The 99% on-time delivery rate doesn’t mean the system works; it just means the routing algorithm finally stopped leaving the efficiency dividend with the people who actually drive the route. PepsiCo is running thirty-five driverless trucks on Arizona roads — a twenty-six-thousand-pound box truck loaded with Doritos, no one at the wheel, 99 percent on-time, no accidents. The engineering is impressive. The 99% figure is real. Point-to-point freight with fixed loading docks is exactly where lidar and computer vision earn their keep. The trucks learn the routes, they don’t fatigue. I’m not anti-machine; I’m anti-extraction, and those are entirely different things.
Jim Farrell, PepsiCo’s supply-chain senior vice president, said the quiet part out loud: “One of the things that we can do is be able to grow the business without having to add as many employees.” Some drivers will be retrained to manage the new equipment; others will handle unloading. The net direction is unmistakable — a smaller slice of the workforce, doing a larger volume of work, while the productivity gain flows up. The International Brotherhood of Teamsters is lobbying state legislatures to require human operators in autonomous commercial freight, and the FedEx Freight CEO’s own admission that regulation lags behind the technology confirms the offensive reality: the routing contracts are about to concentrate.
The suppressed variable is who gets the check for the precision. When you strip forty thousand annual labor hours out of a fixed delivery network, that money doesn’t evaporate. It moves. The American default sends it straight to the corporate balance sheet, then offers the displaced driver a polite brochure about retraining and redeployment. That isn’t a transition plan. It’s a waiting room with a better printer. The shareholder who bought the Isuzu chassis and the software license pockets the gain; the worker absorbs the friction.
Denmark solved the exact problem by protecting the worker, not the job. They fire people more easily than most American states. But when a Dane loses her job to automation, the system pays her most of her salary for up to two years and spends roughly two percent of GDP on active labor-market programs. The productivity gain is captured by the firm; the adjustment cost is socialized. The security comes from the safety net, not from trying to outlaw the steering-wheel-less cab. We don’t have that machinery here, but we have the starting material. The recent union drive by ride-hailing drivers in Massachusetts shows the defensive line is already being drawn from the bottom up.
An autonomous mile is not oil, but it is a productivity windfall that was not created by PepsiCo’s existing shareholders alone — it was built on decades of public investment in roads, research, regulatory frameworks, and the workforce that trained the engineers. There is a version of this story where every autonomous mile PepsiCo’s fleet logs deposits a tiny fraction — a cent a mile, a tenth of a percent of the transportation cost saved — into a public or worker-owned fund. The fund pays retraining, income support during transitions, or an actual dividend to the drivers who made the system work while they were in the cab.
The build is ownership. Look at the rural electric cooperatives that wired the American South. They didn’t wait for distant investor-owned utilities to decide whether a poor county was profitable enough to run a line. They owned the wires, member by member. A logistics cooperative does the same for freight. Drivers, municipalities, and local shippers pool the routing contracts, buy the autonomous fleet outright, and keep the efficiency dividend paying local wages and funding retraining locally, instead of mailing it to a quarterly report. The capital is just as patient as the tech. The math closes.
When the Fagor cooperative collapsed in 2013, the Mondragon network absorbed or retrained roughly 1,700 displaced members. If a Basque worker co-op can absorb its own displaced people, a soft-drink company that posted $11.5 billion in operating profit for fiscal 2025 can figure it out. The steering wheel is obsolete. The bargaining table isn’t. We can run a driverless network that treats the displaced driver as a line-item to be managed, or we can run one where the efficiency gain pays for the transition. The burden is on the people who made the choice to defend it — or, better, to explain why the workers who built the system that made the autonomy possible have no claim on what the autonomy leaves behind.