Peter Howard plans robot paths for a living.

The Boston office is planning your factory

Somewhere in a Fortune 100 automaker's plant, a robot arm is following instructions that a Boston server generated a few milliseconds ago. The arm does not know this. It also does not know that the path it is following did not exist before it was needed, and that a different path would have been generated if the operator standing near it had happened to shift a shoulder. Peter Howard runs the company that generated the path. He calls this "motion planning." Everyone else in the industry has, for about four decades, called it "programming the robot," which is roughly like calling GPS navigation "unfolding the map."

Howard is the co-founder and chief executive of Realtime Robotics, headquartered at 27 Wormwood Street in Boston's Seaport district, staffed by roughly fifty people, and funded to the tune of about $82.9 million as of a Series B round completed in May 2024. The company's core product does something that sounds trivial when described and turns out to be extremely hard when attempted, which is generate collision-free paths for multiple robots working in the same physical space, and do it fast enough that the robots can react to each other and to humans and to obstacles as those things move. The word Realtime is not marketing. The industry standard, for a very long time, was to write a robot's motion sequence by hand, test it, adjust it, redeploy it, and repeat the process anytime the workcell layout changed. That process takes weeks. Realtime's software takes hours. This is the entire pitch, and it is enough to have gotten the company into some of the largest manufacturing lines in the world.

The call from Duke

The origin story is the sort of thing that happens to serial entrepreneurs who have spent three decades collecting phone numbers. Roughly a decade ago, Howard got a call from Duke University. Researchers had come up with a new approach to motion planning, one that used specialized hardware to compute collision-free paths orders of magnitude faster than conventional methods. Would he want to look at it. He flew down. He met the inventors. He wrote the first check.

He also had, at that point, some prior context. Fifteen years earlier, he had run a sensor-technology venture that worked with BMW on quality control for the X5 production line, which is another way of saying he had once already been in a room with a car manufacturer trying to figure out how to make robots see what was in front of them. The technology stack had changed. The question had not. When Duke called, he already knew what problem was worth solving.

Five companies, one operating principle

Realtime is Howard's fifth company. Two of the prior four went public. One was sold. One licensed out its core technology. The through-line, according to Howard, is not sector. It is the moment when a technical breakthrough exists but the business around it does not, which he describes as "formative chaos." He seems to enjoy this stage more than the running-a-mature-business stage, which is why he tends to leave.

His credentials for doing so are an MS in Management from MIT and a Professional Director Certification from the American College of Corporate Directors, which is a very specific credential to hold and suggests a founder who takes board work seriously. His pre-Realtime career included leadership at outsourced R&D and manufacturing services companies with operations in the United States, Japan, France, the Netherlands, Singapore, and China. This is a background that shows up in the operating detail of Realtime, which sells software to industrial customers who are unforgiving about downtime and require the sort of contractual seriousness that first-time founders often learn about the hard way.

The $15 lawnmower

The anecdote Howard tends to tell about himself involves being eight years old and borrowing $15 from his father to buy a used lawnmower, then paying his father back with interest out of mowing fees. The lesson his father was trying to teach was: you can spend money you earn on anything, but you have to earn it first. Whether or not this is the actual origin of Howard's approach to capital, it is what he says it is. Founders tell the stories they need to tell. The fact that he still cites the loan in interviews decades later suggests either the loan really did shape him or he has, at minimum, chosen to be shaped by it in public, which for the purposes of running a Series B robotics company amounts to about the same thing.

"This changes everything"

The demo Howard tells reporters about took place in the MassRobotics lab in Boston. A senior executive from a major robot manufacturer was in town. Realtime set up its multi-robot cell, ran a task, and invited the visitor to interfere physically with the robots. The system rerouted in real time around the intrusion. The executive is said to have used the phrase this changes everything, which is exactly what one hopes a robot manufacturer's CTO will say and is otherwise the sort of thing you rarely get on the record. Howard has repeated the story in interviews, which is why we have it. The MassRobotics demo has since become part of the company's origin mythology, one of those cleanly memorable events that founders return to when explaining, without saying it directly, that the product actually works.

What the software actually does

To describe Realtime Robotics as motion planning is technically accurate and mostly unhelpful. What the software actually does is take a factory workcell containing multiple industrial robots, model it in a cloud-based simulation environment, generate the paths those robots need to take to accomplish their assigned tasks, coordinate the paths so the robots do not collide with each other, insert safety interlocks so they do not collide with humans, and then optimize the whole thing for cycle time. Some of this happens before the cell is turned on. Some of it happens continuously, in operation, as objects move and conditions change. The customer-facing effect is that the cell can be redesigned and redeployed in hours rather than the weeks a hand-programmed cell would require, and that the redesigned cell tends to run faster.

The customer list, by industry, tracks with where flexible high-volume automation matters most: automotive, general manufacturing, machinery. The technology stack, according to the public record, involves cloud AI tools, simulation integration with tools like Process Simulate and Visual Components, Python and C# on the software side, and a healthy amount of internally developed algorithms for path generation and collision avoidance.

Career, briefly, in the format that CEOs prefer

A few unrelated facts

What he wants

Howard's stated professional ambition is to make deploying industrial robots as easy as setting up a PC, which is a large ambition and also the sort of thing that, in robotics, has been credibly said out loud by very few people. His stated non-professional ambition, offered in an Authority Magazine interview, is to see a constitutional amendment overturning Citizens United. This is a specific policy view for a CEO to volunteer in a business Q&A. He volunteered it anyway.

Frequently asked