The future of cars is within sight. Autonomous or self-driving cars are coming to American streets. And Silicon Valley will ensure they arrive sooner than major car companies want.
In addition to a ten-minute dissertation on rocket science and arguing that all life on Earth is probably not real, Tesla’s charismatic chief executive Elon Musk made a startling assertion at the Recode conference in early June: Self-driving cars already drive better than humans in most situations and in about three years they are likely to hit the roads with the full blessing of governments across the globe. The only obstacle now is collecting enough data to prove autonomous cars are significantly safer.
Tesla has been collecting such data since it turned on its Autopilot feature in Model Ss in October 2015. With the help of its enthusiastic owners, these vehicles currently collect 1,000,000 self-driving miles every 10 hours, mapping public streets all over the world with low cost ultrasonic and high definition camera gear from MobileEye (MBLY). The data is then whisked back to Tesla servers where it is crunched, optimized and worked into future software updates. The results are impressive and carmakers worldwide have been scrambling to make up the data deficit.
Doing that is not going to be easy. Car companies really have to choose between partnering with technology companies like Alphabet (GOOG), ramping up their own efforts to collect and make sense of data or some combination of the two. None of these options appeal to an industry that became quite comfortable with incremental upgrades and a decidedly longer-term view of technology. Pushed by growing competition from Tesla, BWM recently announced an aggressive campaign to build self-driving cars over the very near term. It plans to increase the number of software engineers in R&D from 6,000 to 15,000 in just five years.
Others are looking to partnerships. Volvo will bring unsupervised, self-driving cars to London in early 2017 in addition to its extensive testing in the less regulated streets of China. And ride hailing companies will play a role too. With their large fleets traveling tens of thousands of miles every day the potential for real world data collection is huge.
General Motors (GM) made an investment in Lyft in addition to buying Cruise Automation, a SiliconValley startup building self-driving car software. Toyota announced a deal with Uber. Volkswagen is working with Gett in Germany. They’ll need all of this data to convince legislators that autonomous vehicles are safer than those driven by us absentminded humans.
The investment opportunity of this paradigm shift is exciting. Although it’s difficult to build cars, there is very little fragmentation in the supply chain for the bits and pieces that make autonomous cars work. That consists of radar, LiDAR and ultrasonic sensors, video cameras and a central computing system. Mobileye has partnered with 90% of the firms in the industry so, assuming it can meet demand for its cameras and software, its future looks bright.
The bottom line is that self-driving cars is no longer a question about technology. Tesla proved it can be done. The next step is collecting enough data to prove the vehicles can make a serious dent on the 30,000 US traffic fatalities each year. That shouldn’t be too difficult. Funeral homes, insurance companies and emergency rooms could be seriously disrupted.