Major Axel Gilliot is a senior officer in the French Army, currently a fellow at the Belfer Center for Science and International Affairs at the Harvard Kennedy School as part of his École de Guerre program. He writes on defense innovation in a personal capacity.
It is 3:14 a.m. A French infantry platoon has held its position for thirty-six hours in a contested sector, with no vehicle and no means of recharging its equipment. The platoon leader takes stock of his resources: two observation systems at 12% battery, a primary radio and the platoon’s drones at 8%. The precision marksman’s night-vision goggles went dark an hour ago. The combat system is intact; its energy source, however, is dead. A modern infantry platoon becomes blind and mute after thirty-six to forty-eight hours—less in extreme cold.
This is not because the technology doesn’t exist, but rather because of the lack of an adequate mechanism to translate it from laboratory to field.
Modern armed forces operate within an unprecedented paradox: they have at their immediate disposal a technological reservoir of historically unmatched richness—universities, laboratories, start-ups, engineering schools. The technological building blocks exist: advanced materials, embedded systems, distributed energy, resilient communications, robotics, often only a few kilometers from the barracks. And yet the link between the technology developed in a laboratory and the technology that concretely transforms the life of an infantry platoon remains long, uncertain, and costly.
It is this concrete and urgent question—not an abstract or doctrinal one—that I brought with me to the Belfer Center of the Harvard Kennedy School in 2025 as a military fellow.
Cambridge as a Laboratory of Observation
The Belfer Center is not just a specialized think tank. It constitutes a singular space where students, political figures, military veterans, international security scholars and entrepreneurs converge. Importantly, these entrepreneurs are convinced that national defense demands the same standards of agility and experimentation as the most innovative sectors of the civilian economy. For a senior officer of the French Army, schooled in long acquisition processes and vertical chains of validation, exposure to this ecosystem produces a striking effect of contrast.
Within the Belfer Center, the QLab1—founded in 2023 by two veterans, one from the U.S. Army, the other from the U.S. Air Force—offers a singular illustration of the model’s power: thirty-four defense start-ups created in eighteen months, on an annual budget of just three hundred thousand dollars. Thirty-four projects born of operational needs that no one had yet translated into requirements, and whose very existence attests that the most fragile phase of the innovation process can indeed be industrialized.2
A few kilometers away, at MIT, a U.S. Air Force reserve officer directs Mission Innovation X,3 a structure whose mission is disarmingly simple: to connect military problem owners with academic capability providers, through short, iterative cycles. The outcome is not always the validation of a prototype; it is often something even more valuable: a team assembled around a real problem, a hypothesis confirmed or refuted, or a documented failure sparing the expenditure of vastly more resources in a formal program bound for a dead end.
Farther afield, at Stanford University, the Gordian Knot Center for National Security Innovation approaches the question from a complementary angle. It does not directly produce technologies or companies but rather trains officers and civilians in a competence found in no military manual: how to frame an operational problem in actionable terms. Its flagship program, Hacking for Defense (H4D), presents multidisciplinary student teams with topics submitted by military units or government agencies. Within a few weeks, the students develop the reflexes of iteration and testing on a concrete operational irritant. The officers, for their part, become familiar with an ecosystem that had been almost entirely unknown to them, while sharpening the armed forces’ expression of need.4
The Missing Link in France
France ranks among the nations that have invested the most, institutionally and financially, in structuring their defense innovation ecosystem. The French Defense Innovation Agency (Agence de l’Innovation de Défense), under the authority of the General Delegate for Armament (Délégué général pour l’armement), has structured a coherent and recognized ecosystem. The French Future Combat Command (Commandement du Combat Futur) oversees experimentation with methodological rigor. Exploratory hubs attached to brigades establish a short, operational feedback loop between field needs and available equipment. This whole architecture works, but only once a project has already taken shape—with an identified team, a demonstrable and scalable prototype, and a designated institutional sponsor.
The point of fragility lies upstream, in what Americans call zero-to-one: turning a field observation or a laboratory technology into a constituted team and a testable prototype. It is precisely in this zone of maximum fragility—upstream of the case file, funding, and specification sheet—that a considerable share of innovation potential dissipates. So, we end up with operational hunches that are never articulated, university technologies whose potential military relevance is never flagged, or civilian engineers who never identify the entry point to report recurring dysfunctions they observe in daily contact with their defense clients.
This is more than a problem of will; it is a problem of mechanism.
The Pre-Zero: The Ground the French Army Must Conquer
What has struck me most in the Cambridge ecosystem is not just the technology but also the culture of early exposure. MIT students work on United States Special Operations Command (USSOCOM) problems from their first year of master’s study. Doctoral candidates in composite materials spontaneously ask themselves whether their research might be of interest to an infantryman. This connection is neither innate nor native: it is the product of years of institutional presence of the American armed forces on campus—physical, regular, and embodied by service members and veterans who speak the language of researchers because they have been trained to do so.
A Concrete Proposal
Out of this year of observation has emerged a proposal that is simple in principle but demanding in execution: the creation of an Army start-up studio. Not another incubator, nor another accelerator, nor an additional service window. But a deliberately lean structure, attached to the French Army Staff and physically embedded on a university campus. Its sole mission would be to transform operational needs or technological opportunities into constituted teams and testable prototypes—before a structured handover to the existing institutional mechanisms.
The funding model is deliberately frugal. The challenge is not budgetary, it is structural: to have a mechanism to deploy its resources in the right place, at the right moment, with explicit “kill” criteria.
The capacity to swiftly terminate a project that fails to demonstrate its operational value is every bit as important as the capacity to launch one. The “kill” is not an admission of failure. It is a governance decision.
Herein perhaps lies the most counter-intuitive lesson of the American ecosystem: the capacity to swiftly terminate a project that fails to demonstrate its operational value is every bit as important as the capacity to launch one. The “kill” is not an admission of failure. It is a governance decision, taken with method and on the basis of pre-established criteria, which frees resources for the next project.
What France Can Develop in Return
After a year spent in Boston, I now have a firm conviction: France does not need to replicate the American model. But it has the capacity to draw inspiration from it to build its own. The French model would be more austere in form, more anchored in existing institutional structures, and more attuned to a military culture that combines methodological rigor with an appetite for boldness. The Army start-up studio is not an American import. It is a reasoned translation, in the service of a distinctly French ambition.
The most demanding translation is not a technical problem but a cultural one: to recognize that an imperfectly worded observation by a logistics non-commissioned officer is sometimes more valuable for operational innovation than a perfectly documented research program. But this recognition is not enough. It must be translated into physical presence, into explicit mandates and into reproducible mechanisms. Systematizing the presence of the armed forces at every stage of the innovation chain—and particularly at its origin—is less an organizational choice than a strategic imperative. Only if this is achieved will France stop missing out on the technologies that were destined for it without its knowing.
1. Ryan Holte and Nick Maynes, founders of the QLab, Belfer Center for Science and International Affairs, Harvard Kennedy School. Program launched in 2023, attached to the Intelligence Project.
2. David Ignatius, “Harvard has a defense-tech incubator. Pete Hegseth should visit it,” The Washington Post, May 14, 2026.
3. MIT Mission Innovation X (MIx), a dual-use program directed by Gene Keselman, Executive Director and reserve officer of the U.S. Air Force.
4. Gordian Knot Center for National Security Innovation, Stanford University, founded by Steve Blank and Joe Felter in 2021. The Hacking for Defense (H4D) program, created in 2016, is today taught in more than 70 universities. Seventy start-ups have emerged from it, having raised more than 350 million dollars.