Dictated my thoughts to AI. TLDR: Xiaomi will continue to decline. But two years from now, might be a great buy.
The Decline of Xiaomi
I'm getting more excited about Xiaomi as an investment precisely because I think the next 2-3 years could be ugly. Xiaomi is trying to be both Apple and Tesla at once, and right now both sides of that bet are under pressure.
The Apple side: Smartphones
Xiaomi is still the world's #3 smartphone maker, shipping 165.2 million phones in 2025 for a 13.3% global share. But its position is slipping at the worst possible time. In Q2 2026, its shipments fell 26% YoY to 31.2 million while the overall market fell 7.4%. IDC has its global share dropping from 14.2% to 11.3%.
Memory is a big part of the story. More than half of Xiaomi's phones sell for under $200, and that end of the market gets hit hardest when memory prices spike. IDC says memory costs were up nearly 300% YoY by Q2 2026 and made up more than 65% of the bill of materials for low-end phones. Counterpoint estimates smartphone memory prices rose over 80% QoQ in Q2 alone.
The damage is already showing. Q2 smartphone revenue fell 7.5%, and smartphone gross margin dropped from 11.5% to 8.5%.
Samsung has a built-in hedge here. It makes phones and memory, so when memory prices rise it loses on one side and gains on the other. Xiaomi has no such offset. Apple, meanwhile, sells almost entirely at the premium end with real pricing power. Its Q2 shipments grew 3% while Xiaomi's fell 26%.
Xiaomi's answer is to become more like Apple by designing its own chips. Lei Jun has committed at least RMB50 billion (about US$6.9 billion) over 10 years, with a new chip every year. The first, the XRING O1, is built on TSMC's 3nm process and has shipped over one million units. The new XRING O3, unveiled in August, is the first mobile chip to support LPDDR6 memory. Xiaomi still relies on MediaTek for 5G modems, so it isn't fully independent yet.
So Xiaomi phone hasn't become a bad division. It's just sitting in a bad part of the cycle.
The Tesla side: Cars
2025 was a breakout year. Xiaomi delivered 411,082 EVs, up more than 200%, with EV revenue of RMB103.3 billion (up 222%) and RMB0.9 billion in operating income.
2026 has been much harder. The EV and new initiatives segment's gross margin fell to 19.2% from 26.4% a year earlier, pushing it back into the red with a Q2 operating loss of RMB2.6 billion, after RMB3.1 billion in Q1. Volume is still growing: Q2 deliveries rose 28.2% to 104,199. The problem is price, not demand.
China's EV market has turned into a brutal price war.
BYD, Geely, NIO, Xiaomi and many others keep cutting prices and offering cheap financing. Xiaomi itself ran a promotion on the YU7 with three years of zero-interest financing. Beijing has stepped in repeatedly, calling for an end to "involution," and the market regulator warned that carmakers selling below cost to push out rivals face significant legal risks. The price cuts have kept coming anyway.
The deeper issue is that Chinese companies are all very good at building hardware. When everyone can make a good car cheaply, there's little to set anyone apart and far too much supply. AlixPartners expects only around a dozen Chinese carmakers to remain financially viable by the end of the decade. This is a last-man-standing game, and it could take 3-5 years to play out.
The Ecosystem: What ties it together
What keeps me interested is the Human × Car × Home strategy. Xiaomi isn't selling a phone, a TV, a washing machine and a car as separate products. It's building one ecosystem around all of them, and it already has 754 million monthly active users and more than 1.08 billion connected IoT devices. That's an enormous base to build on.
My view
The phone business will struggle while memory stays expensive, and the EV business, though growing fast, faces consolidation and margin pressure. I'm confident in Xiaomi's ability to build the chip, software and ecosystem layer. I'm far less sure about the long-term economics of making cars in China.
So I'm happy to wait. If the stock falls another 30-40% over the next couple of years while the ecosystem keeps getting stronger, I'd be very interested. That's the setup I want: 165M+ phones a year, 750M+ users, a billion-plus devices, its own chips and a fast-growing EV business, bought after the industry's problems have already crushed the valuation.