The DARPA challenge is somewhat of an orthogonal use case. The DARPA challenge was based off of an autonomous vehicle that doesn't have extensively pre-mapped roads; it is in rough terrain. The vehicle had to alter its route on-the-fly if it encountered impassable obstacles. It also didn't have to concern itself with traffic. I presume the use case is getting materiel to troops without risking human drivers. This is important considering IEDs were a major source of injury and death for coalition troops in the most recent wars.

Google/Waymo and its successors are using pre-mapped courses with many heuristics and edge case tweaks. Routes are generated from existing resources (Google Maps, etc). Much effort is devoted to avoiding other vehicles and pedestrians. Much of the rules are based off of U.S. traffic rules, such as speed limits, stop signs, traffic signals, and lane markings.

They both share technology (computer vision, momentum/traction control), but I conjecture the bulk of the work for commercial autonomous driving was not related to the DARPA challenge and wasn't paid by its grants.

That was true for the first two Grand Challenges, but once that prize was claimed DARPA simply cranked up the difficulty level; the 2007 race was all about navigation in urban environments, and compliance with traffic laws was a condition of success.

Sure, of course you're right that "the bulk of the work" is not related. But it's also no secret that the "sudden spike in interest in this technology" lucker referred to above happened because the US government paid for it to happen, as a means of advancing military vehicle automation technology.