Rest of World published a compelling piece on the collapse of several Chinese EV startups, and there’s a glaring lesson for the entire industry: Car startups shouldn’t be developing their own software. They should rely on established software companies to build, license, and deploy that software.

Think about it: You shell out tens of thousands, maybe more, for a car—an investment you expect to last for a decade or longer. And for EVs, the software isn’t just a dashboard convenience, it’s central to the entire driving experience. From battery management to over-the-air updates and self-driving features, software makes or breaks the car.

Now, let’s imagine you buy a shiny new EV from a flashy startup. Three years later, that startup folds. What happens to your software updates? What happens to the core functionality of your vehicle if the startup disappears? Spoiler: you’re screwed. 

When an EV company collapses, it’s not just a question of no more updates or no more customer support. It’s much worse. Your vehicle could be left with outdated software that becomes incompatible with new systems, or worse, it could stop functioning altogether. This is especially true when startups decide to build their own custom software ecosystems from the ground up. It sounds like a smart idea to stand out in a crowded market, but it’s more like building a sandcastle next to the ocean—it’s not going to last, and it’s the customers who end up with sand on their shoes.

Take Byton, for example. They were around for four years, from 2017 to 2021, with big dreams of luxury electric SUVs featuring fancy tech. It was a Tencent – Foxconn joint venture, but all their resources couldn’t save them from the software pit they dug themselves into. They poured talent, effort, and money into creating a massive dashboard screen with a custom UI, promising AI-driven features. And where did it get them? Bankruptcy.

Or consider Bordrin Motors, another four-year wonder from 2016 to 2020. They developed their own vehicle operating system and digital cockpit platform. Sounds cool, right? Well, it would be if they hadn’t run out of money trying to maintain it all.

Established software companies like Apple, Google, and Microsoft have been making software for decades. They know what it takes to keep an ecosystem alive, stable, and, more importantly, secure. The idea of rolling your own software is not new—remember when every gadget maker wanted to make their own OS? It was a disaster then, and it’s a disaster now. Why should carmakers fare any better?

Instead, EV startups should focus on what they’re supposed to be good at: building great electric cars. Let the software experts handle the software. Tesla, for all its faults, is still in the game partly because of its strong software focus. They’ve managed to build a robust platform that, so far, has stood the test of time. But here’s the catch: not every startup can be Tesla, nor should they try to be. 

Some Chinese EV companies are getting it right. Look at Xpeng Motors and Li Auto. These guys are smart. They’re doing a mix of in-house development and licensing from established tech providers. Xpeng partnered with NVIDIA for AI computing and works with Desay SV Automotive for some software components. Li Auto isn’t too proud to license components for specific functionalities. And guess what? They’re still in business!

Licensing software from a more established player means that even if your car startup fails, your customers aren’t stuck with an expensive, bricked paperweight in their driveway. Their car can still receive updates, still work, and they aren’t left holding the bag. It’s akin to separating hardware from software in the tech world: Apple doesn’t build its processors, TSMC does. Apple doesn’t make its screens; Samsung does. Division of labor works for a reason.

Apple is an example of a tech company that wants to do the whole widget and they mostly do these days, but still not everything. They design their hardware but they don’t build them. The manufacturing and assembly go to partners like Foxconn and Pegatron. They didn’t design their own processors until they have the resources to put together the teams for it. And in their early days they didn’t even design their own products. Apple hired frogdesign (now just Frog) to design their computers and come up with a design language to be followed by the company’s lines of products so they all have the same style.

The problem is that too many startups have founders who think they can do it all. They want to control every aspect of the experience, which is admirable until reality sets in. Building a car is hard enough. Making great software is equally hard. Trying to do both? It’s a fool’s errand. And who pays the price for that arrogance? The consumers.

It’s one thing to purchase a phone or computer and no longer receiving updates or support after 3-4 years but when it’s a car that costs tens of thousands of dollars, you damn sure want to be able to use it for more than just a few years or at least sell it at a decent price when you need to.

No startup founder builds their company expecting to fail, so of course they will spend resources to do everything. However, when a company is just starting up the leaders need to be able to determine what their areas of strengths are, what sort of resources can they pull, and what factors can or should be outsourced to leverage outside expertise and increase internal efficiency. Once a company is strong enough to maintain a solid core and grow a business from there, then they can begin to consider building or developing non core elements internally. Of course, they also need to be able to identify what their actual core strengths are, lest they focus on the wrong things and end up accelerating their own collapse.

The key takeaway here is simple: EV startups need to know their limits. No matter how much venture capital you have or how many big names are on your board, you’re not a software company just because you hire a few software engineers. You’re a car company, so act like one. Leave the software to those who know it best. Because in the end, if you go under, it’s the customers who will feel the real crash.