The Brompton, Sashimi and Unmanned Fighters

Defence Tech · Field Notes

What a retired Air Marshal told me over fatty tuna about why the manned fighter jet is finished — and what defence tech founders should be building instead.

I was having sashimi with a retired Air Marshal in central London when he casually dismantled everything I thought I knew about air power. Ed Stringer — former head of Joint Ops at the UK's Permanent Joint HQ — had cycled across on his Brompton, sat down, ordered the fatty tuna, and within about ten minutes had walked me through an argument that I genuinely haven't stopped thinking about. We were supposed to be talking about working together on some defence tech projects. What I got instead was a first-principles demolition of modern fighter aviation over some very good egg fried rice and an Italian red that had no business being in a Japanese restaurant. I'm sharing it because I think every defence tech founder and executive needs to hear this.

I should tell you what kind of guy Ed is. He's not one of these retired officers who does the after-dinner circuit telling war stories. He's the kind who casually mentions he used to fly reconnaissance missions at 500 knots — that's roughly 600 miles per hour — low enough to see people on the ground, and then just moves on like it's nothing and starts explaining why everything about the way we build fighter aircraft is economically insane. Between slices of toro and the fried rice, he took me through the maths. And the maths is devastating.

Why manned fighter jets are becoming economically indefensible

Here's the thing. A manned fighter jet carries an enormous amount of engineering overhead for one reason: keeping the pilot alive. That means oxygen generation systems, fire suppression, ejection seats rated for zero-zero escape, triple-redundant hydraulics and electrics, and every single component tested to a failure probability of 10⁻⁷. That last number is important. It means months on the test rig before a single system gets cleared for flight. Months. For each system. Just think about that for a second.

Then there's the airframe. A modern fast jet is engineered to last 7,000 flying hours across a 30-year service life. Not because any war demands that — but because you need those decades of peacetime flying to justify the obscene cost of training pilots. Every fastener, every structural joint, every composite panel is designed for longevity. The aircraft isn't built for combat. It's built for the training programme that precedes combat. That's mad when you actually think about it.

Ed put it simply. Strip the human out, and you strip all of that. One hydraulic system. One electrical system. Designed to fly — not designed to survive a certification regime that takes years. And that's where it gets really interesting.

What happens when you design a fighter that was never going to carry a human

This is the critical distinction and most people miss it. I'm not talking about a manned platform with the cockpit removed. I'm talking about something conceived from day one as uncrewed. Designed from a blank sheet. No ejection seat. No canopy. No environmental control system. No peacetime safety certification consuming years of engineering before the thing ever touches a runway.

An unmanned combat air vehicle designed from scratch needs to fly for maybe 200 to 300 hours. That's enough for any plausible campaign. The airframe can be designed for pure performance and then thrown away. Fasteners bonded rather than bolted. Reliability standards drop by orders of magnitude. You're not building something to last 30 years. You're building something to win.

This is where Ed's argument gets uncomfortable for the establishment. He isn't making a capability argument. He's making a maths argument. And the maths doesn't care about your feelings or your Air Force culture.

The fleet-vs-fleet attrition maths that changes everything

Right, so this is where it gets wild. The question everyone keeps asking is whether an unmanned fighter can beat a manned one in a dogfight. That's the wrong question. Like, fundamentally the wrong question. The right question is what happens when you model fleet against fleet over a full air campaign — with repair cycles, attrition rates, and replacement timelines factored in.

Ed walked me through it over the egg fried rice and I swear my chopsticks were just hanging in mid-air. The maths is so simple it's almost offensive.

Assume the cost reduction gives you a 4:1 ratio. Four unmanned platforms for every manned fighter. Now assume the manned platform is genuinely superior in individual combat — give it a 2:1 kill ratio. Two unmanned kills for every manned loss. Pilot skill, better sensors, whatever you like. Be generous.

Run the campaign. You have 100 manned fighters. I have 400 unmanned. Even at your 2:1 exchange rate, you've exhausted your fleet while I still have 200 aircraft in the fight. And I can manufacture replacements in weeks. You can't build an F-35 in weeks. You can't train a pilot in weeks. Game over.

"Don't get into the game of asking whether your unmanned fighter beats their manned fighter one-on-one. Do a war game. Fleet versus fleet. Model the campaign with repair, recuperation, replacement. That's where the answer lives." — Air Marshal (Ret.) Edward Stringer

I sat there thinking: that's the entire argument. Right there. It's not about who wins the dogfight. It's about who runs out of aircraft first. And the side that can produce cheap, disposable, high-performance unmanned systems at scale will always — always — outlast the side that's sinking billions into platforms designed to keep one human alive for 30 years.

Why the loyal wingman is a stepping stone, not a destination

The loyal wingman concept pairs an unmanned drone with a manned fighter like an F-35 or Typhoon. The drone acts as an expendable escort — it carries sensors or weapons, and critically, it absorbs the missile so the pilot doesn't have to. Ed is realistic about where defence establishments currently sit. The loyal wingman is the concept they're comfortable with because the pilot remains in command, the human stays in the loop, and the Air Force's entire culture — which is built around pilots, right? — stays intact.

But he's explicit about it. It's a halfway house. The real step change comes when you design from a blank sheet for an aircraft that was never, under any circumstances, going to have a human anywhere near it. That's when you unlock fundamentally different design constraints, performance envelopes, and production economics.

The loyal wingman is the door through which autonomous mass enters the force structure. But it's not the room.

If you're a defence tech founder working in autonomous systems, this distinction matters. Build for the transitional doctrine and you get short-term contracts. Build for the end state and you get a company.

The dual-use trap that kills defence tech companies in peacetime

I asked Ed about dual use — because everyone in defence loves saying it. His answer was blunt. Who keeps these companies alive in peacetime? Because "somehow" is not a business model.

And he's right. I've been spending time with early-stage defence companies over the past few months and the pattern is the same every time. The tech is incredible. The founders are brilliant. And the commercial pipeline is a slide in a deck. That's not dual use. That's a funding gap with a civilian veneer.

THE HONEST QUESTION..

If your defence tech company's unit economics require a war to work, you don't have a dual-use business. You have a capability demonstration with a funding gap. Be honest about whether your commercial positioning is a genuine revenue strategy or narrative convenience for investor decks. Because investors will figure it out eventually — and by then you'll have burned through the runway.

Ed's been involved in this since 2018, including early work with Improbable on military applications of synthetic environments. He's watched the cycle play out. Extraordinary technology, brilliant people, no peacetime revenue model. His point is that without real commercial traction, you're asking investors and governments to fund a standing capability that only becomes valuable when the shooting starts. History suggests that doesn't end well.

Why the military's best people are leaving — and why that's your hiring advantage

Here's the thing that genuinely struck me. Ed basically said it's now more patriotic to leave the military and join a defence tech company than to stay inside an institution that's going to marginalise you the second you fall off the promotion track. Coming from a retired Air Marshal, that's a hell of a statement.

The promotion system funnels all high-performers into a single race for command. Those who don't make each cut get pushed out. Those who remain are — by definition — the ones who weren't selected. Ed's words, not mine: second-division players carrying on everywhere else. The result is a tiny number of senior leaders, a large number of frustrated people who were told they were the best right up until they weren't, and a system that actively ejects its most capable talent.

He referenced four officers who all independently reached the same conclusion and left the Air Force. Two of them he'd had long conversations with before they made the decision. And they're now doing more for national security from inside defence tech companies than they could ever do fighting for a desk job inside the institution.

For defence tech founders, this is a direct talent acquisition thesis. The military is producing a steady supply of exactly the people you need — operationally experienced, technically literate, frustrated, and looking for meaning. But here's the thing that nobody tells you. These people are not updating their LinkedIn and applying for jobs. They need to be found while they're still inside, often before they've admitted to themselves that they're leaving. That's where I come in. That's what cold calling was built for.

Technical founders are winning the contracts that used to go to sales teams

This came up repeatedly over dinner — not as a theory but as something Ed and I are both seeing play out in real time. Technical founders who can walk a customer through the architecture themselves are closing the biggest deals in defence tech. The old model — non-technical BD selling the output of an engineering team to a prime contractor — is dying.

At Helsing, the leadership blends geopolitical expertise with genuine technical depth. Ed pointed out that Chinese defence companies have operated this way for years — founders and CTOs driving the big deals personally. Western defence tech is catching up, but slowly.

If you're a defence tech founder who can't go deep on your own technology, you're increasingly at a disadvantage against founders who can. And I see this constantly in my work. I've watched a single placement — someone the client nearly rejected — go on to close a deal worth nearly $100m within six months. The old enterprise companies were supertankers where no individual crew member could shift the trajectory. These new defence tech companies are frigates. One or two people change everything.

What the next 18 months look like

European defence budgets are expanding rapidly. A new generation of early-stage companies in autonomous systems, decision support, and AI-enabled intelligence is proliferating. Helsing is scaling across Europe. Applied Intuition is building out its UK presence — I helped build that team from scratch. Improbable has pivoted fully into defence as Skyral. New seed-stage companies are appearing every month.

Many of them will need to build entire teams from scratch in the next 12 to 18 months. They'll need people who understand both the technology and the operational context…and eligible for security clearance — and those people are disproportionately available from within the military's broken promotion pipeline. They'll need advisory structures that go beyond a letterhead and a LinkedIn endorsement. And they'll need to solve the dual-use problem honestly or risk being another demo that never scales.

The campaign maths is clear. The talent is available (if you know where to hunt and how to reach them). The capital is flowing. The question for founders is whether you're building a company that can survive peacetime economics long enough to matter when the fleet-versus-fleet war game starts for real.

As for Ed — he finished his cheesecake, noted that he'd need to go easy on the red given his half-hour Brompton ride home, and then casually mentioned he'd flown over Buckingham Palace twice in his career. Just dropped it in between the mochi and the bill. Like that's a normal thing to say.

We're going to work together. I think there's something genuinely valuable in connecting the frustrated talent inside the services with the founders who need them — and I'm probably the only person mad enough to cold call a serving officer and say, mate, let me tell you about what's on the other side. More on that soon.

Frequently asked questions

Why are unmanned fighters cheaper than manned jets?

Manned fighters carry the full engineering cost of keeping a pilot alive — life support, ejection seats, oxygen, fire protection, triple-redundant systems, and components tested to 10⁻⁷ failure probability. They also need 7,000-hour airframe lives for decades of training. Unmanned fighters designed from scratch need none of this, requiring only 200–300 hours of airframe life with single-redundancy systems and relaxed certification, producing a roughly 4:1 cost advantage.

What is the loyal wingman concept?

The loyal wingman pairs an unmanned drone with a manned fighter like an F-35 or Typhoon. The drone carries sensors or weapons and absorbs threats to protect the pilot. It's a transitional doctrine that keeps the human in the loop while introducing autonomous platforms into the force structure — a halfway house rather than the end state for unmanned combat aviation.

What does dual-use mean in defence technology?

Dual-use means a technology company serves both commercial and military markets. The problem is that military capability is only needed at scale during conflict, so defence tech companies need genuine commercial revenue to survive peacetime. Without it, "dual use" becomes narrative convenience rather than a functioning business model.

How does fleet-vs-fleet war gaming change the unmanned debate?

Individual dogfight comparisons favour manned fighters. Fleet-level war gaming over a full air campaign — modelling attrition, repair, and replacement cycles — favours unmanned systems. At a 4:1 cost ratio, 400 unmanned aircraft can absorb a 2:1 kill-ratio disadvantage against 100 manned fighters and still have 200 remaining when the manned fleet is exhausted.

Why are experienced military officers joining defence tech startups?

Military promotion systems funnel all talent into one narrow race for command, ejecting frustrated high-performers at every stage. These officers increasingly see more impact — and arguably more patriotic contribution — in joining national security technology companies than in staying inside institutions that marginalise anyone not on the senior leadership track.

────────────────────────────────

• Steven Bourne is a distinguished headhunter and documentarian. 22 years operating behind the scenes from London to Silicon Valley, building companies and leading covert executive searches for the world's most formidable DeepTech, AI, and Defence companies. stevenbourne.com

• Air Marshal (Ret.) Edward Stringer CB CBE served as Director General Joint Force Development and Head of Joint Operations at UK Permanent Joint Headquarters. He now advises defence technology companies and research organisations.