Moats aren't designed
Originally written in Chinese. Translated with AI, reviewed by Gary. Read the Chinese original →
I've been watching Meituan a lot lately.
At its 2021 peak, Meituan's market value was close to HK$2.7 trillion. Today it has fallen about 70% from there. There are several reasons behind it, including regulation, a market-wide valuation reset and this round of the food delivery war. But it made me rethink an idea the startup world has been repeating for more than twenty years:
Asset-light.
I used to believe asset-light was the best model for a startup.
Uber has no cars, Airbnb has no rooms, Grab has no cars, Meituan has no restaurants.
Everyone said: build a platform, connect supply and demand, and you can scale fast without investing much in assets, with high capital efficiency. That was the story VCs loved to hear for the past decade and more.
I also used to believe a platform would always be worth more than infrastructure.
But watching Meituan recently, I've started to see it a little differently.
The companies once called too heavy
In this delivery war, Meituan is up against JD and Alibaba.
The interesting part is that both were long considered too heavy.
JD spent more than twenty years building warehouses, logistics and supply chain. Many people thought Richard Liu had it wrong. Everyone else was building platforms; why did he keep doing these slow, expensive things?
Looking back today, those investments that looked so clumsy became JD's biggest advantage.
Because these things can't be copied tomorrow just because you decide to do them today.
Alibaba is the same.
Over the years it kept investing in Cainiao logistics, payments, cloud computing and local services. It all looked expensive, but in the end every piece became part of its ecosystem.
That's when it hit me: maybe the real question was never asset-light versus asset-heavy.
It's what value you have created that nobody else can copy.
Not building a moat for the sake of a moat
In startups we used to talk about growth, fundraising, valuation and burn rate. Many people would even say, quite proudly, "We're spending OPM, other people's money."
I like that line less and less.
The real goal of a startup shouldn't be spending other people's money.
Nor designing a business model just to raise money, and certainly not building a moat for the sake of having a moat.
A moat was never the goal. It is only a result.
When you keep creating value, keep solving users' problems, and keep investing in what others won't invest in and won't stick with, over time a moat appears on its own.
JD didn't build warehouses in order to have a moat.
If on day one Richard Liu had told his team, "Let's build warehouses, because one day they'll become a moat," I don't think anyone would have bought it.
He built warehouses because he believed it would get goods to customers faster, make the supply chain more efficient, and solve the biggest pain point in Chinese e-commerce at the time.
Only later did that value slowly turn into a moat.
That is also the biggest lesson I take from Meituan's war.
For years everyone believed platforms were the most powerful business model.
But today the market is repricing.
Capital markets no longer look only at growth.
They've started asking another question:
If someone came at you with unlimited money today, what would you have left?
If the answer is only subsidies, your advantage may not be as deep as you think.
If the answer is a set of capabilities nobody can copy in ten years, that's real value.
A startup shouldn't begin by thinking about fundraising, or valuation, or a moat.
Create value first, and do it seriously.
Because a real moat is never designed.
It's the result of doing the right things for a long time and creating value consistently, until others simply can't catch up.