A report released today, however, brings George Jetson back Earth. Yes, self-driving cars will be an estimated $87 billion market by 2030, according to Lux Research prognosticator Cosmin Laslau. But even by then, Google-like autonomous cars will only account for 8 percent of the automotive market, according to the report.
Laslau labels such cars “Level 3” vehicles—capable of operating autonomously with the aid of high-resolution maps and other technology but still in need a human in the seat as a backup system. What he calls Level 4 cars are so completely autonomous that automakers can do away with steering wheels and pedals. Those cars won’t even exist 20 years from now. (Note that this is a different level-based schema from the one employed by the National Highway Transportation Safety Administration.)
Lux has no doubt robot cars are coming down the road but in 2030 the highways will be jammed with Level 2 cars that boast technology already showing up today: adaptive cruise control, lane departure warning and collision avoidance braking. More than half of cars will offer such features by 2020, and a decade later 92 percent of cars will roll off the assembly line able to operate semi-autonomously.
Level 1 cars – think anti-lock brakes and electronic stability control – are common today. That’s the technology that keeps my 2012 Ford Focus from stalling out on a 30 percent grade section of the Berkeley Hills as I follow a Level 0 car – the ’72 Volvo driven by a retired University of California professor at 10 miles per hour. Technology-free Level 0 cars are largely obsolete, according to Laslau, though not necessarily in my neighborhood.
“The pathway to commercialization remains unclear, not only in terms of adoption timeframe, price point, and business model, but also in surprisingly fundamental questions like how the cars will look like,” writes Laslau. “Whether automakers can actually technically develop mass-market autonomous cars—and how quickly they can do so—is a separate question that is plagued by uncertainty.”
That’s because of the incredibly complex technology challenges behind building such vehicles. (Not to mention the cost: A lidar rangefinding system currently can cost between $35,000 and $80,000, according to Laslau.) For instance, as my colleague Alexis Madrigal revealed last week, one reason Google’s robot cars are able to navigate so effortlessly around Silicon Valley is that the company has developed incredibly precise and data-rich maps of the area to help guide them.
How precise? As Madrigal wrote:
They're probably best thought of as ultra-precise digitizations of the physical world, all the way down to tiny details like the position and height of every single curb. A normal digital map would show a road intersection; these maps would have a precision measured in inches.
While Google aims to replicate such maps for the entire country, Laslau contends that such challenges means truly autonomous, Level 4 cars will be decades off.
“All told, remapping all (or even most) of the world’s roads with much higher resolution is a formidable and costly challenge; keeping all these maps update to date even more so; and ensuring they contain no errors is nearly impossible,” wrote Laslau.
Then again, conventional wisdom once had it that was sheer folly to think that a Silicon Valley startup could build a long-range electric car that would outsell fossil fuel-driven luxury sedans from automakers like Audi, BMW and Mercedes.