Every now and then, a critical mass of clever, ambitious folks excited about a particularly good idea coalesces, often in a particular geographic region, and humanity gets lucky. The American colonies in the late 18th century and Detroit in the early 20th century are historical examples. Silicon Valley, starting in the 1980s, is probably our best contemporary example.
In recent years, those modern titans of technology have turned their futurist eyes towards personal transportation. Whether explicitly or in sotto voce tones, they’ve indicated that the traditional auto industry personified as “Detroit” was a dinosaur about to go extinct. Not knowing the auto industry metaphor of becoming an obsolete buggy whip manufacturer, the tech industry saw Detroit’s future as “making handsets” — i.e. low tech assemblers.
Tesla was going to show us the new electron driven future, Google was going to make cars that drove themselves, and the Apple of the tech world’s eye was going to do nothing less than completely reinvent the automobile, just as it had done with music players and telephones. The push towards self-driving, autonomous cars and trucks was only going to accelerate the ascendancy of Silicon Valley as the new Motor City.
Just because you’re good at one thing, however, doesn’t mean you’re good at another.
The history of business and industry has many examples of successful people and firms that have failed when trying new ventures. Sticking to what you know is sound reasoning in business. You may not be the black swan that you think you are.
Now, while Tesla is still moving forward and slowly expanding its lineup of EVs, the profit it is reporting this quarter is the first black ink the firm’s had in three years, and most of that is attributable to the sale of clean air credits to other automakers, not making and selling Tesla cars. Both Google and Apple have retreated from their ambitious plans to seemingly become automobile manufacturers. As Bill Ford, the great-grandson of one of those guys in early 20th century Detroit put it, “The conversation has really shifted.”
According to Bloomberg, Google co-founder Larry Page told a private meeting earlier this year that he was “heartbroken” by the lack of progress of the company’s autonomous car project, though a Google spokesman denied that Page was disappointed by the project’s accomplishments to date.
Silicon Valley may have bitten off more than it could chew. Besides developing autonomous driving technology, a serious engineering undertaking all on its own, making an automobile is about as complicated a process as manufacturing can get. For each individual car or truck model, between 20,000 and 30,000 parts have to be designed, engineered, tested and mass produced (using almost every material science and manufacturing process known to humanity), and then the assembled whole has to work reliably in a severe environment for ten to twenty years. There is a lot more to making a car than emailing Foxconn some schematics and blueprints.
James Kuffner has worked in both the tech and automotive industries, in both Silicon Valley and in southeastern Michigan. He used to run Google Robotics and is now CTO of the Toyota Research Institute, Toyota’s artificial intelligence lab not far from their main North American R&D center in Ann Arbor. Kuffner says, “Despite what a venture capitalist tells you, making a car is hard.”
Instead of using automakers as low tech assemblers for their own branded cars, it looks like Google and Apple are retreating to the position currently held by traditional automotive suppliers. They’ll supply software for autonomous driving, not build automobiles, and the auto and tech industries will establish technical partnerships. Uber is testing Volvo SUVs as robot taxis in Pittsburgh, and Fiat Chrysler is providing 100 Pacifica minivans for Google to autonomize. The Google-FCA hookup was consummated after Google and General Motors could not reach a deal over proprietary issues concerning technology and data.
It’s not surprising that ownership of data proved to be the sticking point. Owning mineable data is becoming almost as important as owning technology.
Tech firms want access to vehicle performance and manufacturing data, saying they could use that data to get better profit margins than the traditional automakers earn. Automakers don’t want their proprietary data ending up in the hands of competitors, and with Silicon Valley having made noises about getting into the car biz, competitors don’t necessarily have to have the word “Motor” in their corporate name.
Swamy Kotagiri, chief technology officer of Magna International, an automotive vendor that does everything from making components to assembling entire cars under contract, like the BMW 5 Series, says that things will likely wash out based on the industries’ areas of competence. Tech firms will supply infotainment and some self-driving tech but automakers will ultimately control anything that could touch on safety.
While they may not compete directly with product going forward, the industries do compete in hiring talent. Though Tesla and other Silicon Valley firms made a big deal about “not being Detroit”, many of the engineers they’ve hired for their automotive ventures have been auto industry veterans. Going in the other direction, GM spent a billion dollars buying Cruise Automation in part, according to CEO Mary Barra, to help the car maker attract software engineers. Ford has made similar, though smaller, investments in outside tech firms.
While the domestic auto industry (including Toyota with their Ann Arbor R&D center), the state of Michigan and area universities are making major investments in turning the Detroit area into the center of self-driving vehicle research, the industry and region understand that they are competing with California for talent.
However, I wouldn’t underestimate the ability of the automakers to compete. The traditional automakers are old hands at dealing with geographic preferences of talent. I once asked a car designer how come so many car companies have styling studios in California and he said, “Designers like pretty girls and the beach too.”
Even though for the price of a 1,000 sq ft fixer-upper bungalow in Palo Alto you can be living in a large home on a lake in Oakland County, Michigan, software engineers still apparently prefer Silicon Valley to Detroit. Just a couple months ago, Ford announced that they were doubling hiring at their Palo Alto facilities.