What Future Spillovers Will Defence Tech Create?
From strategic necessity to civilian infrastructure, today’s defence technologies could reshape industries far beyond geopolitics.
Key findings
- Defence-funded innovation has historically generated major civilian technologies, from GPS to the Internet.
- Autonomy, secure communications, and advanced manufacturing are emerging as the next major spillover arenas.
- Companies that survive beyond geopolitical cycles are often those designed for civilian relevance from the outset.
- The challenge for founders and investors is not only building strategic technologies, but anticipating their “third life.”
At the World Economic Forum in Davos this January, Canadian Prime Minister Mark Carney declared what many in that room had been carefully avoiding saying out loud: that we are living through the end of what he called the pleasant fiction, a decades-long story in which the rules-based international order could be relied upon to hold, and that a brutal new reality of unconstrained great-power geopolitics had taken its place.
History repeats. Not in the sense that events are identical across centuries, but in the sense that the underlying structures recur with enough regularity that ignoring them is a choice. The pattern of defence-funded research generating unexpected civilian value has appeared after every major period of geopolitical mobilisation in the modern era, and there is no structural reason to believe this moment will be different. The question is not whether the current wave will produce transformative technology; it will. The question is whether the people building it have thought seriously about what happens when the geopolitical temperature drops again, and the pleasant fiction reasserts itself, as it always does, for a while.
For deep tech founders, the temptation is to optimise entirely around today's procurement wave. The harder question is what the company looks like in twenty or thirty years: will the technology still matter outside a classified programme, or will there be a warehouse of highly specialised prototypes and a cap table that only works in a permanent crisis? Treating the historical record of defence spillovers as a design space is one way to build companies that survive peace.
How defence innovation became civilian infrastructure
ARPA-funded research networks became the early Internet. GPS, originally a military navigation and targeting system, became the default positioning layer for logistics and smartphones. The cavity magnetron, developed for British wartime radar, became the kitchen microwave. In Europe, Galileo and Copernicus are explicitly dual-use: precise positioning for finance and aviation, and strategic observation for defence, from a single constellation. Israel built an entire cybersecurity industry by re-architecting military capabilities from Unit 8200 for regulated civilian markets worldwide.
Rolls-Royce offers a similarly powerful industrial example. The company built its core capabilities through military aircraft propulsion, especially wartime aero engines such as the Merlin and Griffon. That engineering base carried into the jet age after World War II: the Avon turbojet was developed beginning in 1945, first ran in 1947, entered military service, and was then adapted for civil aviation. Civil versions of the Avon went on to power the de Havilland Comet, one of the first commercial jet airliners. This was not a one-off transfer but the basis of a durable civilian aerospace franchise. Rolls-Royce’s Civil Aerospace business now supplies engines for large commercial aircraft and business aviation, with a market share in the large commercial aircraft space amounting to 54% according to a 2023 company statement.
Computer graphics also illustrate this pattern. In 1965, ARPA funded a center of excellence at the University of Utah under Ivan Sutherland to solve problems of simulation, cockpit displays, and making complex environments computable. That ecosystem produced the foundational primitives for interactive 3D graphics, companies like Evans and Sutherland and Silicon Graphics, and the intellectual lineage from which Pixar and Adobe later emerged, Pixar via Lucasfilm, Adobe via Xerox PARC. Neither was a defence contractor. Both were built on top of capabilities that defence money had seeded a generation earlier. The same rendering stacks that drove Toy Story and InDesign now power synthetic training environments and battlefield rehearsal systems. Defence bought the primitives, lost them to the commercial world, and bought them back as finished products.
Three Areas Where Spillovers Are Almost Guaranteed
Autonomy
Bringing autonomy to the edge is fundamental in military operations. In the civilian space, Internet connection and access to cloud services are within reach. That is not the case on the battlefield. This reality is pushing drone developers to build AI models that can run directly on a drone and autonomously navigate toward a target without relying on continuous external connectivity.
Testing has been one of the main constraints in developing drone autonomy, together with the scarcity of accessible use cases in the civilian world. In most European countries, testing is highly constrained and approval heavy1. At the same time, there wasn't a compelling enough commercial case to deploy capital into drone autonomy startups at scale, compounded by the complexity of getting regulatory approval.
One example is the flying taxi sector: startups such as Volocopter have been pushed into bankruptcy by complex regulations rather than the lack of a working product or customers.
Now, autonomous technologies are being deployed on the battlefield, where the only limits are technological. We still don't have total autonomy, but we have already reached effective battle autonomy on the Ukrainian battlefield. Drones are capable of navigating to predesignated locations, acquiring targets, and striking fully autonomously. The same applies to terminal guidance on very small First Person View (FPV) drones used in short-range operations, where onboard AI allows the drone to lock onto and hit a target in the last mile of flight without human input.
When it comes to autonomy, an overlooked space evolving at breakneck speed due to military demand is fully autonomous Unmanned Ground Vehicles (UGVs). The ground domain is far more complex and unstructured from an AI perspective than the air domain, where a drone navigates through a substantially obstacle-free 3D environment. On the ground, the system has to deal with terrain variation, physical obstacles, and unpredictable surfaces in real time.
This is also a space that could bring significant changes to industries that are less regulated when it comes to autonomy, such as mining and civilian construction. These sectors, where labour shortages are intensifying due to declining working-age populations and rising education levels in wealthy countries, trends pulling people from physically demanding jobs, will require new investments, with mining alone needing $2.1 Trillion2 by 2050. Therefore, autonomous ground systems could fill that gap.
The ultimate effect of deep, effective multi-domain autonomy would be the large-scale adoption of autonomous systems in civilian life. This could bring the strongest shock to our way of life and economy ever experienced, as well as deeply challenge the ways in which we as humans contribute to society.
Cryptography and Secure Communications
Secure communication has always been a defence concern and is now a systemic industrial one. Post-quantum cryptography, satellite links that remain trustworthy in contested environments, resilient mesh networks, and authentication that works when infrastructure is degraded are exactly the challenges defence ministries are currently funding at scale. The civilian side is equally large: every bank, logistics network, hospital, and cloud provider will eventually need to migrate away from cryptographic primitives vulnerable to quantum attacks.
The Israeli cybersecurity experience is instructive here. The strongest companies in that ecosystem translated intuition cultivated in Unit 8200, then rebuilt products almost entirely from scratch for regulated civilian customers, treating interoperability and compliance as core design constraints from day-one. That discipline is what allowed them to sell into global financial institutions while remaining relevant to governments, and it is the model worth studying for any cryptography company being seeded by defence demand today.
Teletactica is a Ukrainian company developing drone communication antennas and devices whose architectures were deployed under dense electronic warfare conditions on Ukraine's frontline, and which has since established an EU presence to supply allied defence customers, with civilian applications still ahead of them, but the underlying technology pointing clearly in that direction.
Yevhen Zhbeko, Co-Founder at Teletactica: “Our core focus is making communication reliable in the worst conditions, under jamming, limited bandwidth, and intermittent connectivity. While today that’s driven by defence needs, the spillover into civilian use is very natural. You see the same constraints in disaster response, mountain and maritime rescue, border and critical infrastructure security, industrial machine-to-machine communication in noisy environments, and even large-scale events where networks get overloaded. In all these cases, the challenge is not speed, it’s reliability and trust in the connection.”
“Our core focus is making communication reliable in the worst conditions, under jamming, limited bandwidth, intermittent connectivity. Today that's driven by defense, but civilian spillover is natural.”
- Yevhen Zhbeko, Co-Founder at Teletactica
Advanced Manufacturing
Roboze produces high-precision additive manufacturing systems for super polymers, including carbon PEEK, reducing component mass substantially for aerospace, defence, and energy. Tetmet, co-founded and led by Tom Vroemen, HEC Paris EMBA alumnus and CDL Paris mentor, takes a different approach with its patented Adaptive Spatial Lattice Manufacturing: AI-guided robotic arms and laser welding assemble metal rods into large lattice structures up to seventy per cent lighter than conventional metal sheet equivalents, using standard recyclable materials. Already working with Safran, Ariane, and Stellantis, the same structural logic that makes a lighter satellite also makes a lighter airframe and a more economical solar installation frame.
The frontier of advanced manufacturing extends further still. As materials science converges with neurotechnology, the next generation of components will be customised not just for structural performance but for biological compatibility: flexible substrates, biocompatible polymers, and lattice geometries designed specifically for brain-computer interfaces and neural implants. Cogitat, an Imperial College London spinout and CDL alumnus, decodes brainwave data from standard EEG headsets using machine learning and has already been tested in NHS-linked stroke rehabilitation research. Its inclusion in NATO DIANA's 2025 cohort points to plausible defence uses such as cognitive readiness monitoring, workload assessment, and training support, rather than direct weapons control.
Dimitrios Adamos, Co-founder and CTO at Cogita:
“Neurotechnology is one of the areas where defence investment is building capabilities that will transform civilian life. Non-invasive brain-computer interface technology is being developed for defence training today, measuring cognitive performance from brain signals in real-time. This will reshape how we approach professional training and workplace safety in the coming decade. Building our technology through the NATO DIANA programme, we see this first-hand at Cogitat: the same technology that serves defence applications has direct potential in civilian domains like air traffic control and pilot safety.”
“Neurotechnology is one of the areas where defence investment is building capabilities that will transform civilian life."
- Dimitrios Adamos, Co-founder and CTO, Cogitat
Designing technologies for their “third life”
Most technologies that matter have at least three lives:
- The first is in research prototypes, funded by defence or public capital because the upside is strategic rather than commercial. In this stage, companies not only receive public support but can bypass stringent certifications and security concerns that would be considered unacceptable in the civilian sector, but are fine for the military. This allows for rapid iteration and hyperquick R&D.
- The second is shaped by demanding early adopters with intense and very particular needs.
- The third is when the technology disappears into everyday infrastructure: ARPANET became the Internet, the cavity magnetron became the microwave, and Galileo became an invisible timing layer that most of its users will never think about.
Founders and investors can either design for that third life or ignore it. Ignoring it is easier: take the bespoke contract, let the first customer dictate the architecture. There is money in that strategy, but it rarely produces companies that matter after the crisis cycle has moved on. As we see this tension up close, sitting across tables from founders trying to reconcile the clarity of a defence contract with the ambiguity of everything that might come after. The ones who treated the defence context as a proving ground rather than a destination tended to build something more durable.
When Demis Hassabis co-founded DeepMind, he argued that the people closest to a powerful capability carry a specific responsibility to think harder than anyone else about where it goes. Defence tech founders are in an analogous position. The capabilities being built under current procurement pressure are real, they are dual-use almost by definition, and the people building them are best placed to determine whether they eventually land in a hospital or a logistics network, or remain locked in a classified programme until they are obsolete.
Founders and investors can either design for that third life or ignore it. Ignoring it is easier: take the bespoke contract, let the first customer dictate the architecture. There is money in that strategy, but it rarely produces companies that matter after the crisis cycle has moved on. As we see this tension up close, sitting across tables from founders trying to reconcile the clarity of a defence contract with the ambiguity of everything that might come after. The ones who treated the defence context as a proving ground rather than a destination tended to build something more durable.
When Demis Hassabis co-founded DeepMind, he argued that the people closest to a powerful capability carry a specific responsibility to think harder than anyone else about where it goes. Defence tech founders are in an analogous position. The capabilities being built under current procurement pressure are real, they are dual-use almost by definition, and the people building them are best placed to determine whether they eventually land in a hospital or a logistics network, or remain locked in a classified programme until they are obsolete.
As Carney reminded at Davos: “nostalgia is not a strategy”3. Neither is optimising entirely for a crisis that will, eventually, pass. The real question for every investor reading a defence tech pitch deck and every founder choosing an architecture is simpler and harder than any regulatory framework: what is the third life of what you are building, and are you designing for it?
History has shown us, repeatedly, that the spillovers no one planned for are often the ones that matter most. So, what future spillover do you see?
This article has been written by Livia Kalossaka & Paolo Trecate.
Sources
European Commission, Joint Research Centre. (2026). Drones, counter drones and autonomous systems.
BloombergNEF. (2024, October). Mining industry needs $2.1 trillion in new investments by 2050.
Cecco, L. (2026, January 21). 'Nostalgia is not a strategy': Mark Carney is emerging as the unflinching realist ready to tackle Trump. The Guardian.