Tesla Says Millions of Older Cars Need Hardware Updates for Self-Driving

Tesla CEO Elon Musk said this week that millions of Tesla owners will have to upgrade the computers inside their cars if they want the company's next version of Full Self-Driving software. This is a big change from what Tesla told customers before: that their current hardware would be good enough.

The cars affected are those sold between 2019 and 2023, which use what Tesla calls Hardware 3. TechCrunch reports that Musk said Tesla would open small factories in major cities to swap out these older computers. The company made $1.4 billion in free cash flow last quarter and is planning to spend $25 billion on capital projects this year—money that likely includes these upgrade factories.

Why the Older Computers Aren't Enough

Tesla's Hardware 3, which launched in 2019, was built to handle self-driving. But as self-driving software got more advanced, it needed more computing power than Hardware 3 can provide.

Think of it like upgrading the brain of a car. The old brain worked fine for simpler tasks, but the new software is more complex and demands more processing speed—kind of like asking a computer from ten years ago to run today's video games smoothly.

Tesla had promised customers that Hardware 3 would be sufficient for full self-driving. But the demands of the technology turned out to be higher than expected, especially when the car has to analyze video from multiple cameras all at the same time, in real-time.

This kind of situation has happened before in consumer tech. When smartphones first became popular, phones bought just 18 months apart often needed completely different software. The difference now is that Tesla is trying to retrofit existing cars—something that is normal in the car business, but far less common in consumer electronics.

How Tesla Plans to Do It

The company's plan is to build small factories in major cities. These wouldn't be huge plants like a main factory. Instead, they would be set up to handle thousands of car upgrades efficiently.

Retrofitting millions of cars through regular service centers would be complicated. You'd need special tools, trained workers, and a huge supply chain. Small, focused factories in cities would let Tesla group the upgrades together and make sure they're done right.

Tesla hasn't yet said when these factories would open, or how many cars would need the upgrade. The company also hasn't announced whether the upgrades would be free for customers who bought Full Self-Driving, or if there would be an extra charge.

What's Happening Elsewhere in the Car Industry

Other carmakers saw this coming. Companies like Mercedes, BMW, and Lucid put more powerful computers in their new electric cars from the start. They built in extra processing power to handle whatever self-driving software might come in the future.

Tesla had an advantage by getting to market first with semi-autonomous features. That gave the company millions of real cars collecting driving data. But it also meant Tesla had to support older hardware that now looks outdated.

Other News in Electric Cars and Robotics

Several other companies made moves this quarter. Redwood Materials, a battery recycling company founded by a former Tesla executive, cut about 10% of its workforce. The company also lost its chief operating officer and several other leaders.

Rival companies in the self-driving and robotics space raised money: Humble Robotics brought in $24 million for developing robots, and Reliable Robotics secured $160 million. Rivian began delivering its new R2 SUV from its Illinois factory. Amazon is adding 75 electric heavy trucks to its freight network.

What This Means Going Forward

The broader context here is that the car industry is learning what we know from phones and computers: hardware and software don't age at the same pace. Software gets better faster than hardware can keep up.

If Tesla pulls off these city-based upgrade factories, it could show other carmakers a working model for how to keep older cars current. It's also an opportunity for Tesla to build trust with customers by honoring a commitment—even if that commitment is now more expensive than anyone expected.

On the technical side, if the microfactory approach works, Tesla could use it for other hardware updates down the road as self-driving continues to evolve. For an industry still figuring out how long cars and their computers should last, Tesla's approach may end up setting a pattern that others follow.