He spent thirteen years shipping electronics into orbit. Then he went looking for the least glamorous machine in America and decided to make it drive itself.
Picture a freight car with no locomotive in front of it. No diesel engine, no crew, no long chain of wagons waiting on a single set of brakes. Instead it rolls under its own power, battery-electric, and when it meets others like it they lock together into a small pack - a platoon - to save energy. Then, mid-route, one peels off toward a different destination while the rest continue on. That is the machine Matt Soule has spent since 2020 building at Parallel Systems, and it is the reason a company most people have never heard of has quietly raised more than $91 million.
Soule runs Parallel from Los Angeles, a city that worships the automobile and treats the railroad as a relic. The irony is deliberate. His argument is that the American freight network already exists - roughly 140,000 miles of track sitting underused while trucks clog the interstates - and that the missing piece was never more steel. It was autonomy, batteries, and a different idea about scale.
Most of the rail industry chases bigness: longer trains, heavier loads, more cars behind one locomotive to spread the cost. Soule went the other way. Parallel's railcars are individually powered and individually smart, which means the economics work at small scale, on short routes, for the kind of loads that today ride on trucks. "We're using a different physical architecture to accomplish truck-competitive economics at small scale rather than big scale," he has said. It is a sentence about physics as much as business.
By mid-2026 the idea had left the whiteboard. Parallel was running a commercial pilot near the Port of Savannah under Federal Railroad Administration supervision, working with short-line operator Genesee & Wyoming, with a backlog of more than 300 railcars on order and a third-generation vehicle in development for full commercial service.
“We are much closer than most people realize. Autonomous freight is shifting from experimental testing to commercial revenue generation right now.”Matt Soule
Soule is an electrical engineer by training - a bachelor's from Northwestern, a master's from the University of Southern California - and by temperament. Before Parallel he spent about two decades around systems that punish sloppiness: laser weapon control loops at Northrop Grumman, electronics work at Lockheed Martin and Moog, and then the long haul at SpaceX.
At SpaceX he worked in avionics, the electronics and software that keep a rocket pointed the right way and alive under enormous stress. He rose to lead the avionics department, running a group of roughly 300 engineers designing and testing the electronics that flew. It is the kind of job where a bug is not a support ticket. It is a launch.
That background matters because freight rail, for all its 19th-century image, is fundamentally a controls-and-sensors problem once you strip out the human crew. Cameras, lidar, adaptive braking, continuous sensing, software that decides when to couple and when to let go: this is avionics vocabulary applied to steel wheels. Soule did not wander into rail as a hobbyist. He arrived with the exact toolkit the problem demanded.
The founding logic, by his own account, came from curiosity as much as opportunity. After years at the frontier of aerospace, he went looking for an industry that advanced technology had barely touched - and freight rail, one of the oldest networks in the country, was hiding in plain sight. In 2020 he started Parallel Systems with a cluster of fellow ex-SpaceX engineers who agreed the railroad was overdue for a rethink.
Short-haul freight - the trips too short for today's mile-long trains but long enough to matter. Parallel wants to convert a slice of the roughly $700 billion U.S. trucking market to rail, and has talked about cutting port trucking volume by as much as 30 percent.
Self-propelled railcars that platoon to save energy and split to reach multiple destinations, without a locomotive and without a crew. Autonomous coupling replaces a manual task so dangerous it has historically injured and killed rail workers.
Cleaner air, less highway congestion, and a freight system that behaves less like a scheduled train and more like a network. Battery-electric power means zero tailpipe emissions on routes that trucks run on diesel today.
Ask Soule what success looks like and the answer is not a single gadget. It is a network - autonomous railcars talking to autonomous trucks, robotic warehouses and AI routing, all handing cargo off to whichever mode is cheapest and cleanest for each leg. The railcar is the first piece.
It is a patient bet. Rail is regulated, capital-heavy, and slow to trust newcomers, which is why the FRA pilot and the partnership with an established short-line operator matter as much as the technology. Soule is not trying to replace the railroad. He is trying to give it a nervous system.
The rocket engineer's instinct shows through in the framing: build the hard physical thing correctly, prove it under real oversight, then let the network effects compound. Whether freight rail is ready to become the internet of freight is still an open question. But for the first time in a long time, someone with the right toolkit is asking it in earnest.