Upstarts & downfalls

All it took was one letter to derail George Hotz's self-driving car promises.

In September 2016, Hotz took the stage at TechCrunch Disrupt, a competition in San Francisco for startups, and announced he was mere weeks from selling a device that could turn a regular car into a self-driving car.

He whipped out a device that looked like an original Nintendo DS gaming system and said the Comma One would be "harder to come by than a pair of Kanye West sneakers."

Six weeks and a lot of media attention later, Hotz announced that he was canceling the project after receiving a letter from NHTSA questioning the device's safety and efficacy and urging Hotz to delay releasing it until he could prove it would be safe.

"Would much rather spend my life building amazing tech than dealing with regulators and lawyers," Hotz tweeted at the end of October 2016, after announcing the cancellation. "It isn't worth it."

Brash Silicon Valley upstarts such as Hotz see the auto industry as ripe for disruption. The industry is slow to change, they say, and they argue it's time to rethink the ways in which vehicles are powered, sold and owned.

But after several years of trying to buck norms, the disrupters are running into the same problems that the traditional auto industry has faced for decades: safety regulators, environmental regulators and consumer protection agencies.

“A car is not exactly like a computer on wheels,” said Kristin Schondorf, global automotive and transportation mobility leader at EY.

Automakers, suppliers and the ecosystem surrounding them have dealt with growth-

restricting obstacles for decades, including regulations, automotive-grade standardization, product liability and affordable mass manufacturing. As industry disrupters seek to ensure their longevity, they are running into these problems for the first time, and their ability to learn from the past may affect their future.

“The thing these companies struggle with the most is, is their product durable enough?” said Mike Ramsey, a technology analyst at Gartner Research. “Can you guarantee that you will be around and capable of servicing millions of vehicles in the future if something goes wrong?”

It can be difficult to ensure that software in a vehicle is automotive grade, meaning it can withstand 10 years of wear and tear.

"From a hardware perspective things are really measurable," said Walter Sullivan, head of the Silicon Valley lab for software supplier Elektrobit. "Software is where it starts to get more vague. Software isn't as measurable."

This ambiguity leaves companies few ways to ensure their technology can hold up in real-world situations or withstand cybersecurity attacks, Sullivan said.

Automotive startups also face the multifaceted challenge of mass producing their products. Some, such as lidar company Luminar, work with automakers, such as Toyota Motor Corp., to bring their processes up to automotive speed.

Others, such as Tesla, choose to forge their own paths. For Tesla, the result has been a slow and unsteady road to increasing output of the Model 3 sedan. The automaker has already pushed back its target of producing 5,000 Model 3 vehicles a week twice, and missed its goal of 2,500 a week in the first quarter of 2018.

The ability to mass produce automotive grade vehicles has weeded out weaker players over the years, said Ron Harbour, a manufacturing consultant with Oliver Wyman.

"It's no different now," he said.

How can Silicon Valley startups get around this problem? By forging partnerships with the companies they are trying to disrupt.

"This industry has developed detailed processes and procedures for a reason," Schondorf said. "I'm relatively certain new entrants don't have that, and if they think they do, they're short-lived."