How a DARPA-funded series of contests a decade ago boosted autonomous driving tech; engineers from its top teams are leading today's self-driving startups
Dana Hull / Bloomberg : Tweets: @a_riley17 Tweets: Anne Riley Moffat / @a_riley17 : Pretty much everyone working on self-driving cars competed in this 1 contest 10yrs ago. It's a who's who of autonomy https://www.bloomberg.com/... pic.twitter.com/RE8Amy5gmy
Context & Ripple Effects
This piece is a retrospective on how the field got its people: DARPA funded a series of autonomous-driving contests roughly a decade ago, and engineers from the top teams went on to lead today's self-driving startups — as Anne Riley Moffat put it, a who's who of autonomy from one competition. It lands mid-arc in our coverage: the race between Silicon Valley and Detroit to build a self-driving car was already in full swing by mid-2016, and this explains where much of that talent came from.
What makes the retrospective worth reading alongside later coverage is the gap between pedigree and progress: by 2022, investors had put an estimated ~$100B into these startups with little to show for it, and even the big players — Uber, Tesla, Waymo — had a disappointing 2018 while only minimum-viable-product startups advanced.
First-order effects
- The immediate effect is reputational and genealogical: the self-driving startup landscape can be mapped back to a handful of DARPA contest teams, meaning hiring, partnership, and credibility in autonomy still run through that alumni network.
Second-order effects
- Capital followed the contest lineage — the estimated ~$100B investors later bet on these startups was underwritten by the same founder pedigrees this piece traces, so valuation risk concentrated wherever those teams clustered.
- Contest-era rivals re-emerged as commercial combatants over infrastructure: the fight over mapping for self-driving vehicles among Google, GM, and Uber shows former challenge participants now competing to own the data layer rather than just the vehicle.
Third-order effects
- If the pattern holds, prize-style government contests function as an industry-formation tool: DARPA seeds a field with a competition, the top teams spin out into companies, and private capital industrializes what public money prototyped — a playbook DARPA has since reused in AI security and other programs.
- Because so much of the sector's leadership descends from one event, the industry's technical culture and assumptions are unusually homogeneous, which helps explain why the whole cohort hit the same wall of mounting losses and shortcuts rather than diverging approaches.
The trend: State-funded prize competitions are becoming the standard ignition mechanism for new technology industries, with DARPA repeatedly converting contest cohorts into venture-backed sectors.