Defense Disruptors: ‘Making Smart Things Smarter’ With Parry Labs

The F-35 fighter jet represents one of the most ambitious — and expensive — military programs in history. Yet for much of last year, the aircraft remained largely grounded, not due to mechanical failures but because of software integration challenges that have plagued the program for years.

John Parkes, CEO of Parry Labs, sees the F-35’s struggles as emblematic of a broader problem facing the Pentagon: how to rapidly deploy and update software across increasingly complex military systems.

“We had all the nuggets of the system that were needed,” Parkes recalled of his earlier work inside the Office of the Secretary of Defense for Research and Engineering, “but we could never get it instantiated on the platform because it wasn’t software-addressable.”

That experience led Parkes to leave government in 2016 and found Parry Labs, a company focused on what he calls “making smart things smarter.”

His mission: create the software infrastructure that allows everything from drones to tanks to ships to rapidly integrate new capabilities — turning military hardware into platforms as adaptable as a smartphone.

The stakes, according to Parkes, couldn’t be higher as adversaries advance their own technological capabilities while the US military struggles with legacy systems and vendor lock-in that can take years to modify.

Building Operating System for ‘Computers That Move’

Parkes described his company as building “a foundational software operating system for platforms and sensors” designed to accelerate how software gets deployed across what he calls “computers that move:” military systems that operate in contested environments with unique safety and security requirements.

The company’s product platform, called Stratia, aims to simplify software delivery into systems “that have safety criticality, cyber criticality, that have all sorts of networked interdependencies,” Parkes explained.

The technology spans multiple domains, from legacy drones like MQ-1s and MQ-9s to newer loitering munitions, unmanned ground vehicles, maritime platforms, and space systems.

“Think of it like what AWS did for the cloud — making software delivery simple and scalable,” Parkes said, referencing Amazon’s cloud computing platform. “We want to bring that same ease of integration to complex military platforms, but purpose-built for the safety, security, and mission speed requirements of the warfighter.”

The challenge, as Parkes sees it, lies in the complexity of modern military platforms. A single drone, for instance, might have separate computers controlling flight, managing networks and radios, processing camera feeds, and handling signals intelligence. Coordinating all these systems to respond to a simple command like “go look at that thing” has traditionally required extensive custom integration.

“Historically, that environment has not been made simple,” Parkes argued. “In fact, it’s almost been intentionally made complex because the people who own those complex systems were able to create long-enduring vendor lock.”

Demystifying Open Architecture

Much of the Pentagon’s current reform efforts center around something called Modular Open Systems Approach, or MOSA — a concept Parkes acknowledged can be confusing.

“MOSA is many things to many people,” he said. “If you’re in the government, it’s often an acquisition strategy or a policy. If you’re a technologist, it might be the way that you separate physical, logical, and data representations.”

At its core, Parkes argued, MOSA represents “a principled approach of how you interconnect systems” that allows diverse vendors to work together without requiring custom integrations for every combination of components.

The math, as Parkes explained it, is daunting without such standardization: “If you have four autonomy vendors, 20 payload vendors, and three platform vendors and you want to create a marketplace, the interconnectedness of all those systems becomes three times three times four times 20 times 10.”

That creates hundreds of different integration challenges that MOSA aims to simplify.

“It reduces the cost of innovation,” Parkes claimed. “It actually makes it lower cost and faster for us to be able to go and accelerate these new capabilities.”

What’s Working and What’s Not

Parkes credited Congress and military leadership with creating more flexibility for innovation. He praised acquisition authorities that allow services to create new programs under $1 billion without extensive Congressional approval, and highlighted well-funded innovation hubs like the Defense Innovation Unit and DARPA.

The challenge, he argued, lies in how the Pentagon structures emerging markets for software-centric defense systems.

“I don’t see them well defining oftentimes, how these companies can come together,” Parkes said, describing efforts to encourage “self-forming teams.”

“These efforts don’t always reward technical merit as clearly as they could, which can make it harder for innovative teams to break through,” he added.

The government, Parkes contended, needs to better define “major system components” so companies and investors “know where they can win, where they can lose.”

Without clear market definitions, he argued, “there’s no sort of end state of winning, and the investor community will ultimately need to see that.”

Redefining ‘the Edge’

Parkes admitted reluctance to use the overworked term “edge computing,” but distinguished his company’s focus as the “maneuver edge” — the computing environment inside military platforms themselves, rather than servers in command posts.

“The edge is also a tank, plane, ship, or satellite,” he explained. These environments face unique constraints, including mission safety requirements, contested communications, multiple radio systems, and sensors that must coordinate across many subsystems.

The complexity extends beyond individual platforms to networks of connected systems. “Tanks talk to loitering munitions, which talk to UAVs flying overhead,” Parkes noted. “You now have three systems that are all complex. And so how do you deliver software to all three?”

Security Without Compromise

Addressing concerns about cybersecurity in open architectures, Parkes rejected the notion that openness and security conflict.

“We certainly do not believe that open architecture and cybersecurity are conflicting,” he stated. “We think they’re very well aligned.”

He claimed Parry Labs implements “the latest zero trust standard” and incorporates “post-quantum encryption” concepts, along with precise time management across distributed systems.

The company, he said, treats robust cybersecurity as a prerequisite for the kind of data-sharing that open architectures envision.

The Economics of Speed

Beyond technical benefits, Parkes argued that modular approaches paired with commercial technology could deliver dramatic cost savings through three mechanisms: speed, competition, and reuse.

“When there’s a monopoly, people charge more,” he observed, contending that competitive markets for well-defined system components would drive prices down by “orders of magnitude.”

He also highlighted potential savings through software reuse across military services, something he claimed the Pentagon currently struggles to achieve.

“The same targeting AI software for counter-drone targeting as an example could be brought to five or 10 different weapons systems across the Department of Defense,” Parkes said.

“Right now, the government has had little ability to go cross-service and say, hey, this group needs it, that group needs it. Let’s just buy it once at the top.”

AI as Plumbing, Not Product

On artificial intelligence, Parkes positioned his company as focusing on infrastructure rather than algorithms. “Parry Labs is focused really on the first principles of AI, which is having strong data understanding, being able to deploy software against that data,” he explained.

Drawing another comparison to commercial technology, Parkes said, “Much like Palantir focuses on providing the plumbing for AI rather than the algorithms themselves, Parry delivers that same kind of infrastructure for the edge — enabling faster, more reliable autonomy for the warfighter.”

He predicted that many AI algorithms would become commoditized within continuous delivery pipelines, making the real challenge “how do you make the data and the software delivery for autonomy as fast as possible?”

From Pentagon Frustration to Silicon Valley Solution

Parkes traced his journey from engineering at Goodrich Aerospace through corporate development roles before joining the Pentagon’s Special Capabilities Office, where he worked on urgent operational problems, including counter-IED efforts, by combining technologies from across the intelligence community and DARPA.

The recurring obstacle, he recalled, was getting promising technologies actually deployed on operational platforms.

“Oftentimes, we would have all the nugget of the system that was needed, but we could never get it instantiated on the platform or the system that was needed because it was vendor locked or because it wasn’t software addressable,” Parkes said.

That experience drove him to start Parry Labs, initially working with Special Operations Command on drone programs.

The company doesn’t build weapons or autonomous systems itself, Parkes emphasized, but rather aims to “unlock that entire ecosystem to be accelerated” across air, ground, maritime, and space domains.

His bet is straightforward: if the Pentagon can treat its most complex machines more like software platforms, it can move faster, buy smarter, and field capabilities at the speed of modern conflict.

“We’re really just trying to make that whole ecosystem simpler, easier, and more quickly addressable by software,” Parkes concluded.

As military systems become increasingly software-defined and adversaries accelerate their own technological development, the question isn’t whether the Pentagon will embrace more flexible approaches to system integration but whether it can do so fast enough to maintain its technological edge.