When it comes to autonomous vehicles, you can move things, or you can move people without drivers. The logistics applications already are way ahead of the human ones, but May Mobility is sticking with moving people.
And the Toyota-backed startup founded by Edwin Olson has moved into the vanguard of the tricky business of transporting human beings autonomously, where the promise is bright but the reality remains a slog. The company based in Ann Arbor, Mich., has achieved near-unicorn status after just four years in existence by autonomously operating shuttles that are integrated with public-transportation systems in a handful of cities.
Even so, May’s narrative remains one of the tortoise and the hares so far — with the hares being flashy autonomous-driving champions such as Tesla. Elon Musk’s startup is sui generis, but Tesla has been promoting the autonomous-control features of its vehicles before they’re ready for prime time, and some fatalities have resulted from drivers’ over-reliance on those capabilities.
“The hare can be flashy,” May Mobility CEO and co-founder Edwin Olson told me. “People get excited about the story the hare is telling. But at the end of the day, we’re here literally to transform cities, to make cities work better and to provide access with equity, that sort of thing. We’re excited about the impact we’ve had already.
“We have delivered more than half of the revenue-generating rides across the entire [AV] industry,” Olson said. “All of them are on public roads in mixed traffic; we’re not focusing on low-speed parking-lot trams. All of ours are street-legal vehicles driving with traffic and dealing with the full customer experience.”
May is pursuing people-moving even though the quickest and broadest commercial applications of driverless technology are coming in moving goods instead. “There’s a greater business opportunity with people,” Olson said. “People-moving is generally a higher-margin business. People are the most valuable things there are, and they care how long they’re waiting, whereas if your grocery delivery is delayed by five minutes, it’s no big deal. With logistics, autonomy doesn’t have to work as well.”
So far, May’s clients are mainly “municipalities trying to solve transportation problems” in cities including Arlington, Texas; Indianapolis; Ann Arbor; and Hiroshima, Japan. In Hiroshima, the company is operating vehicles on a 1.9-mile route on the campus of Hiroshima University. Additionally, in Grand Rapids, May operates a 3.2-mile fixed route, and local employers “have kicked in to help provide transportation because it benefits their employees,” Olson said. May also participated in pilot projects in Detroit, Columbus and Providence.
At this point, all of May’s 30-some shuttles in operations so far are backed up by a human “safety driver” on board in an application of “Level 4” autonomous driving as defined by the U.S. government; “Level 5” would be completely driverless operation.
“Our system is designed so there’s no expectation that the safety driver will take over,” Olson explained. But he or she is there just in case. “We’ve been proceeding on plans to remove safety drivers. Some other companies have talked about much more aggressive plans for doing that, which haven’t come to be.”
Olson is an MIT graduate who led a University of Michigan winning team in a 2010 mapping competition and worked with Ford on AVs before joining the Toyota Research Institute in Ann Arbor. Toyota staked May to an initial $50 million in funding in 2019, and May has raised about $86 million through Series B, including funding from Toyota AI Ventures, owned by the automaker and my Maven Ventures, whose portfolio also includes General Motors’ majority-owned AV outfit, Cruise. May adapts Lexus RX450h hybrid SUVs and wheelchair-accessible Polaris Gems for its current vehicles and will add an AV based on the Toyota Sienna minivan.
May’s differentiator, Olson said, is that “we give the vehicle an imagination instead of trying to program it in advance. What would happen, for example, if [the vehicle] yielded to a pedestrian or another vehicle blew through a yellow light? By simulating 10 to 15 seconds forward, we can see the likely outcomes of those situations, in scenarios that we’ve never seen before. That looks a lot more like how people drive.”
Olson said May’s simulators benefit in this way “from a form of reinforcement learning, allowing us to build our system faster. We don’t need to have thousands of engineers coding in.”
He dismissed some pundits’ concerns that a thorough autonomous-transportation system for humans could be dystopian because of how it might allow for centralized control of access.
“Today we largely have a two-tier transportation system: one for people with means, the other for people who use public transit,” Olson said. “Because it’s for the poor, it ends up being not very well funded and usually very poor-performing. That leads to a dyistopian world now. We have two systems that don’t support each other but are intentioned for same road geometry, and that leads to bad outcomes for everyone..
“The way that you get to a utopian ideal is to create public transit that’s so damn good that people who use other options will use it anyway,”he added. Plus, “If people in the system are happy, you’re creating one system that makes far better use of real estate in the city.”
“How make it great experience? If people in system the’rey happy and youre’ creaeting one system that meakes far better use of RE in city.”
In the meantime, Olson’s strategy has led to May’s status as “the only AV company that generates positive gross margins,” he said. Positive net income, Olson said, is “a few years out,” even as May Mobility’s valuation nears the $1-billion status that defines a unicorn.
Olson declined to specify the company’s current valuation. Olson said that May so far hasn’t generated “huge amounts of revenue. But the point is to demonstrate that our technology works and that we’re delivering real value for customers. Scale will come when these vehicles start coming off assembly lines at high volume.” Olson wouldn’t say when that will happen.