[Avoiding Hype, Serving the Mission]
Sidenote: I had the pleasure of doing a (German) podcast interview with an incredibly smart host and experienced interviewer. It just aired today .. which I was not sure of it ever would, because (A) I don’t think I was very succinct; and (B) there was very clear feedback that my strong feelings about the responsibilities of private equity and venture capital in defense tech are admirable, but they are also not very tangible. I realized that I talk a lot about HOW to do things, not WHAT to do. Let’s balance my frontier observations with pragmatic execution.
Enduring innovation in defense does not follow the hype cycle. It follows the kill chain.
In an environment where failure is fatal—and success is imitated by adversaries—real progress is neither press-released nor overfunded. It emerges where the mission demands it. Below, I outline six active domains of innovation where startup technologies are demonstrating utility, adaptability, and survivability at scale .. and are able to talk about it. Each is a signal worth tracking. [BONUS: at the end I mention a few of my “un-convictions”]
1. Autonomous Aerial Systems & AI Pilots
AI-powered UAVs from firms like Shield AI and Anduril are shaping how autonomy coexists with mission command. These systems challenge our assumptions about risk, control, and accountability. Other examples who demonstrated utility, adaptability, and survivability at scale:
- Skydio: A U.S.-based company specializing in autonomous drones for reconnaissance missions, notably contributing to the Army’s Short Range Reconnaissance (SRR) program.
- Elroy Air: Developing autonomous cargo drones for aerial resupply missions, with recent demonstrations for the U.S. Marine Corps.
- Quantum Systems: A German startup integrating drone hardware, software, and AI to deliver real-time aerial intelligence.
2. Edge AI & Tactical Situational Awareness
Decision-making at the tactical edge cannot wait for a cloud uplink — it needs real-time inference in contested space. Systems built by Rebellion Defense and Tactical AI integrate sensor fusion and threat identification in denied environments. This isn’t “AI for defense”—it’s cognitive infrastructure, built to survive latency, ambiguity, and human exhaustion. Other examples who demonstrated utility, adaptability, and survivability at scale:
- Reveal Technology: Provides AI-powered 3D modeling and computer vision solutions for enhanced battlefield awareness.
- Vantiq & C4i Systems: Collaboratively launched VANGUARD AIX, an AI-powered edge solution designed for rapid defense response.
3. Resilient Communications & Spectrum Adaptation
Jamming is the new normal. The electromagnetic spectrum is the first terrain to be contested in modern conflict. Silvus Technologies and others are leading with mobile-networked MIMO radios that thrive in chaos, not in labs. Survivable comms are not an accessory—they are the precondition for command. Other examples who demonstrated utility, adaptability, and survivability at scale:
- Himera: A Ukrainian startup producing electronic warfare-resistant communication systems, recently supplying secure radios to the U.S. Air Force.
- Rampart Communications: Offers resilient wireless communications with physical layer encryption, protecting against cyber and EW threats.
4. Modular Munitions & Loitering Weapons
Precision, cost asymmetry, and autonomy converge in new forms of loitering munitions and modular strike platforms. It demonstrates adaptability over mass. Firms like Anduril are building reprogrammable effectors, not just weapons. The ability to change payloads or missions mid-flight reframes what we mean by “flexibility under fire.”
- Cummings Aerospace: Developed the Hellhound, a 375-mph, 3D-printed turbojet-powered loitering munition designed for the U.S. Army’s Low Altitude Stalking and Striking Ordnance (LASSO) program.
- UVision: An Israeli company supplying loitering munitions like the Hero series to various armed forces, with plans to expand in the U.S. market.
5. Command, Control & Networked Autonomy
JADC2 is not a platform—it’s a philosophy. Lattice OS and related systems seek to unify the battlespace—not by centralizing it, but by enabling decentralized actors to make sense and act fast. Autonomy here is not a drone trait; it’s a system-level property of how intent scales. Other examples who demonstrated utility, adaptability, and survivability at scale:
- Scout AI: Raised $15 million to develop its Fury system, enabling human operators to command fleets of military robots across various domains.
- L3Harris Technologies: Introduced AMORPHOUS, a software allowing control of thousands of autonomous assets through a single user interface.
6. Defense Industrial Base & Manufacturing
But speed without scale is a gimmick. Anduril’s Arsenal-1 facility and others signal a return to serious capacity. Building modern kill-web systems requires precision manufacturing, modular architecture, and production agility. These are not boutique labs. They are strategic assets—and any nation that lacks them is a nation at risk. Other examples who demonstrated utility, adaptability, and survivability at scale:
- Hadrian: Aims to revolutionize defense manufacturing by producing parts rapidly and efficiently, supporting defense startups in scaling production.
- X-Bow Systems: Collaborating with Lockheed Martin to become a major supplier of solid-rocket motors, addressing domestic supply chain concerns.
What unites these domains is not the allure of novelty but the urgency of operational fit.
The startups mentioned here aren’t just building products. They are adapting to friction, constraint, and consequence—often faster than primes, and closer to the edge. That is where innovation becomes credible. And in defense, credibility is the only currency that matters.
[BONUS: WHAT NOT TO DO]
Above six domains reflect my convictions .. what about my “un-convictions”? If you find yourself in one of these areas, please reach out to me: Maybe I’m wrong, and I like to hear good arguments why. I usually prioritize credible, constrained, and consequence-aware innovation. However, I red-flag companies that:
🚩 Externalize ethical or operational failure
🚩 Prioritize scale over safety
🚩 Ignore end-user feedback or deployment frictions
🚩 Exploit narrative momentum without domain knowledge
DONT: “Spray and Pray” AI Integration Firms
Startups pitching AI-enabled everything: predictive logistics, autonomous C2, computer vision for soldier biometrics—without explaining trade-offs or failure modes. They are often backed by generalist funds or accelerators, lacking end-user contact. These firms commodify decision-making without accountability. They treat defense as a dataset problem, ignoring the edge-case-driven reality of warfighting. Their abstraction layers externalize ethical and operational risk—often to frontline operators.
DONT: Companies Building for Surveillance-Capitalism Transferability
Dual-use platforms originally designed for commercial surveillance (facial recognition, crowd analytics, behavioral nudging) being “ported” to defense applications. They are often eager to sell to both adversarial regimes and democratic allies, depending on who pays. These ventures externalize moral and democratic cost—amplifying surveillance infrastructures that can be used against civilians. They treat “compliance” as ethics and avoid hard conversations about authoritarian misuse, power asymmetries, and consent.
DONT: Tactical NFTs, Web3-for-War, or Blockchain Bloat
Blockchain tech absolutely does have a place on the battlefield. But often startups are proposing immutable battlefield logs, tokenized supply chains, or blockchain-based ID for forward operators. They are frequently light on technical viability and heavy on buzzword convergence. These ideas often exhibit narrative inflation—selling theoretical resilience while introducing complexity, latency, or fragility. They substitute ideology for utility and invite brittle systems into high-consequence environments.
DONT: Zero-accountability Weaponization Startups
Loitering munitions or autonomous targeting systems pitched with “man-on-the-loop” language but without concrete human-in-control governance. These are often startups that treat lethal autonomy as a software scalability problem. These firms externalize decision risk. They build tools that can kill without ensuring the architecture of responsibility, escalation thresholds, or treaty compliance. Autonomy without doctrine is not progress—it’s precarity.
DONT: Defense Companies Built as PR Plays or Financial Arbitrage
SPAC’d defense tech companies with weak IP, no deployments, and a polished pipeline of marketing decks and DoD logos. Their business model is less “serve the mission” and more “sell the story before reality catches up.” It violates my stance against narrative inflation. These ventures commodify national security to satisfy quarterly expectations and often degrade trust in the broader defense innovation ecosystem.