Steve Messinger had just quit his job.
Not because anything had gone wrong. Applied Intuition, the autonomous vehicle company where he worked, was well-funded and well-regarded. His career was moving in the right direction. He quit because his college friend Jason Cornelius was in a 900-square-foot apartment in San Francisco running customer interviews with the military and building rocket prototypes by himself, and Messinger had decided it was time to stop watching from the sidelines.
He put $10,000 into the company. Packed his belongings into a closet in Cornelius’s apartment. One moved onto the couch. The apartment had one bathroom and a roommate. The combined founding capital of what would become Perseus Defense was $30,000.
That was the spring of 2025. Within a week of Messinger’s arrival, the two had built a rocket and launched it from a drone. They brought the hardware to their Y Combinator interview, set it on the table in front of the partners, and described what they had built in seven days. YC offered them a spot.
Today Perseus Defense operates out of a 17,500-square-foot warehouse in Buda, Texas, just south of Austin. The team has grown to ten full-time engineers. The company has raised $6.5 million. In January of this year, they achieved fully guided flight with their micro-missile and have been doubling down on direct hit to kill ever since. They are now launching five to ten missiles per week out of that warehouse, refining the design with each flight, and preparing for a live-fire demonstration with 1st Cavalry Division later this year.
Not long ago, Messinger received a text from a soldier who was preparing to deploy to the region.
The soldier wanted to know when the product would be ready.
It was the kind of message that clarifies, quickly, whether you are working on the right problem.
THE KIND OF PEOPLE WHO QUIT GOOD JOBS
Messinger and Cornelius met at Penn State as graduate students in aerospace engineering, though they worked on different sides of the discipline. Cornelius was pursuing his PhD in rotorcraft aerodynamics and machine learning-based aerospace vehicle design optimization. Messinger was completing his master’s in guidance, navigation, and control, the field concerned with how an aircraft moves its actuators in real time to reach a target and stay on course when things go sideways.
After Penn State, their paths split. Cornelius went to NASA Ames, where he led vehicle design for the Dragonfly mission, a rotorcraft that NASA is sending to Saturn’s moon Titan. He developed machine learning methods for optimizing rotorcraft design for future Mars missions. He created an international drone competition. He began teaching aerodynamics at Stanford.
Messinger took a different route. He joined Boeing’s Aurora Sciences subsidiary and worked on autonomous landing systems for commercial airliners, specifically for scenarios in which one or more pilots become incapacitated mid-flight. He then moved to an 80-person company developing autonomous boat swarms for the Navy, which was acquired by Applied Intuition not long after he joined. It was at the Applied Intuition acquisition event, over dinner in San Francisco, that he and Cornelius started talking seriously.
Around the same time, the two began sitting in on a Stanford course called Hacking for Defense, giving student teams feedback while quietly taking notes for Perseus. The experience sharpened their approach to customer discovery. By the time Messinger quit his job and showed up at Cornelius’s door, the interviews were just beginning to ramp up. They have now conducted hundreds of conversations across the joint force. One early idea that came into focus was a micro-missile launched from a vertical takeoff drone. The interviews killed it, at least the drone part of it.
“Don’t build another drone. We already have 100 million of these things. Focus on the counter.“
That feedback, delivered by end users who would actually have to operate whatever Perseus built, shaped everything that came after. The pivot happened before there was a company, before there was an office, before Messinger had moved onto the couch. By the time Perseus formally existed, its founders already knew what it needed to build.
A $250,000 ANSWER FOR A $500 DRONE
The arithmetic of the modern battlefield is not complicated. In the Ukraine-Russia conflict, drones now account for somewhere between 70 and 80 percent of battlefield fatalities. They are small, fast, cheap, and proliferating faster than any defense procurement cycle can track. Group 1 and Group 2 systems, the class Perseus is targeting, are essentially commercial quadcopters and small fixed-wing platforms carrying explosive payloads. They cost between $500 and $10,000.
The United States military’s primary kinetic solution for intercepting them is the Coyote, which costs roughly $250,000 per unit. Directed energy weapons remain years from reliable fielding at scale. Electronic warfare countermeasures work until adversaries adapt their frequencies and flight patterns, which they now do routinely. Last year, a Ukrainian commercial drone destroyed a Russian strategic bomber worth more than $100 million. The problem is not that the military lacks sophisticated weapons. It is that the weapons it has are the wrong ones for this particular fight.
Perseus is building a 24-inch guided micro-missile designed from the start around the counter-UAS mission. Each missile is priced at under $10,000. Each pod carries fifteen of them. The system is portable and configurable for launch from UAVs, ground vehicles, boats, and dismounted infantry. A low-collateral variant is in development for use in populated environments.
With a range exceeding 1,000 meters, Perseus targets the same Group 1 and Group 2 systems that have reshaped what ground combat looks like. At the current price point, a unit commander can engage a drone swarm without doing the mental math on whether the engagement is worth the cost. That calculus, right now, often leads to waiting too long or not engaging at all.
Every design decision inside the company is evaluated against one constraint: can this be manufactured at scale. The internal production target is 10,000 missiles per month. Perseus is not building a prototype and then figuring out how to make it. It is building for the factory from the first sketch.
“OCTOBER SKY” SPEED
I first heard of Perseus Defense from a Harvard Business School classmate who had worked with them during their time at YC. This week was the first time I actually spoke with Steve Messinger. He is exactly what you would hope for in a founder building something this consequential: unhurried, precise, and completely unbothered by how hard the problem is.
The conversation brought back a memory I had not thought about in years. Like a lot of kids who grew up obsessed with space, I was shaped by October Sky, the story of Homer Hickam, a coal miner’s son in Coalwood, West Virginia, who watched Sputnik cross the sky one night in 1957 and decided he was going to build rockets. Homer and his friends launched again and again, failed repeatedly, learned from every failure, and eventually earned their way out of the coalfields and into careers as engineers and scientists. It is a story about what American ingenuity looks like before it gets organized into something slower.
Most days it feels like that version of America is harder to find. The bureaucracy built around the last generation’s successes has a way of crowding out the people who just want to build things and see if they work. Talking to Messinger made that feeling go away for a while.
