But the cost of delivering packages can be more expensive than delivering people, according to Ashley Nunes, a fellow at Harvard Law School who studies innovation. Revenues derived from moving human passengers in human-driven ride-hailing services can reach 97 cents per mile, he said, and that's a target that autonomous ride-hailing hopefuls want to emulate.
He cautioned that making comparisons to goods delivery can be complicated, because of fluctuating per-truck volumes.
There are operational considerations. People will expect a robotaxi to take the shortest, quickest route to their destination, Abuelsamid said. But package delivery has more flexibility. The delivery service can optimize the route for efficiency rather than speed. That could cut all the non-revenue, deadhead miles to pick up passengers.
"If you've got a vehicle that's doing 10 to 20 deliveries on one trip, you can plan a route so that the last drop-off is the closest one back to your distribution center," Abuelsamid said.
Other costs are associated with carrying passengers, like keeping the vehicle clean.
"You have to ensure nobody's getting sick in the vehicle," Abuelsamid said.
While Zoox may launch with passenger service, he said it wouldn't be long before they carry packages.
"They're going to use that technology in any way that benefits Amazon," Abuelsamid said.
Other analysts agree that Zoox will alter its strategy.
"They're definitely going after both ride-hailing and delivery," said Gary Silberg, who heads the automotive practice at the KPMG consulting firm.