Click here to jump to part two of this post, published on July 13, 2016.
Among automakers, Toyota had stood out for its stubborn refusal to develop autonomous vehicles. Even as Google’s prototypes took to the streets of California and Tesla’s Autopilot features made their way to consumers, the world’s largest automaker, helmed by the racing enthusiast Akio Toyoda, remained steadfast that cars were meant to be driven by people.
In late 2014, a company executive even went so far as to say that human beings were the “ultimate in sensor fusion” capable of making judgements better than any driverless car.
But just a year later, the company reversed course. In late 2015, it announced it would spend at least $1B on a Silicon Valley research center for autonomous driving and robotics, as part of a plan to put a self-driving car on highways by 2020.
The auto industry has long been considered an oligopoly, with high barriers to entry including the need for massive capital expenditures and navigating a mystifying array of regulations. But in recent years, the threat posed by self-driving upstarts like Apple, Uber, Google and Tesla has become too great for Toyota and its big auto brethren to ignore.
Indeed, a growing chorus of VC and tech pundits has been predicting the demise of the major automobile manufacturers at the hands of Silicon Valley heavyweights. In the era of the self-driving car, these critics argue, incumbents like Ford and General Motors will be unable to compete with the fast innovation cycles, software expertise and unorthodox thinking of their disruptors.
Big auto breaks the tech monopoly on autonomous car buzz
Earlier this decade, self-driving vehicles were closely associated with tech companies, and Google in particular. But our Trends tool, which sorts through millions of media articles to quantify attention to developments in tech, shows that the Google-autonomous vehicle association is loosening.
Looking at GM, we can see the impact of its recent moves in the autonomous space, actually pulling ahead of Apple (whose autonomous vehicle efforts remain unclear).
It may still be true that Google’s AI expertise will carry it to commercialize fully-autonomous vehicles ahead of competitors, or that Uber’s mountain of trip data will give it a leg up in coordinating the routing of autonomous vehicle fleets. Despite this, big auto manufacturers still have billions in cash and sufficient R&D muscle to make the race a close one.
Though the incumbents might not have the software expertise or coastal cachet of the tech firms, they do have decades of experience manufacturing cars, managing a complex web of suppliers, and millions of loyal customers around the world. On top of that, these automobile giants are moving quickly to put the pieces in place that will prepare them for our self-driving future. Here’s why big auto still has a chance to shape the future of personal transport.
Auto executives have recognized the threat of auto tech
A key piece of the narrative is the auto complex’s acknowledgement of the threat (or opportunity) facing the industry. This may seem like an obvious hurdle to clear, but incumbents in other industries like print media and smartphone manufacturing had remained stubbornly confident that their basic business models would withstand new technologies as the walls came down around them.
By contrast, many key auto executives have signaled their willingness to adapt existing business models in order to survive.
Notably, General Motors has pursued the “endgame” of a self-driving taxi fleet with an urgency coming directly from its executive leadership. In an op-ed this past January, GM CEO Mary Barra emphatically embraced tech-driven trends in electrification, sharing, and automation.
“We are moving from an industry that, for 100 years, has relied on vehicles that are stand-alone, mechanically controlled and petroleum-fueled to ones that will soon be interconnected, electronically controlled, and fueled by a range of energy sources,” Barra wrote.
While Barra optimistically described the coming changes as an “opportunity” to reshape the sector, others in the industry acknowledge that the erosion of individual car ownership threatens to reduce them to commoditized OEMs.
To this end, BMW’s board member for R&D Klaus Froehlich told Reuters in March that the company needs to rapidly improve its artificial intelligence and machine learning capabilities to prevent itself from becoming a white-body manufacturer for tech giants deploying autonomous software. Over the next five years, BMW aims to aggressively hire software engineers to the point that developers will comprise half of its R&D department.
“For me it is a core competence to have the most intelligent car,” Froehlich said. “Our task is to preserve our business model without surrendering it to an internet player. Otherwise we will end up as the Foxconn for a company like Apple, delivering only the metal bodies for them.”
And as noted, even the industry’s most strident holdout in Toyota has made a quick about-face to bring its self-driving technology up to speed.
“I believe there is a point in Toyota participating in the field of autonomous driving. And we’ve got the resources to do so,” Toyota Motor Corporation president and CEO Akio Toyoda said.
To be sure, automakers such as Toyota are late to the game and have considerable ground to make up. But even with first-movers such as Google years from commercializing fully-autonomous vehicles, it would be premature to rule automakers out entirely.
Big auto investing heavily both organically and inorganically
Of course, it’s one thing for an incumbent to boast about adapting to technological change, and another thing entirely to execute on it. Indeed, cynicism is not entirely unwarranted given that many corporate executives have held superficial “innovation” meetings with startup founders in displays of “corporate innovation theater.”
But big auto companies are indeed putting their money where their mouths are to narrow their current gaps in automation and beyond. Car manufacturers worldwide are investing heavily in internal R&D as well as deploying capital to tech startups.
In recent years, automakers have established formal in-house auto tech initiatives at a rapid pace. We’ve covered many of these in the past; besides the BMW software recruiting initiative mentioned above, GM also announced plans to hire 700 engineers for a new software development center, and Toyota has pledged at least $1B for its advanced Toyota Research Institute.
Other automakers like Volvo have also been quietly working on self-driving initiatives for years. Besides major tech corporates, Volvo itself may be a front-runner in the autonomous race, being the first automaker to both accept liability for its self-driving cars and commit to deploying fully-autonomous cars by 2020.
Not content with to compete solely through organic R&D, players like GM have also been aggressive in leveraging the startup ecosystem, paying upwards of $1B this March in the landmark acquisition of self-driving tech startup Cruise. In addition to more than seven other startup investments through its GM Ventures arm (see our infographic above), the company has made two ride-sharing plays that could be crucial to helping it establish a self-driving vehicle fleet: a highly publicized $500 million investment in Lyft this past January and the acquisition of Sidecar’s assets later in the month.
In addition to investing in companies that are working on self-driving technology, GM and other big auto manufacturers are pouring dollars into next-generation manufacturing, powertrain, and materials technologies. In March, Daimler announced it would invest $543 million in a new factory for producing the lithium-ion batteries, the backbone and most cost-intensive component of modern EVs (electric vehicles). General Motors has also been active, investing $3.2M in the lithium-ion battery company Sakti3, with the goal of producing solid-state batteries with higher energy density and lower costs than prior technologies.
Mapping and traffic data, seen as vital pieces of the self-driving equation, have also been a focus for automakers. A little under two years ago, the Porsche family spent $55M for a stake in Inrix, a startup using real-time data from sensors on more than 4 million GPS-enabled vehicles to build traffic tracking, telematics and mapping products.
As mentioned, perhaps the most notable startup deals have come from General Motors. In addition to a number of startup investments through its GM Ventures unit, its highly-publicized $500M January investment in Lyft touched off similar commitments by Toyota (Uber) and Volkswagen (Gett). These moves also highlight the next arm of the multi-prong strategy that automakers are pursuing to diversify their development risk — partnerships.
Automakers also embracing tech partnerships & startup acquisitions to fill gaps
Despite these significant expenditures, critics might still assert that no amount of investment could help big auto overcome the knowledge base of tech players in fields like artificial intelligence, or the large networks already established by ride-hailing startups.
In some sense the auto industry has already implicitly conceded this point, as many leading players have not been above collaborating with tech partners to account for deficits in such areas. GM’s partnership with Lyft is just one example; tie-ups from Toyota-Uber to BMW-Mobileye-Intel and Microsoft’s offer of auto tech services illustrate that the two sectors have begun actively reaching across the aisle for opportunities.
As Venrock’s David Pakman acknowledged in his bearish view on the auto industry, modern automakers are also very comfortable with the sourcing and integration of parts produced by a variety of independent suppliers. From a hardware perspective, the near future of auto tech looks to be much the same, with autonomous hardware from the likes of Delphi and Mobileye (along with tech firms like Nvidia and Intel) proliferating across vehicles from Tesla to Volvo.
To be sure, open questions still surround these cross-industry engagements. Firms like Toyota have been leery of sharing knowledge with Google in the past. Even in the Mountain View giant’s research partnership with Fiat Chrysler, it has yet to be determined as to which party will own the data generated by the trial.
As mentioned, for car companies leery of getting too cozy with tech giants like Google, startup acquisitions represent a better alternative. GM has obviously been aggressive in this sense with its $1B March 2016 acquisition of autonomous tech startup Cruise Automation, in addition to buying up Sidecar’s assets earlier in January.
Tikhon Bernstam, an angel investor in Cruise, has pointed out that other autonomous tech startups could present similar targets for automakers (and by extension, exit opportunities for investors):
“Does acquiring Cruise make GM worth 2% more by helping turn it into more of a software company? Absolutely, I’d say… what’s the cost of falling behind in the transformation of the auto industry into software companies?”
Indeed, a host of other early-stage startups are working to crack the self-driving puzzle, from a16z-backed comma.ai to rivals like nuTonomy, Drive.ai, ADASWorks, and Nauto. And even beyond these companies, the big auto manufacturers will have no shortage of well-capitalized partners and acquisition targets to choose from. Investors poured $409 million into auto tech companies in 2015, up 154% from the year before.
Intra-industry cooperation has picked up as well, with the German consortium of Audi, BMW and Daimler coming together to spend $3 billion for HERE, Nokia’s high-definition mapping technology.
“By acquiring a global platform in HERE, it also may establish a de facto standard, giving the industry a unique opportunity to rally around specifications or processes that could in fact hasten the onset of a connected and automated future mobility ecosystem,” the market research firm IHS said in a note quoted by CNBC.
Indeed, with the support of its three major backers, HERE is now moving forward with its open specification for car-to-cloud data communication, aiming to create a common language for capturing and processing the vast amount of data generated by autonomous vehicles.
In aggregate, it seems that automakers have at the very least mounted a robust campaign to meet disruption risks by simultaneously budgeting for internal R&D, finding common ground for collaboration, and working to acquire autonomous and connected technology companies.
Tech players lead in data collection, but room for automakers to narrow gap
As with other machine learning-based applications, the algorithms driving self-driving cars — along with other auto tech applications, such as suggested destinations — are expected to rely upon copious amounts of real-world driving data. Certainly, Tesla and Google have the edge here at the moment, having deployed autonomous or semi-autonomous vehicles on public roads for some time now.
Google has fielded a few dozen prototypes for several years, collectively driving more than 1.7 million fully-autonomous miles, including complex urban-environment data. On the other hand, Tesla has gathered over 100 million miles of mostly highway data in the first 18 months of its ongoing Autopilot “beta,” through roughly 70,000 Autopilot-capable vehicles.
While these figures are impressive and well ahead of the field in these early days, they represent varying levels of data fidelity, and are still a far cry from the potentially billions of miles needed to concretely establish the safety of autonomous vehicles. It’s not unreasonable to expect that incumbent automakers can rapidly build their own data pools, even with incremental market penetration of their self-driving tech starting with top-range luxury models.
In comparison to, say, Tesla’s relatively small and US-centric install base, General Motors sells nearly 10 million new cars globally every year, including over 270,000 Cadillacs. Germany luxury marques BMW, Mercedes, and Audi each move nearly 2 million vehicles annually. These are just a few of the brands that have planned or begun to roll out semi-autonomous systems in the vein of Tesla’s Autopilot.
Apart from fielding self-driving cars themselves, large automakers also have access to other existing data streams. Continuing with the GM example, the firm already has 12 million connected vehicles on the road through its integrated OnStar telematics systems. GM is pursuing plans to combine crowdsourced data from its OnStar customers with Mobileye mapping technology to help create the sort of high-fidelity, continuously-updated maps that might power a driverless vehicle.
Meanwhile, the German manufacturers that purchased the mapping company HERE are also working to collect anonymized customer driving data to improve their own mapping technologies.
“The high-precision cameras and sensors installed in modern cars are the digital eyes for updating mobility data and maps; in this way, information such as speed limits or critical driving situations are already recognized today,” Audi chair Rupert Stadler said in a statement following the acquisition.
These factors have led Raj Rajkumar, a Carnegie Mellon professor who co-directs a GM-funded research lab, to assert that data is an area where incumbent automakers are strongly positioned against their would-be disruptors.
Of course, gathering robust datasets is only half the battle, but automakers are making strides in analytics as well. As one example, Ford led Pivotal’s Series C this May, investing over $180M to advance its Smart Mobility subsidiary’s capabilities by leveraging Pivotal’s cloud and big data tech.
As a side note, recent remarks from the White House’s Council of Economic Advisers on artificial intelligence postulate that government agencies may need to begin scrutinizing large proprietary datasets from an antitrust perspective, given their potential to foreclose market entry to competitors.
Incumbents still have advantages in manufacturing at scale
In many of these fields on the frontier of auto tech, incumbents are admittedly playing catch-up to various degrees. However, automakers still retain one obvious advantage: their sprawling networks of capital assets and existing processes for vehicle testing, supply chain management (SCM), and manufacturing.
One of big auto’s core competencies remains its ability to source components from an array of suppliers and assemble vehicles expected to perform reliably around the globe.
Indeed, despite a recent spate of high-profile recalls, on the whole automakers have quietly been turning out the safest and most reliable vehicles seen to date. In addition to having the infrastructure to stress-test their cars across climates from deserts to tundras, these firms have deep experience complying with vehicle safety and environmental regulations in each of the markets they do business in.
From a manufacturing standpoint, automakers have been on the forefront of adopting industrial automation technologies, and have also managed to renegotiate many cumbersome union labor deals. Finished vehicles require the implementation and management of a complex supply chain that tech newcomers will have to replicate in order to commercialize their products. John Krafcik, the current CEO of Google’s self-driving car project and a respected Hyundai executive, has pointed to this as a hurdle that tech companies may be underestimating.
“In general—and I would say this specifically does not apply to Google—I think there isn’t sufficient understanding for just how hard it is to make cars. There’s not sufficient depth of experience in the complexity of the automobile.”
Tesla has often been lauded for being the first technology-minded firm to break into the exclusive club of “full-stack” automakers, and rightfully so. It managed to capitalize on big auto’s post-recession woes in 2010, snatching up the former GM-Toyota NUMMI plant at a bargain price.
However, Tesla has struggled at times to maintain quality control and manufacturing pace over what is currently still a low-volume, luxury-segment operation. With ambitious goals, its models have often been late to market, and its latest Model X has suffered a number of teething issues and missed production targets.
As Tesla scales for production of its first volume seller in the upcoming Model 3, these challenges while likely grow in tandem. The more affordable sedan saw record numbers of pre-orders, but some analysts have questioned Tesla’s ambitious targets of 500,000 vehicles by 2018 (from a projected 80,000 for 2016). It also remains to be seen whether mass-market consumers will be as forgiving of potential quality issues as its loyal early-adopters.
Industry experts said Tesla’s new goals were extraordinary and raised doubts it could meet them. The handful of North American auto plants capable of building 500,000 vehicles a year are all run by automakers with decades of experience, they said.
An alternative approach for tech firms is contract manufacturing, which Apple and Google have been rumored to be considering. Apple, for one example, is famously masterful in retaining design control of its products while handling production through third-party manufacturing and supply networks.
However, an automobile obviously represents a different order of complexity than a smartphone, and the magnitude of quality control slip-ups would be correspondingly higher, with costly recalls and of course human lives at stake. And even Magna Steyr, the world’s largest automotive contract manufacturer and a rumored partner for Apple, currently has an annual capacity of just 200,000 vehicles at its primary plant (most of which is already spoken for). This represents less than half a percentage point of the nearly 69 million vehicles produced in 2015.
With traditional automakers like BMW adamantly refusing to become the “Foxconn” for technology companies working on auto tech, it remains to be seen where new entrants will find excess capacity for contract manufacturing in the near future. New plants (perhaps from opportunistic partners in markets like Asia) are certainly possibilities, but long lead times would provide a window for big auto to muster their response.
Of course, none of these barriers are insurmountable. And in theory, electrification should simplify some of these issues (e.g. bypassing most emissions regulations and decreasing powertrain complexity with fewer moving parts). But in the near-term, infrastructure as well as decades of experience in SCM and quality control are arrows that incumbents do have in their quivers to fend off newer rivals.
Automakers also have global distribution networks and brand equity of their own
Another piece giving major automakers a more straightforward path to commercialization is their extensive network of dealerships for delivering vehicles into consumers’ hands. Though consumers are turning heavily to internet research to avoid the hassle of multiple dealer visits, the final physical touchpoint of a test-drive remains vital to the car-buying process, with 82% of global consumers performing at least one. And interestingly, J.D. Power finds that consumer satisfaction with dealers is higher than ever before, at least in the US.
“People are more satisfied with dealers today than they have been,” says John Humphrey, senior vice president for global automotive operations at J.D. Power. “They’re more satisfied when they buy and when they go for service.”
Despite this, the direct-to-consumer showroom model pioneering by Tesla is understandably more attractive than haggling with an independent salesperson. But for Tesla, Apple, or Google to build like-size networks of owned and operated showrooms would incur large capital costs for the necessary real estate and personnel. According to Tesla’s website, the company has 231 stores, compared to 12,000 dealers for Ford alone as of the end of 2014.
In the future, applications of novel technologies like VR showrooms could boost consumer confidence in remote purchasing (e.g. through channels like Apple’s existing retail network, or purely online). But as these innovations will likely trickle down from luxury segments, for the time being the incumbent automakers still have many more global mass-market touch-points to reach their customers.
Finally, though major automakers are often derided within tech communities, it’s worth noting that many still retain significant brand equity in the eyes of the average global consumer. Technology giants do usually top lists of the world’s most valuable brands, but major automakers follow closely behind.
As mentioned earlier, many automakers have earned the trust of consumers by turning out increasingly reliable and safe vehicles. Even if one believes in an ownership-free autonomous future, an uncertain timeline and lengthy replacement periods ensure that consumer brand loyalty will still be relevant for at least a few cycles.
As for which brands consumers will associate with car manufacturing 15, 20 or 50 years from now? Well, that remains to be seen. But don’t count out big auto just yet.
Looking for more data on auto tech? Sign up for CB Insights tech market intelligence platform free below.
If you aren’t already a client, sign up for a free trial to learn more about our platform.