A Toyota Sienna rolled to a stop in the parking lot of the University of Michigan's "Big House" after the minivan's cameras, sensors, and lidar detected an oncoming vehicle in the other lane.

The interesting part is what happened next: The Sienna, owned by autonomous vehicle startup May Mobility, seemed to recognize that although the other vehicle had the right of way, its driver deferred to the minivan. After a moment, the May vehicle turned left in front of the stopped car.

The move was received with enthusiasm from May Mobility CEO and founder Edwin Olson, who was Tech Brew’s tour guide during a recent demo ride from May’s production facility in Ann Arbor to downtown and back.

“We were able to change what we were going to do by running the simulations and learning what the right thing to do in that situation was,” Olson said. “And we did that by running thousands of simulations right there to figure out all the permutations.”

It’s an example of May’s proprietary AV system, “Multi-Policy Decision Making,” in action.

Real-time learning

Our ride’s soundtrack was the whirring of a fan from a computer situated on the front passenger side.

That computer works hard during each trip, running thousands of simulations per second, Olson said.

“This is one of the main differences about May’s technology: We don’t tell the car what to do,” Olson said. “We give it all the best strategies that we can learn offline, but then allow it to learn what’s actually best for the situation as it’s driving around.”

As we came to a four-way stop, a car approached the stop sign on the right. The May vehicle waited to ensure the other driver hit the brakes.

It’s “very likely,” Olson said, “that that specific timing was not in our training data. But the computer was able to run lots of simulations saying, ‘I think they’re probably going to stop, but they’re going pretty fast. I think there’s a 20% chance that they might blow through the stop sign.’”

Go-to-market strategy

May, which was founded in 2017 and whose backers include Toyota and BMW, has completed more than 13,000 rides in Ann Arbor. Its customers are not individuals, but businesses and local governments: It has contracts with numerous other municipalities, including Grand Rapids, Minnesota, and Arlington, Texas, where it provides microtransit services aimed at bolstering the reliability of public transit.

The startup also started offering driverless rides in Sun City, Phoenix, earlier this year. The startup has a fleet of 35 vehicles, with plans to grow this year.

May’s mission, according to Olson, is to “outcompete personal car ownership” by providing more reliable, accessible, and efficient public transit that users can hail on demand. May is competing in this segment because its executives see better revenue opportunities compared to other AV applications like robotaxis.

“The goal of most cities is, can we provide the transit?...We just want the transit to get delivered—which is a great situation for us, because it makes public transit one of the highest revenue opportunities in all the AV go-to-market spaces,” Olson said. “It’s better than trucking; it’s better than ride-hail; it’s better than middle-mile.”

Another key part of May’s model is accessibility: It gives municipalities the option to include as many wheelchair-accessible vehicles in their fleet as they want; users can select those vehicles when requesting a ride.

Olson predicts that May could be the first AV company to reach profitability, with a current target of around 2027. The company has more than $300 million in venture-capital funding, according to Crunchbase. The startup cut 13% of its workforce earlier this year following the closure in November of a Series D funding round led by Japanese telecom firm NTT Group.

May is taking a measured approach to scaling up its business, Olson said. One of its next moves will be to launch a shuttle service under a $2.4 million contract with the city of Detroit.

Eventually, the goal is to expand May’s reach into other geographic areas and market segments.

But for now, the focus is on incremental advancements—and doing so with positive margins. The startup also is eyeing an expansion in the next couple of years across the percentage of its fleet that operates without a safety driver, from 5% to 75%.

“We want to make public transit so good that people who can afford cars and people who can’t afford cars choose to use public transit anyways,” Olson said. “Because that’s going to make cities safer, more accessible.”