A four-month-old company in Beijing just shipped its first product, and it is not a phone. It is a ROM. MIUI, a reworked build of Android you flash onto somebody else’s hardware, turning a Samsung or an HTC into a nicer phone than the one Samsung or HTC actually sold you. Xiaomi was founded in April by Lei Jun and six others, and this week, it released software for devices it does not make. I find that backward in the best possible way. How do I find it? Because I’m monitoring all the job offers that call for Windows experience, and this company, from their space at the university, posted a bunch of them. Plus, I can give it a try as my phone is “compatible”.
Lei Jun is not a kid in a dorm doing this for fun. He ran Kingsoft; he was behind Joyo.com, which Amazon bought for $75 million in 2004, and he has spent years as an angel investor in Chinese tech. So when he looks at the Android phones shipping in China right now, a mess of half-finished manufacturer skins sitting on hardware that never sees another update, he sees a software problem nobody is bothering to fix. MIUI is him fixing the software layer first and worrying about hardware later.
What MIUI actually is: a heavily customized take on Android 2.2, lifting a lot of its visual language straight from iOS, with a real theming engine, finer permission controls, and performance tuning aimed at the mid-range hardware most Chinese buyers actually own. A team of around thirty engineers builds it. The part I keep circling back to is the cadence, because there is a fresh build every Friday, shaped by a forum full of users who file the bugs and ask for the features that go into the next one. Something like a hundred thousand people are already flashing it. That weekly loop is the entire marketing budget, and it costs almost nothing.
Here is where I get skeptical. An Android skin is not a moat. Anyone can fork AOSP and ship a prettier launcher (I mean, even I did it in my free time), and the graveyard of custom ROMs is full of projects with rabid fans and no business attached. CyanogenMod has a devoted following and zero revenue. The difference Xiaomi is betting on is that the community itself becomes the moat: spend a year turning a hundred thousand enthusiasts into evangelists (I won’t lie, even I’m looking at the company quite closely for the past year too, hence this post), and you have a launch-day crowd for whatever phone you eventually build. I am not sure that converts. Loving a free ROM and paying for hardware are very different acts.
I am not wading into the “MIUI is just an iOS ripoff” argument, because that one is going to trail this company for its entire life, and it bores me already. The question that actually matters is whether a software-first outfit can out-execute Huawei, ZTE, Lenovo, and the rest of the local Android crowd on hardware it has never once manufactured. China’s phone market is fragmented and chaotic, Samsung is running away with the premium end, and a startup whose only real asset today is a Friday software build and a busy forum is a genuinely strange thing to put money behind.
The normal way to do this is to build a phone and bolt the software on at the end. Xiaomi turned it inside out: build the software, gather the audience, then go find a factory. If a phone ever ships and it is even halfway decent, this company starts from a place none of its rivals occupy, with a crowd that already picked it. Whether that crowd opens its wallet is something I cannot say yet. Ask me when there is something to actually buy.