I’ve been doing some research on the history of the Defense Advanced Research Projects Agency (DARPA) Driverless Car Challenge. Actually, it was officially called DARPA’s Grand Challenge for Autonomous Vehicles. The challenge took place in two stages: the first in March 2004 and the second in October 2005 (here’s an excellent oral history of the challenge from Wired).
In the first round, none of the 15 teams participating could complete the 142 mile desert course from Barstow, CA, to Primm, NV; thus, no one won the $1 million prize that DARPA was offering. From that failure, though, came astounding success. That $1 million has now spawned an industry that, in 2017, was worth $5 billion. It is estimated to grow to over $25 billion by 2024--and that’s mostly likely a low end estimate.
As part of this we will see huge disruptions to our transportation and logistics systems, with autonomous trucking, in addition to cars, drones, fire trucks(?), and a whole host of other autonomous vehicles.
This is the perfect example of the kind of market/industry creation story outlined in Mariana Mazzucato’s book The Entrepreneurial State. She sees government’s role as going beyond just correcting market failures (like the global financial crisis or monopolies) to creating industries. The gist is that early investment and support from the government spurs additional private sector investment and commercialization.
In the case of DARPA’s Driverless Car Challenge, because government was willing to tolerate failure and bring non-traditional providers into the mix, it was able to create huge gains. I see at least four key reasons the program was a success.
Accessing New Technologies and Non Traditional Participants: DoD had been working on autonomy through its traditional contractors and making little progress. The challenge allowed DoD to leapfrog traditional providers and bring new technologies and people unaffiliated with DoD to the table.
Building the Human Capital: The team members that competed are a who's who of executives in autonomy today. Stanford’s team was led by Sebastian Thrun who started Google’s driverless project now a separate company called Waymo. The head of Ford’s driverless program competed. Other competitors have gone on to run autonomy research programs at university or founded companies to supply the key technologies and components.
Grounding in a Requirements with a Deadline: The military had a goal to transform one third of it surface warfare fleet to autonomous vehicles by 2015 in an effort to reduce potential casualties. That requirement/mandate and timeline drove DARPA to get creative. Sure R&D can work on basic technologies that might not pay off for decades but a good portion of it has to be grounded in some fundamental need that has a big pay off if met.
Leveraging the Government Brand: Winning a government sponsored contest, especially one from the agency that created the Internet is good PR. As you can see from the subsequent careers of those who participated, that PR gets attention. The promise of that stamp of approval attracts the best talent.