Xiaomi Car Achieves First Annual Profit Turnaround, Lu Weibing Discusses Memory Pressure: Will Raise Prices If Can't Sustain

Question AI · How does Xiaomi delay price hikes through diversified strategies amid rising memory prices?

On March 24, Xiaomi Group (HK01810) released its full-year 2025 financial results. Last year, Xiaomi’s total revenue reached 457.3 billion yuan, a 25.0% increase year-over-year; adjusted net profit was 39.2 billion yuan, up 43.8%, both hitting record highs. Among them, revenue from innovative businesses such as smart electric vehicles and AI reached 106.1 billion yuan, a 223.8% increase, turning profitable for the first time with an operating profit of 900 million yuan. In the fourth quarter of last year, this segment accounted for 31.8% of total revenue, approaching the 37.9% share from smartphones during the same period.

During the earnings call, Xiaomi Group Partner and President Lu Weibing responded to topics including rising phone memory prices and Xiaomi’s AI deployment. Lu Weibing said Xiaomi faces significant pressure to raise prices, and the trend is unavoidable. Xiaomi tries to bear some of the cost for consumers for a while; if it becomes unsustainable, prices will go up—just more than competitors can hold out. Regarding AI commercialization, he stated it is still in early stages, and even the team has no KPIs yet.

Xiaomi Cars Turned Profitable for the First Time, Achieving 900 Million Yuan Operating Profit

The financial report shows that in 2025, Xiaomi’s total revenue was 457.3 billion yuan, up 25.0%; adjusted net profit was 39.2 billion yuan, up 43.8%, both setting new records.

Looking at segments, revenue from innovative businesses like smart electric vehicles and AI reached 106.1 billion yuan last year, a 223.8% increase. This segment turned profitable for the first time with an operating profit of 900 million yuan. In Q4 2025, revenue from this division rose to 37.2 billion yuan, a growth of over 123%, accounting for 31.8% of total revenue, gradually approaching the 37.9% share of smartphones during the same period.

In the smartphone and AIoT sectors, last year’s revenue was 351.2 billion yuan, a 5.4% increase. Smartphone revenue for the full year was 186.4 billion yuan, a 2.8% decrease, mainly due to reduced shipment volume and lower average selling price (ASP). In Q4 2025, smartphone revenue was 44.3 billion yuan, down 13.6% year-over-year, mainly due to decreased shipment volume and ASP.

Recently, the topic of “full-scale price hikes for phones” has attracted much attention, with some brands raising prices. During the earnings call, Lu Weibing said the actual increase in storage prices this round is more aggressive than initially expected, affecting not only the phone industry. He understands that competitors are adjusting prices due to memory cost pressures, and believes price hikes are inevitable. Xiaomi also faces significant pressure in this regard. “The trend of price increases is unavoidable. We try to bear it for consumers as long as possible; if we can’t, prices will go up. We hope users can understand that we’ve just held out longer than others.”

Lu Weibing also mentioned that Xiaomi has some advantages in “withstanding price pressure.” First, Xiaomi’s product categories are diverse, with all major segments progressing simultaneously, allowing better coordination and resolution of memory issues. Second, Xiaomi’s demand for memory is substantial, and over the past few years, it has established trust-based long-term supply agreements with the top global memory suppliers, eliminating stock shortage risks and providing pricing advantages. Third, Xiaomi’s past forecasts of rising memory costs were pessimistic, leading to more aggressive stocking, so current inventory levels are sufficient.

Future AI Investment Will Exceed 60 Billion Yuan in the Next Three Years, Lu Weibing: AI Commercialization Still in Early Stages

In terms of R&D, the financial report shows that Xiaomi’s R&D investment in 2025 was 33.1 billion yuan, nearly matching annual profit, a 37.8% increase. The total number of R&D personnel reached 25,457, a 20.1% rise. According to the report, over the past five years (2021-2025), Xiaomi has invested a total of 105.5 billion yuan in R&D. Recently, Xiaomi announced that R&D investment will be increased to 200 billion yuan over the next five years, with at least 60 billion yuan allocated to AI in the next three years.

Lu Weibing also discussed during the earnings call that after the launch of the “Lobster” product, Xiaomi quickly deployed “Lobster” in smartphones, receiving positive feedback. He believes the larger market will be AI moving from virtual to physical worlds, with autonomous driving and humanoid robots as the ultimate carriers. Based on this idea, Xiaomi has made strategic layouts, aiming for AI to serve the entire ecosystem of people, vehicles, and homes. He emphasized that AI is still in early stages, and Xiaomi currently has no specific commercial plans or KPIs for AI. “We only set KPIs when the business matures; right now, it’s a very fast-changing period.”

Regarding model development, Xiaomi recently launched three self-developed large models designed for the Agent era: the flagship base model Xiaomi MiMo-V2-Pro, the multimodal model Xiaomi MiMo-V2-Omni, and the speech model Xiaomi MiMo-V2-TTS. This marks a significant step toward AGI (Artificial General Intelligence). Lei Jun stated at the launch that Xiaomi is relatively low-profile in AI, and progress may be much faster than what’s visible.

Large AI models are accelerating Xiaomi’s “full-ecosystem” deployment across people, vehicles, and homes. In March, Xiaomi officially released the mobile intelligent agent Xiaomi miclaw, initiating a closed beta that sparked widespread discussion. In February, Xiaomi showcased its future smart home exploration plan Xiaomi Miloco at MWC, which is the first real-world application of the Xiaomi MiMo large model, bringing AI from screens into daily life.

In embodied intelligence, Xiaomi has disclosed several key achievements: the phased research result TacRefineNet, and the open-source robot VLA large model Xiaomi-Robotics-0. In March, Xiaomi’s humanoid robot was officially “interning” at Xiaomi’s automotive factory, autonomously operating for three hours on screw fastening stations, with a 90.2% success rate for dual-side installation, meeting the fastest 76-second production cycle, marking a key step toward large-scale application.

Reporting: Nandu N Video Reporter Lin Wenqi