You probably haven’t heard of Mobileye, but in the past couple of years it has become one of the most important players in the emerging self-driving car industry. And that means Intel’s $15 billion acquisition of the company, announced Monday, makes Intel one of the most important tech companies in the car industry.
Intel powered the PC revolution by providing the computer chips that allowed PCs to become ever more powerful. Mobileye has positioned itself to do the same for self-driving cars: Currently, conventional car companies put Mobileye sensors, chips, and software into their vehicles in order to give them advanced driver-assistance capabilities.
The company says its hardware is already embedded in more than 300 car models from 27 different car companies, and the company is reportedly working with giants such as GM and Volkswagen.
Right now, Mobileye’s technology does things like help cars on the highway to automatically stay in their lane and maintain their distance from the car in front of them. But the company has larger ambitions. As its technology grows more sophisticated over time, it wants to enable its carmaking partners to build the same kind of fully self-driving vehicles that tech companies such as Uber, Tesla, and Google’s Waymo unit are working on.
For Mobileye, the benefits of the Intel acquisition are obvious. The Israeli company will benefit from Intel’s vast financial resources and its strong global reputation.
The deal is also hugely important for Intel, because it could cement the company’s leadership of a big, emerging market that might have otherwise passed it by. A decade ago, Intel passed on an opportunity to make chips for the iPhone and ultimately lost the smartphone chip business to companies using the rival ARM chip standard. By making a big, early bet on one of the self-driving car industry’s rising stars, Intel is hoping to avoid making the same mistake again.
The big question, however, is whether Mobileye has chosen the right strategy to win the self-driving car wars. Mobileye has won contracts from conventional carmakers by building technology systems that carmakers could add to their existing cars alongside components made by many other suppliers. But Mobileye and its automaker partners might discover that this modular approach isn’t the best way to build self-driving cars.
Mobileye could be the Intel of the self-driving car revolution
Most people have heard of Intel thanks to the “Intel Inside” ads the company has been running for more than 20 years. But Intel’s business model has never really been consumer-facing. The company has always primarily sold computer chips to other companies, which incorporate them into their own products. That helps reassure PC-making partners who might otherwise worry about Intel eventually upstaging them with its own PCs.
Mobileye has pursued a similar strategy in the car market. Car companies know they need a self-driving car strategy, but they’ve been wary of working with technology companies like Google that could gain control of their customer relationships, a cautionary tale picked up from the smartphone industry.
Carmakers would desperately like to avoid the fate of smartphone makers that — over the past few years — have been relegated to the status of commodity hardware suppliers on Google’s Android platform. Many customers feel a stronger sense of loyalty to Android than they do to individual smartphone suppliers such as HTC, LG, or Samsung. The result has been bitter price competition among smartphone makers and low or nonexistent profits, while Google reaps a lot of the benefit from the Android platform.
Carmakers want customers to feel a sense of loyalty to their own car brands the way they have in the past. But if they supplied cars for a Waymo-owned ride-hailing service, for example, their vehicles could become almost interchangeable, reducing their bargaining power and ultimately their profits.
This makes Mobileye’s Intel-like business model attractive to conventional car companies. Customers who buy cars with Mobileye chips and sensors inside may not even know they’re using Mobileye technology. They’ll just know they bought a car from Ford or Volvo or whoever. Which is exactly how the carmakers like it.
Intel is aligning itself with carmakers — and against Uber and Waymo
Mobileye’s approach appeals to carmakers in another way too. Most car companies expect self-driving technology to be developed gradually. They use a five-level system created by the auto industry trade group SAE International to describe the shift to fully self-driving vehicles. Using chips from Mobileye or competitors such as Delphi, many automakers already offer driver-assist features like advanced cruise control — in which cars detect when they are approaching other vehicles, and modify speed — and stay-in-lane capabilities (SAE levels 1 and 2). The next step is partially automated systems (levels 3 and 4), culminating in “level 5” systems that are fully automated.
Mobileye has bought into this vision. It’s supplying the chips and sensors carmakers use to build level 1 and level 2 self-driving systems, and it’s working on more powerful technology that will, over time, allow the car to handle more and more situations without human intervention.
Carmakers like this vision because it keeps them at the center of the customer experience. Customers go to a traditional dealership and buy or lease a traditional car for their exclusive use for a decade or longer.
In contrast, Uber and Google’s Waymo unit have a very different vision — one in which their cars are fully self-driving from the start and never allow the customer behind the wheel. In this model, customers would likely summon cars on demand instead of owning a dedicated vehicle in a typical year.
This isn’t a vision that leaves very much room for a technology supplier like Mobileye, since Waymo and Uber are building a lot of the necessary hardware and software themselves. Uber and Waymo probably will need partners in the car industry, but those companies are likely to be relegated to more of a supporting role, with a much weaker relationship to the customer and hence a much weaker bargaining position.
The big challenge for the Intel-Mobileye partnership, however, is that the traditional auto industry approach to building a car might not work that well for self-driving cars. Self-driving cars are going to have a lot of complex software, and within the traditional car industry model, that could require a lot of different suppliers — most of them non-software companies — to write software that works together seamlessly.
In practice, that hasn’t always worked well. Even today’s traditional cars have been shown to be vulnerable to hackers, in large part because it’s hard to secure a computer system when the software is written by many different companies. The software on a fully self-driving car is going to be much more complex, and managing that complexity might prove crippling for a conventional car company that’s not used to doing this kind of work.
One result could be that Mobileye becomes more and more influential behind the scenes, gradually gaining more and more control over carmakers’ designs as they realize they need guidance from a skilled technology partner. There were some hints of this approach in a January announcement that Mobileye was working with BMW and Intel to develop a “scalable architecture” for self-driving cars that could be licensed to other automakers.
The other possibility, though, is that the complexity of integrating software from many different suppliers hampers conventional car companies’ efforts to get fully self-driving vehicles to market. That could cause them to lose the race to Waymo, Uber, or some other startup down the road.
It's worth noting that Tesla, which straddles the line between tech companies and automakers, hopes to combine the best of both models. Until recently, Tesla relied on Mobileye technology to power its Autopilot feature. But after a Tesla customer died in a car crash, the two companies went their separate ways, and Tesla is now trying to develop its self-driving capabilities in house.
Tesla already relied less on third parties than conventional automakers, and its origins in Silicon Valley give it greater software capabilities than conventional automakers. So it could prove to be the only company to occupy the “sweet spot” between Detroit and Silicon Valley, with the in-house capability to both build its own cars and create sophisticated software for them.
Disclosure: My brother works at Google, which like Waymo is owned by Alphabet.