AV Technology Why are companies investing so much in Autonomous Vehicle technology?
A profusion of news stories talk about how much the market for autonomous vehicles (AVs) will be worth in the future while ignoring a basic truism—that the amount of money a consumer will spend on a car likely won't be rising by that much (adjusted for inflation). Also, the number of cars on roads will almost surely be decreasing as time goes on—not increasing—due to intelligent and shared mobility. Thus, the large sums of capital going into R&D for AVs may be more about chasing deep-pocketed competitors (such as Google sister company Waymo or GM's Cruise) than about long-term returns on investment.
Most of us have read stories that feature plum predictions for the auto industry—those saying that the market for vehicles with autonomous capability will be worth as much as USD$1 trillion in 2030 by some estimates. By further approximations, it may be valued as much as USD$3 trillion or more by 2040 or shortly thereafter. Of course, some of these numbers assume that the majority of all cars will continue to be owned by individuals and that intelligent- and/or shared-mobility services won't make much of a dent in the marketplace.
Probably, however, they will make an impact—and an outsized one at that. According to PricewaterhouseCoopers (PwC), by 2030, the number of cars sold in Europe, for example, will fall from roughly 280 million today to about 200 million in 2030, a decrease of more than 25 percent. In the United States, the number of cars sold is expected to decrease by 22 percent, from 272 million to 212 million.
In China, the number of cars that will be sold is projected to rise by 50 percent due to consumer demand, from 184 million to 276 million, but this market is an exception to the general worldwide trend.
The above declines will be largely due to the increasing use of public transportation, car-sharing, ridesharing/ride-hailing, robo-taxis, and other intelligent/shared mobility services. PwC estimates that by 2030, slightly more than one out of every three kilometers traveled in a passenger automobile will be within an intelligent- or shared-mobility vehicle. Because cars in these categories will experience more frequent use (and disposal), car sales will rise between 2020 and 2030, even as overall car inventories decrease.
Hence, the shrinking market for individually owned automobiles means less "pie" for automakers to split. In the interim, many of these car companies are beginning to manufacture (or at least develop prototypes of) intelligent/shared mobility AVs, or they're starting ridesharing or robo-taxi services themselves. Some are doing all of the above.
By 2030, it's expected that fully 95 percent of all new vehicles will either be electric-drive (55 percent) or hybrid-drive (40 percent), resulting in significant development and retooling costs for established car manufacturers. PwC theorizes that 19 percent of the financing of vehicle development may become supplanted by investment in AV platforms, particularly on the software side. By 2030, Lux Research has calculated that specialized hardware and software will add an average additional USD$481 cost to each AV.
A great deal of R&D for AVs has, in fact, already taken place. It's estimated that worldwide, at least USD$16 billion has been invested in work on AVs thus far. And yet, outside of a handful of robo-taxi and ride-hailing services, Society of Automotive Engineers (SAE) Level 4 ("high automation") functionality in new production vehicles has not yet materialized.
While only a handful of AV platform developers and new car startups have failed in the market so far, a number of analysts have predicted an industry "shakeout" is coming due to the copious amounts of money that companies are spending trying to achieve the "holy grail" of SAE Level 4 and 5 production vehicles.
The fabled pot of gold
Consider that the average car these days sells for USD$37,000, and last year, 91.3 million vehicles were sold worldwide. This means in the last year, the one-time vehicle sale market was worth USD$3.3 trillion. By 2030, this number is expected to grow to more than USD$4 trillion, according to consulting firm McKinsey.
Between one in 10 and one in four vehicles sold in 2030 will be autonomous, meaning the AV market for individually owned vehicles will be worth somewhere between USD$400 billion and USD$1 trillion by that time. But when you add in aftermarket products and services and recurring revenue from mobility services, that number rises to somewhere between USD$670 billion and USD$1.67 trillion. It's a lot of money.
The illusory 'holy grail'
For many car companies, achieving SAE Level 4 functionality in a production automobile has been tantalizingly out-of-reach for some time.
Research company Gartner is famous for its viral Hype Cycle for Emerging Technologies chart—a graph with a mountain-like curve denoting where various future concepts stand as far as their marketing and promotional hyperbole versus what they can deliver in reality for consumers. In 2012, AVs were climbing up the left side of the mountain; the auto industry had high hopes for autonomous technology to deliver on its promises and for bundles of cash to come rolling in.
First, it was said that the initial car models sporting this autonomous capability would begin to appear in 2016. Then, that date got pushed back to somewhere around 2018. Several companies (including Tesla) have recently insisted that 2020 (or 2021 in a number of cases) would finally be the year that saw multiple vendors offering these features.
By 2015, AVs had reached the "Peak of Inflated Expectations" on Gartner's chart, in terms of what consumers were hoping for vs. what companies could realistically deliver in a medium-term timeframe.
By 2017, AVs were falling rapidly toward what Gartner calls the "Trough of Disappointment." There was a consensus (at least among auto industry cognoscenti) that while SAE Levels 2 and even 3 might finally be achievable (Audi was the first brand to launch an SAE Level 3 production car in 2017), SAE Levels 4 and 5 were still likely further away than anyone had anticipated 5 years previously.
"We overestimated the arrival of autonomous vehicles," admitted Jim Hackett, the CEO of Ford, in April 2019.
The reasons for this are many. Despite millions of kilometers of real-world testing on public roads and billions of kilometers of testing in simulations, not every traffic scenario and circumstance has been experienced or plugged into the "brain" of an AV so it can be learned.
And despite a plethora of sensors and super-fast processing chips in the electronic control units (ECUs) of prototype AVs, conditions such as heavy weather and unexpected obstacles like emergency vehicles gave rise to situations where backup safety drivers had to disengage the autonomous features of their vehicles.
While the 'holy grail' of SAE Level 4 may be possible in limited low-traffic or geofenced areas for now, the risk of accidents is too great for automakers to include Level 4 in a production vehicle and guarantee that it will work on all roads, in all weather conditions, in all countries of the world.
Chasing the deep-pocketed players
The struggle to reach this holy grail hasn't stopped lots of companies from trying to achieve it and spending fabulous sums of money along the way. Waymo (owned by Google parent company Alphabet) has likely spent the most, with USD$3.5 billion invested in R&D so far.
Much of Waymo's money has gone into conducting the most extensive real-world testing of its AV platform of any company in the world, with at least 20 million real-world miles driven and tens of billions of miles in simulation as of 2020. In 2018, the company ordered 20,000 electric I-Pace cars from Jaguar Land Rover that it will convert to its own AVs.
While Waymo originally envisaged creating a consumer-targeted production vehicle to sell to the mass-market, it's unknown if that's still a company goal. For now, Waymo operates one of the few SAE Level 4 robo-taxi services in the world in the suburbs of Phoenix, Arizona. And in the meantime, the company is selling to automaker OEMs something called the Laser Bear honeycomb, a USD$7,500 LiDAR sensor that it uses on its own AVs.
Tesla has also been a big spender, blazing a trail with SAE Level 2 ADAS functionality that's now somewhat commonplace in the industry. CEO Elon Musk has promised SAE Level 4 will arrive later this year, enabling individual Tesla owners to turn their cars into part-time-profitable ridesharing AVs, but like many of Musk's pronouncements, this one may be premature.
It's known that Apple has spent a good chunk of the last seven years working on its AV platform and doing quite a bit of real-world testing, but its AV endeavors (codenamed Project Titan) have waxed and waned over the years. Initially, like Waymo, it seemed the company had dedicated itself to building its own mass-market AV, but in 2017, it appeared the firm decided to change course, instead choosing to develop a platform that could be sold to various carmaker OEMs.
Other enterprises spending jaw-dropping amounts (measurable in billions of U.S. dollars) on AV R&D include Canada's Magna, Germany's Continental and ZF, Japan's Toyota, and the U.S.'s Intel/MobilEye and Cruise. (Cruise alone was reported to be valued at USD$19 billion in 2019 and burnt through some USD$1 billion dollars that year).
Many other ventures have clearly felt pressure to keep up with these efforts.
The rise of intelligent/shared mobility
Now that it's clear that the holy grail of SAE Levels 4 and 5 may be farther out of reach than was originally envisioned, companies such as Waymo and Intel/MobilEye have turned to ride-hailing and robo-taxi services to fund their continued AV R&D.
According to research firm Statista, the market for robo-taxis between 2020 and 2030 could be worth as much as USD$1.2 trillion. Other research groups are estimating that it could be valued by as much as USD$5 trillion by 2030. Even automaker OEMs like Japan's Toyota and Germany's Daimler are establishing robo-taxi services to follow this trend. The total amount spent on R&D for robo-taxi services was USD$27.5 billion in 2018.
But despite these efforts, it's an open question whether all existing players in the AV space will be able to hold out before (and if) the holy grail of SAE Level 4 or 5 autonomy is finally achieved.
Some research companies are predicting that universal operation at these levels of autonomy may not be perfected before 2030, while others are saying it may be many decades before this can happen.
Analysts like Mike Ramsey of Gartner are predicting that perhaps only a dozen of today's AV platform makers may survive in the long-term. Certainly, the recent Coronavirus pandemic has affected the auto industry (and may even have claimed the life of Chinese electric vehicle startup NIO, as that firm's near-term financial viability has been thrown into question).
While a number of firms will compete for business around the world in ridesharing, it's likely that dominant players will emerge on different continents and that standardization of features and platforms will force other contenders to either play catch-up or be merged into larger business entities.