In 2025, Xiaomi delivered 410,000 electric vehicles, connected 1.08 billion IoT devices, and posted 457.3B RMB in revenue — growing 25% in a year when global consumer electronics was largely flat. Lei Jun is doing something the industrial world rarely produces: building a single brand that credibly spans smartphones, smart home, and automobiles. The hardware flywheel is already spinning. The real bet is whether AI turns device ownership into cognitive dependency.
I. Decoding the Business DNA
Xiaomi's business logic rests on a single foundational assumption: if users trust you to deliver best-in-class value across every product category, they'll let you manage their entire device ecosystem.
This assumption is the load-bearing pillar of the "Human-Car-Home" (人车家) ecosystem strategy. The smartphone is the entry point. The home is the scenario. The car is the moving room. Xiao Ai (小爱同学) is the voice hub. And the web of connected devices generates a data flywheel that compounds with each new product.
But the strategy carries an internal tension: value pricing and premium brand positioning are fundamentally at odds.
In 2025, Xiaomi sold EVs at an average of 250,000 RMB and flagship phones priced above 4,000 RMB — evidence that the premium ceiling is being broken. China Mainland market share in the 3,000+ RMB segment rose from 23.3% in 2024 to 27.1% in 2025. But internationally, the "affordable but not premium" label remains the primary ceiling on brand expansion.
Customer Segmentation: Three Segments, Three Relationships
Value-performance buyers (core base): They trust Xiaomi, so they buy Xiaomi phones, then Xiaomi ACs, speakers, and cameras. Their logic is reducing decision fatigue — not researching every product category. They are the engine of AIoT platform depth. The 22.7M users with 5+ connected devices come almost entirely from this segment.
Premium upgrade buyers (growth engine): Moving from mid-range Android to Xiaomi flagship, or from ICE vehicles to an SU7. They care about performance, design, and brand credibility — but are willing to make a calculated bet on a challenger. Most SU7 buyers fit this profile.
Internet services users (profit layer): Xiaomi phone users who subscribe to memberships, use Xiaomi Finance, or engage with Xiaomi gaming. At 76.5% gross margin, internet services is the cleanest profit in the entire model — but it depends entirely on hardware user scale.
II. How the Money Works
Business Snapshot
Metric Value 2025 Full-Year Revenue 457.3B RMB (+25% YoY) Adjusted Net Profit 39.2B RMB (+43.8% YoY) Smartphone Revenue 186.4B RMB IoT and Lifestyle Products 123.2B RMB (+18.3% YoY) Internet Services 37.4B RMB (76.5% gross margin) EV and AI Innovation 106.1B RMB (+223.8% YoY) AIoT Connected Devices 1.08B R&D Spend 33.1B RMB (+37.8% YoY) [Source: Xiaomi 2025 Full-Year Results, March 24, 2026]
Revenue Structure: Four Layers, the Fourth Just Launched
Layer 1: Smartphones (ecosystem entry point)
186.4B RMB in 2025, global top 3 for five consecutive years. The phone's strategic function goes beyond revenue: every Xiaomi handset is a potential AIoT node, an internet services subscriber, and a future EV customer. Smartphone gross margin runs around 12% — a deliberate pricing choice to trade margin for ecosystem entry.
Layer 2: IoT (moat builder)
123.2B RMB, 23.1% gross margin, improving 2.8 percentage points year-over-year. With 1.08B connected devices, Xiaomi has entered roughly one in every seven households globally. The strategic value isn't the revenue — it's the data on energy use patterns, daily routines, and consumption preferences that can increasingly be activated by AI.
Layer 3: Internet Services (profit engine)
37.4B RMB at 76.5% gross margin. This is the cleanest profit line in the model. Advertising, memberships, finance, and gaming all sit on top of the hardware user base. The concern: internet services growth (9.7%) is running well below overall revenue growth (25%), suggesting monetization efficiency per user is compressing at the margin.
Layer 4: EV and AI Innovation (flywheel accelerator)
106.1B RMB, 410,000 deliveries, average selling price 250,000 RMB — growing 223.8% year-over-year. The EV business isn't just a revenue line. It redefines the physical boundary of the ecosystem. Inside an SU7, there are Xiaomi custom chips, Xiao Ai integration, and real-time synchronization with home IoT devices. SU7 buyers are Xiaomi's highest-lifetime-value customers — they stay in the ecosystem longer and buy more connected devices.
III. The Flywheel and the Moat
The Flywheel: Hardware → Data → AI → Better Experience → More Hardware
Xiaomi's flywheel shares structural DNA with Apple's: more devices → more data → better AI → better product experience → harder to leave. But there is a critical difference in execution depth.
Apple's flywheel is driven primarily by high-frequency, high-engagement devices — iPhone, Watch, AirPods. Xiaomi's flywheel is numerically larger (1.08B devices) but includes substantial volume from low-frequency appliances: air conditioners, refrigerators, washing machines. These generate data but don't produce the daily engagement loop that keeps the flywheel spinning fast.
The real flywheel drivers are high-frequency touchpoints: phones, cars, and wearables. As long as these three categories keep growing, the data accumulation compounds. The moment the car is added as a daily-use device, the flywheel gains its most powerful accelerant.
Three Moats, Different Depths
Moat 1: Hardware engineering at price. Xiaomi consistently delivers near-flagship performance at prices below established competitors. The SU7 matched or exceeded premium EVs at 200,000–250,000 RMB — a feat the market didn't expect from a phone company. The in-house Xuanyuan chip is deepening this moat by reducing dependency on Qualcomm and capturing more margin.
Moat 2: AIoT ecosystem lock-in. The 22.7M users with 5+ devices face an enormous switching cost — they'd have to repurchase and reconfigure an entire device ecosystem. This moat is real, but currently applies to only about 2% of the 1.08B connected devices. The gap between 22.7M deep users and 1.08B total devices is the scale of the unconverted opportunity.
Moat 3: Brand positioning (market-specific). In China, the premium pivot is working. In the US and EU, the value-brand ceiling remains intact. The SU7's appearance at European auto shows is a deliberate campaign to rewrite the global brand narrative — but this takes years, not quarters.
IV. Risks and Cracks
Risk 1: R&D Spend Is Approaching the Size of Annual Profits
33.1B RMB in 2025 R&D spend, against 39.2B in adjusted net profit. The commitment to spend 200B+ on R&D over the next five years and 60B specifically on AI in three years puts the profit buffer under structural pressure. If EV unit economics don't reach breakeven within 2-3 years, the cumulative losses will begin eroding the financial cushion that makes the hardware-thin-margin strategy sustainable.
Risk 2: The Smartphone Market Has a Clear Ceiling
Global smartphone shipments run around 1.2B units per year. Xiaomi holds roughly 15% share. The market is in volume saturation. High-end ASP growth can expand smartphone revenue, but volume growth has limits. The ecosystem entry-point function of the phone weakens as global smartphone penetration approaches saturation.
Risk 3: Supply Chain Concentration
Xiaomi phones depend heavily on TSMC (chips), Samsung and LG (displays), and Qualcomm (processors). The Xuanyuan in-house chip is a partial hedge, but full supply chain autonomy is still years away. Under accelerating US-China technology decoupling, these dependencies carry rising tail risk.
Risk 4: EV Brand Concentration on One Model
The entire EV business runs on SU7 (and SU7 Ultra). A single vehicle's market success does not constitute an automotive brand. The upcoming SUV YU7 is the first real test of whether Xiaomi can replicate the SU7 story. Any product failure or major safety recall would inflict systemic damage on a brand equity account still being built.
V. The Endgame
Scale Returns Require the AI Layer to Activate
Xiaomi's theoretical endgame is operating as the "AI operating system for households" — from doorbell to dashboard, all devices running on Xiaomi's software, Xiao Ai as the interface, and a data flywheel that personalizes every recommendation and automates every routine.
In China, this is partially true today. 1.08B connected devices is not a vanity number. But the endgame requires AI to make the experience genuinely better — not marginally better, but "I can't go back" better. If Xiao Ai evolves into a true AI agent that predicts needs and makes decisions autonomously, the ecosystem moat upgrades from "device binding" to "cognitive binding." That's a different class of competitive protection entirely.
Internationally, replicating the ecosystem requires overcoming supply chain headwinds, brand perception gaps, and regulatory barriers simultaneously. That's a multi-decade project.
Stage Assessment: Second Growth Curve Just Launched
The 2025 version of Xiaomi is in a rare position: mature businesses (smartphones + IoT + internet services) generating steady cash flow, while a new business (EVs) is burning cash on the ramp but already at visible scale.
The advantage: there's a financial buffer between old and new growth. The risk: if EV unit economics don't bend toward breakeven in 24-36 months, accumulated losses will start eroding the margin structure of the profitable core.
VI. The Verdict
Xiaomi executed one of the most operationally difficult years any consumer tech company has managed — simultaneously growing smartphones, IoT, internet services, and scaling an EV business to 410,000 deliveries. Execution credit is real and deserved.
The moat, however, is still under construction.
The smartphone moat is value-brand positioning, which has a premium ceiling. The IoT moat is genuine device lock-in, but only for 22.7M of 1.08B connected users. The EV moat is entirely provisional — one good product launch, not a durable competitive position.
The single variable that determines Xiaomi's long-term value is whether AI transforms 1.08B connected devices into a "cognitive dependency" ecosystem. If Xiao Ai becomes a truly useful AI agent in 2-3 years, the device network becomes a structural moat that compounds. If the AI layer remains undifferentiated, Xiaomi is a world-class hardware manufacturer whose moat depth depends on winning the next product cycle — respectable, but not irreplaceable.
The hardware flywheel is real. The AI bet is the one that makes it irreplaceable.
This analysis is based on public financial disclosures, investor relations materials, and industry research. It does not constitute investment advice.