Elon Musk Is Wrong About the Future of Self-Driving Electric Cars in America

"Probably, in 10 years, more than half of new vehicle production is electric in the United States," Musk said. "In 10 years, yeah, half of all cars will be EV. I think almost all cars will be autonomous."

To which we feel compelled to say: Sorry, Elon, but you're wrong. There's no way half of new cars sold in America will be electrically-powered by 2027. And there's even less chance most of them will be capable of driving themselves. 

Let's start with the first part of Musk's statement, the claim that at least 50 percent of new passenger vehicles built in the U.S. will be electric cars in a decade's time. For one thing, we don't have nearly enough batteries, and the batteries we have aren't cheap enough. In 2016, Americans bought 17.6 million new cars; while the exact annual sales number may ebb and flow, projections by IHS Market suggest the market should hold steady around that level for the next decade. Assuming that's the case, Musk's prediction would mean 8.8 million new electric cars would hit the streets in 2027 alone.  

Volkswagen's head of R&D has suggested the carmaker will need 200 gigawatt-hours of battery capacity to build three million electric cars per year. Using that as a baseline, that means it will take plants making around 600 GWh of battery capacity to build those 8.8 million electric cars for the U.S. market. But that output far exceeds most predictions for global lithium-ion production capacity around that timeframe; a 2017 study, for example, predicts just around 500 GWh of global li-ion battery production by 2025. 

The cost of those massive batteries, likewise, is also an issue. They're not cheap. Currently, the cost of a battery pack stands around or below $200 per kWh of capacity, depending on whether you ask McKinsey & Co. or Tesla. The former has claimed that EV battery packs need to drop below $100 per kWh in order to achieve price parity with gas cars—something McKinsey expects to happen closer to 2030. (Those costs are partly offset by the lower costs of the rest of the electric powertrain, however.) 

And even if electric cars reach price parity with gas ones by that time—something that Bloomberg New Energy Finance says is likely to happen by 2025 at the earliest—that does nothing to make up for the fact that electric cars are in many ways still less convenient to recharge than conventional cars are to refuel. New, super-powerful chargers, like ChargePoint's 400-kW units, will help, but at the cost of degrading the car's battery life more quickly than the li-ions would go otherwise. And even with those new chargers, it will likely still take longer to recharge an EV than the four minutes it takes to top off an internal-combustion car at one of the 168,000 gas stations across America. 

But the far bigger obstacle to electric car success in America is likely to be buyers themselves. In spite of massive advances in range and charging speed in recent years, less than 1 percent of U.S. new car sales last year were EVs. The Chevrolet Bolt offers more than 230 miles of range at a price lower than the average new car (once tax credits are factored in), yet dealerships are sitting on a glut of them. Markets from Hong Kong to Georgia to Denmark have proven buyers seemingly only want electric cars if governments reward their purchase with massive incentives, and it's exceedingly unlikely that Scott Pruitt's EPA will be pushing those in the U.S. If anything, we'll be lucky if the Trump Administration doesn't roll back existing regulations that favor green vehicles. 

As for Musk's claim that most new cars will be self-driving by 2027, it's hard to see that as anything short of all-out hyperbole, unless he's including any car that ranks anywhere on the SAE's five-tier scale of vehicle autonomy. Mercedes-Benz is aiming to have its first self-driving robo-taxis on the road somewhere between 2020 and 2025; customer cars capable of driving themselves will come even further down the road. Honda is aiming to simply have Level 4 cars in production by 2025—a far cry from having them make up the majority of the fleet. Kia has put full autonomy in the ballpark of 2030. There's a long list of carmakers working on self-driving cars; none of them are saying they see most new vehicles on the road driving themselves a decade from now.

And we're not the only ones saying this. Goldman Sachs predicts that one-quarter of new cars will be EVs come 2025, and that's globally—including massive car markets like China and Germany, which have already declared their bold plans to push towards an all-electric future. A recent report by the Edison Electric Institute and the Institute for Electric Innovation projects around seven million electric cars on U.S. roads by 2025. And getting full autonomy into most new cars in 10 years' time would require unexpected leaps in both technological development, government regulation, and consumer adoption—three things very unlikely to happen simultaneously. 

We don't begrudge Elon Musk for making such bold claims. He's the CEO of a company that sells electric cars with advanced self-driving features. It's not hard to see why he'd be interested in getting people excited about self-driving electric cars. We'd likely be doing the same thing were we in his shoes—and indeed, we'd be pretty happy if his predictions were right, because that would likely mean safer roadways and a healthier planet. 

But that doesn't mean he's correct.