After a decade of massive investments and runaway hype, Silicon Valley’s big push into electric and autonomous vehicle technology has produced as many struggling startups and outright failures as it has new “mobility tech” icons. But as far as public perception is concerned, there’s a surprising divide between the two technologies: When AV developers like Argo AI go under, the technology itself is derided as a “dead end” in splashy magazine stories, whereas the prominent struggles of EV startups have done little to dim the perception that the electrification of the car is an inevitability.
How is it that two technologies that were both seen as being inevitable have so dramatically parted ways? Why have EVs remained on the hype train, while AVs have been dumped trackside and rolled unceremoniously into the ditch of disillusionment? Does the public discourse actually reflect the real prospects for these technologies, or is there something else going on?
The easy answer is that a well-heeled consumer can go out and purchase an electric car today but can’t purchase a self-driving car any more than they can purchase a flying car. By that standard it seems all too obvious why popular discourse treats EVs as being more real than AVs. The problem with this attitude is that it assumes that these technologies are only real to the extent that they conform exactly to the existing conception of a car, which is hardly the only possible standard for judging a new technology.
The blame for this erroneous conception of success and failure can be laid at the feet of Tesla, the first company to show that electric vehicle technology could be squeezed into the form of a “real car.” By jamming enough batteries into a large, heavy car, and by placing rapid chargers around freeways, the majority of drivers’ use cases could be served with more or less minimal sacrifice. This approach has been wildly popular among those who can afford it; but if the only way to sell an EV is with a 300-mile-range battery, the technology will struggle to break out of its wealthy status-symbol niche.
After all, EVs are already well behind their stated affordability goals. Nearly 17 years after Tesla CEO Elon Musk promised that expensive long-range EVs would inevitably give way to affordable options, U.S.-market EVs have been selling at a significant premium to internal combustion (which are themselves selling at all-time-high prices). The recent decline in average EV selling prices looks more like the premium EV bubble leaking air than a broad expansion of electric car sales across all price points, and mounting supply chain constraints will continue to challenge big battery EV affordability.
If the current crop of Tesla-aping long-range EVs has overstated how well EV-battery technology fits the use case of cars as we know them, driving-automation technology has had the opposite problem. Autonomous vehicles are out on public roads around the country, with companies like Waymo and Cruise giving rides in San Francisco’s complex downtown, and semi trucks from Aurora and Kodiak are hauling goods across the sunbelt, but the technology might as well not exist as far as consumer vehicles are concerned. The mental model we imagine when we hear “self-driving car,” namely a car like any other that drives itself, is as far off as it was before Silicon Valley started hyping mobility technology.
For Americans, a car has two fundamental attributes: you own it, and you use it for all of your mobility needs, whether you’re going one mile or 1,000. Unfortunately, getting modern “AI” to drive at expected levels of safety requires limiting the operating domain and using sensing hardware that costs more than the car itself. In short, the “self-driving car” faces cartoonishly exaggerated versions of the same challenges EVs face: outrageous costs and limited operating range.
If we only bought cars for our 90% use cases, both EVs and AVs would be much closer to widespread adoption. After all, the average American only drives about 40 miles per day, and 95% of our car trips are 30 miles or shorter; long-range trips may be the Achilles heel of these technologies, but they’re not a common occurrence. The problem is that these rare use cases are increasingly top of mind when we buy cars, which is why most Americans now buy trucks and SUVs that trade off efficiency in everyday driving for rarely used capability rather than mere cars.
It’s no coincidence that the company that popularized the version of EVs that is most like cars is also the only company selling the “self-driving car” that Americans imagine. But whereas Tesla has made some progress in reducing the cost of EVs, stoking faith in its ability to continue doing so, faith in its “full self-driving” option has become impossible to maintain in the face of its basic technical implausibility and Elon Musk’s annual Groundhog Day-style assurances that it will be done by the end of the year. The fact that consumers, investors, and regulators haven’t already brought the farce to an end is tribute to how powerfully Americans want to believe in the only commercial offering that actually fits their mental model of a “self-driving car.”
Meanwhile, if you look at electric drivetrain and driving automation technologies outside of the car-dominated consumer perspective, the picture looks very different. Businesses don’t fantasize about heading off into the unknown at the drop of the hat, so a Ford E-Transit with 126 miles of range or a Daimler Truck’s eCascadia with 230 miles of range can easily fill their actual needs. If consumers tracked their actual mobility needs as obsessively as commercial fleet managers, and bought for their 90% use cases instead of one skiing trip per year, the EV affordability problem would largely be solved.
Similarly, taxis and delivery bots that operate in one city only, or semi trucks that endlessly run a single route, are all promising use cases where driving automation is proving real value. You can get robotaxi rides or have goods delivered autonomously in some American cities today, but because each new operating domain requires the automated driving system to be retrained and validated, autonomous services aren’t scaling as rapidly as cars. With wages, vehicles, and freight all facing sustained inflationary pressures, the economics of automated vehicles will only improve along with the technology.
In the first phase of both EV and AV hype, success has been defined by the extent to which these transformative technologies can be squeezed into the shape of the car we’ve become so used to. In part, that’s because the car so totally defines mobility for Americans, but it’s also because popular discourse around emerging technologies was shaped in the era of the internet and smartphones, where consumer adoption was the entire game and there were few entrenched consumer behaviors to overcome. With mobility technologies, the user-facing applications have received much of the hype; but as time goes on, it is becoming clearer that a century of consumer habits mired in the human-driven, gas-powered paradigm are a very real obstacle that simply doesn’t exist in more pragmatic business applications.
It won’t be easy for Silicon Valley to shake off its internet-era habits in the mobility-tech era, as activating consumer hype has been a powerful tool in the venture fundraising game. But it also leads down technological dead-ends (or at least detours), like big battery EVs and “full self-driving” Teslas that never actually deliver on their promises. If we’re going to someday harness the true potential of these technologies, we need to learn to appreciate them on their own terms.
Edward Niedermeyer hosts the podcast The Autonocast and is the author of Ludicrous: The Unvarnished Story of Tesla Motors.
Recognize your brand’s excellence by applying to this year’s Brands That Matter Awards before the early-rate deadline, May 3.
