 2026, with over half of the shipments priced below $200. Low-priced models have thinner profit margins, and users are more price-sensitive. Therefore, Omdia believes that Xiaomi is the most significantly affected among the top five global manufacturers due to rising storage costs. In the second quarter, Xiaomi's smartphone gross margin decreased from 11.5% to 8.5%, with rising prices of core components like storage cited as a primary reason.

IoT and lifestyle consumer performance also declined due to reduced domestic subsidies. This business encompasses smart home appliances, TVs, tablets, wearables, and other IoT hardware. Revenue decreased from 38.7 billion yuan in the same period last year to 31.3 billion yuan, a 19.2% year-on-year drop. Internet services remained relatively stable, with revenue of approximately 9 billion yuan, roughly flat year-on-year, and the gross margin increased from 75.4% to 76.8%.

Combining these three businesses, revenue from 'Smartphone × AIoT' decreased from 94.7 billion yuan in the same period last year to 84 billion yuan.

While the core business faces pressure, automotive sales are driving incremental growth. In the second quarter, Xiaomi delivered 104,200 vehicles, up 28.2% year-on-year; automotive revenue was 23.9 billion yuan, a 15.9% increase, accounting for over 20% of the group's quarterly revenue. Moreover, automotive sales continue to scale. Xiaomi set a delivery target of 550,000 vehicles for this year. Based on the first- and second-quarter financial reports, approximately 185,000 vehicles were delivered in the first half. The second-quarter financial report also disclosed that the first extended-range SUV, the 'Pengcheng' series, will launch in September, further expanding the product lineup.

Image Source: Jiemian News | Screenshot from Xiaomi's product launch livestream

However, the revenue growth rate of automotive sales did not keep pace with the delivery growth rate. The financial report showed that the average revenue per vehicle decreased from 253,700 yuan to 229,300 yuan in the second quarter, which Xiaomi mainly attributed to the decrease in the proportion of SU7 Ultra deliveries. The gross margin of smart electric vehicles, AI, and other innovative businesses also decreased from 26.4% to 19.2%, with operating expenses increasing from 5.9 billion yuan to 7.4 billion yuan, resulting in an expanded operating loss of 2.6 billion yuan.

While automotive sales have already generated quarterly revenue in the hundreds of billions of yuan, the sector is still in the investment and profit adjustment phase.

02 Xiaomi's AI Strategy Converges Around MiMo

Nearly 30% of Xiaomi's R&D investment in the first half of the year flowed into AI, and MiMo also appeared in the financial report revenue for the first time. AI-related investments did not commence this year.

After Xiaomi released HyperOS in 2023, its smartphone system began incorporating large model capabilities, and smart home devices introduced HyperMind. Subsequently, Xiaomi gradually applied AI to intelligent driving, smart cockpits, and home device coordination. In 2024, Xiaomi began increasing investment in AI infrastructure and advanced the development of base large models and multimodal technologies. By 2025, the first MiMo open-source model, MiMo 7B, was released, followed by continued expansion of multimodal, voice, and Agent capabilities.

This year, these changes are further reflected in the organizational structure. In July 2026, reports indicated that Xiaomi reorganized the AI technical divisions behind its AI assistant 'Xiao Ai.' The base model capabilities are now handled by the MiMo team, the cloud team is responsible for model engineering, and the OS teams for smartphones and other businesses handle on-device implementation. This division of labor aligns with Xiaomi's positioning of MiMo during the quarterly earnings call. Management stated that MiMo is at the base model layer of the company's AI system and will continue to integrate with businesses such as smartphone operating systems, home devices, intelligent driving, and robots.

Xiaomi's AI capabilities, previously deployed across different businesses, are now forming a unified underlying model around MiMo.

03 Two Paths for MiMo

After becoming the base model, MiMo first accelerated its external openness.

Externally, Xiaomi started with developers. In April this year, MiMo V2.5 entered public beta testing, and Xiaomi subsequently launched the MiMo Orbit plan, collaborating with multiple Agent frameworks. In May, the API prices for the MiMo V2.5 series were further reduced, and the available quantity of the Token Plan was increased. In June, the AI programming tool MiMo Code was launched.

Image Source: Official Xiaomi MiMo Account

The call volume quickly followed. In the second-quarter financial report, Xiaomi cited OpenRouter data, stating that from July 20 to 26, the Token usage of MiMo V2.5 reached 10.5 trillion, ranking first on the platform for that week. Management believes that the growth in calls is mainly due to the model's performance in perception, Agent, and coding capabilities, as well as its low call prices and open-source strategy. Compared to call volume, the financial report provided a more direct signal of commercialization. In the second quarter, revenue from other related businesses, which increased by 56.5% year-on-year to approximately 1 billion yuan, already included MiMo-related AI businesses as a growth source. Lin Shiwei revealed during the earnings call that the current focus is still on expanding the user base and continuing to iterate the base model.

MiMo's other path leads into Xiaomi's own products. The second-quarter financial report disclosed that Miloco 2.0 is centered around MiMo, and the new generation AI assistant, Super Xiao Ai 2.0, in HyperOS 4 Beta, also integrates MiMo, adding cross-application and cross-device task processing capabilities. Management listed intelligent driving and robots as directions for continued integration in the future.

This path corresponds to Xiaomi's established 'Human-Vehicle-Home' ecosystem. The financial report showed that as of the end of June, the number of connected devices on Xiaomi's AIoT platform reached 1.161 billion, a 17.4% year-on-year increase; the monthly active users of the Mi Home App reached 124 million, and the monthly active users of Xiao Ai reached 175 million. In terms of automotive sales, 104,200 vehicles were delivered in the second quarter, generating 23.9 billion yuan in revenue.

Thus, Xiaomi's AI strategy has formed two clear paths: externally, MiMo enters the developer ecosystem through APIs, Token Plans, and Agent tools; internally, it integrates into the 'Human-Vehicle-Home' ecosystem.

However, scale does not equate to revenue. While MiMo has started contributing revenue, it is still in the early stages of commercialization. The second-quarter financial report showed that smart electric vehicles, AI, and other innovative businesses incurred an operating loss of 2.6 billion yuan in the quarter, with AI-related costs being one of the reasons for the decline in the gross margin of the innovative business segment.

Going forward, whether MiMo can convert call volume and the scale of the 'Human-Vehicle-Home' ecosystem into stable profit contributions in the financial report will determine whether AI can truly become a new growth driver for Xiaomi.

(The images and public data sources in this article are from the internet and public industry platforms. If there is any infringement, please contact us for timely modification.)