What would a European DARPA actually look like?

The US model that produced the internet, GPS, and mRNA vaccines. Europe keeps talking about replicating it. What would it take, and what is actually stopping it?

By VastBlue Editorial · 2026-03-26 · 20 min read

Series: Reindustrialising Europe · Episode 9

The Agency That Changed Everything

In 1958, the United States was in a state of institutional shock. The Soviet Union had launched Sputnik the previous October — a polished aluminium sphere the size of a beach ball, orbiting the Earth every ninety-six minutes, emitting a steady radio pulse that anyone with a shortwave receiver could hear. The technological achievement was modest by later standards. The psychological impact was seismic. The country that had invented the atomic bomb, mass-produced the B-52, and built the world's most powerful economy had been beaten into orbit by a communist state that most American policymakers considered technologically backward. Within months, President Eisenhower signed the order creating the Advanced Research Projects Agency — ARPA, later renamed DARPA — inside the Department of Defense. Its mandate was deceptively simple: prevent technological surprise.

What emerged over the following six decades was not merely an agency but an institutional form — a way of organising the relationship between government funding, scientific research, and technological deployment that has no true equivalent anywhere in the world. DARPA operates with an annual budget of approximately $4 billion, employs roughly 220 people, funds no internal laboratories, owns no equipment, and produces no products. It is, in the most literal sense, a funding body. But to describe DARPA as a funding body is like describing a conductor as someone who waves a stick. The mechanism is the least interesting part of what it does.

The list of technologies that trace their lineage to DARPA programmes reads like an inventory of the modern world. ARPANET, the precursor to the internet, emerged from a 1960s programme to build a communications network that could survive nuclear attack. GPS began as a DARPA-funded programme called Transit in the early 1960s, later transferred to the Air Force. Stealth aircraft technology was developed under DARPA's Have Blue programme in the 1970s. Voice recognition, autonomous vehicles, advanced materials, precision-guided munitions, the RISC processor architecture — all passed through DARPA at critical stages of their development. More recently, DARPA's investments in messenger RNA technology through programmes like ADEPT and P3, beginning in 2013, created the foundational platform that Moderna and BioNTech used to develop COVID-19 vaccines at unprecedented speed. When the pandemic struck, the technology was not invented from scratch. It was waiting on a shelf that DARPA had stocked years earlier.

The Programme Manager: An Institutional Species Europe Has Never Cultivated

Understanding DARPA requires understanding a single institutional innovation that Europe has never successfully replicated: the programme manager. DARPA's programme managers are not administrators. They are not grant evaluators. They are not consensus-builders. They are, in the agency's own language, "technological entrepreneurs operating within a government framework." Each programme manager arrives with a specific technical vision — a theory about how a particular technological barrier might be overcome — and is given between three and five years, a budget typically ranging from $10 million to $50 million per year, and near-complete autonomy to pursue that vision. They select performers — universities, companies, national laboratories — negotiate contracts, set milestones, and make go/no-go decisions. If a programme fails to demonstrate progress, the programme manager can kill it. If a programme succeeds, the programme manager can accelerate funding. There are no standing committees, no peer-review panels, no multi-year evaluation cycles.

The genius of DARPA is not that it picks winners. It is that it empowers individual technical visionaries to make bets, gives them enough resources to test those bets quickly, and accepts a high failure rate as the price of occasional transformative success.

Adapted from William Bonvillian & Richard Van Atta, "ARPA-E and DARPA: Applying the DARPA Model to Energy Innovation," 2011

The programme manager model embodies several principles that are structurally alien to European research governance. First, individual authority. A single person — not a committee, not a consortium, not a panel of national representatives — decides which projects to fund and how to manage them. Second, term limits. Programme managers serve for four to six years, then leave. There is no career track at DARPA. The agency explicitly rejects institutional memory in favour of institutional renewal. Third, tolerance for failure. DARPA expects a significant proportion of its programmes to fail. What it does not tolerate is timidity — programmes that pursue incremental improvements rather than transformative breakthroughs. The agency's own documentation describes its mission as funding "high-risk, high-payoff" research that bridges the gap between fundamental science and military application. The emphasis is on "high-risk." If a programme has a high probability of success, DARPA's institutional logic says it should be funded by someone else.

Contrast this with how Europe funds research. The European Research Council distributes grants through peer review panels that inevitably favour incremental advances by established researchers. Horizon Europe distributes its €95.5 billion through a Byzantine system of work programmes, evaluation committees, and consortium requirements that can take eighteen months from application to contract signature. The European Innovation Council funds companies rather than programmes and lacks the autonomy to make rapid funding decisions without institutional oversight. None of these mechanisms are designed to empower individual technical visionaries. They are designed to distribute resources fairly across member states and minimise political controversy. These are defensible objectives. They are also the precise opposite of what makes DARPA work.

Europe's Attempts: A History of Structural Misunderstanding

Europe has not ignored DARPA. It has discussed replicating it, studied it, commissioned reports about it, and launched institutions that borrow its rhetoric while systematically omitting the structural features that make it function. The history of these attempts is instructive — not because they failed, but because the pattern of failure reveals something important about why European institutions resist the DARPA model.

The earliest significant attempt was the European Defence Agency, established in 2004 in the aftermath of the Iraq War, when the gap between American and European military capabilities had become embarrassingly visible. European defence R&D spending was fragmented across twenty-five national programmes, each too small to fund breakthrough research, each too politically protected to consolidate. The EDA was supposed to fix this. Its annual budget for collaborative research programmes has never exceeded €40 million — roughly one per cent of DARPA's budget. It operates by consensus among member states, funding only projects that all participating nations agree to fund. It has produced useful work in counter-IED technology and unmanned systems, but nothing resembling a transformative breakthrough. It was designed not to.

The European Defence Fund, launched in 2017 and operational from 2021 with a budget of €7.9 billion over seven years, represents a more serious attempt. It funds both defence research (approximately €2.7 billion) and capability development (approximately €5.3 billion), and it operates at a scale that begins to approach meaningful investment. But the EDF distributes funding through calls for proposals evaluated by committees that include representatives of participating member states. Projects must involve entities from at least three EU countries. The evaluation criteria explicitly include "European added value" — a political concept that, in practice, means distributing work across multiple national industrial bases rather than concentrating investment where the technical opportunity is strongest. The EDF is an industrial policy instrument disguised as a research programme. It serves the legitimate purpose of developing European defence capabilities, but it does not and cannot replicate the programme-manager-driven, failure-tolerant, speed-oriented model that defines DARPA.

The European Innovation Council, made permanent in 2021 with a budget of €10.1 billion, is perhaps the most frequently cited "European DARPA." Its Pathfinder programme funds early-stage research with a stated tolerance for failure. On paper, the resemblance to DARPA is deliberate. In practice, the differences are fundamental. The EIC operates within the European Commission. Its programme managers are Commission officials bound by staff regulations, financial regulations, and procurement rules. They cannot hire and fire performers at will or redirect funding without institutional approval. They serve indefinite terms, accumulating institutional caution rather than shedding it. The EIC has funded promising companies, but it has not produced a single technology that has reshaped a strategic domain.

What Actually Stops It: Five Structural Barriers

The failure to build a European DARPA is not a failure of ambition, funding, or talent. Europe has all three. It is a failure of institutional architecture — the deep structural features of the European project that make certain kinds of institutions extraordinarily difficult to create, regardless of political will. Five barriers stand out.

1. The Juste Retour Problem

Every European institution that distributes money faces the juste retour constraint: member states expect to receive a share of funding roughly proportional to their financial contribution. Germany pays roughly 21 per cent of the EU budget and expects to receive approximately 21 per cent of research funding. France, Italy, Spain — each has an implicit expectation of proportional return. This expectation is not written into law (the European Commission formally rejects the principle), but it is embedded in the political economy of every budget negotiation, every programme design, and every evaluation criterion that includes "geographic balance" or "European added value." DARPA faces no equivalent constraint. When a DARPA programme manager decides that MIT and Lockheed Martin are the right performers for a particular programme, there is no requirement to include a laboratory from Mississippi or a company from Montana to ensure geographic balance. The programme manager funds the best performers, wherever they are. A European equivalent would need to do the same — and no EU institution has ever been permitted to operate this way.

2. The Sovereignty Trap

DARPA's authority derives from a single sovereign government with unified command over its military forces. The technologies DARPA develops flow into a single procurement system — the US Department of Defense — which deploys them across a single military. Europe has twenty-seven defence budgets, twenty-seven procurement agencies, twenty-seven sets of classified information handling procedures, and twenty-seven different strategic cultures that determine what technologies are prioritised. A European DARPA would need to develop technologies that are useful to all EU member states or, more radically, to a European force structure that does not yet exist. France's force de frappe has different technology needs from Germany's Bundeswehr, which has different needs from Poland's rapidly expanding military, which has different needs from Italy's Mediterranean-focused forces. The question "what technologies should a European DARPA develop?" cannot be answered without first answering the question "what European military capability are we building?" — and that is a question that EU member states have spent seventy years avoiding.

3. The Classification Problem

DARPA operates at multiple classification levels, including top secret and above. Many programme results are classified for years or decades. Europe has no unified classification system. EU CONFIDENTIAL and EU SECRET exist as categories, but each member state maintains its own national system, and the transfer of classified information between states is governed by bilateral agreements that are cumbersome and inconsistent. A European DARPA would need to operate across these boundaries — sharing results with qualified researchers regardless of nationality while preventing leakage to adversaries. The practical difficulty within the current European security architecture is enormous.

4. The Procurement Gap

DARPA's effectiveness depends not just on funding research but on the existence of a customer — the US military — willing to procure the resulting technologies at scale. When DARPA developed stealth technology, the Air Force bought stealth aircraft. When DARPA developed precision-guided munitions, the services deployed them. This procurement pull is essential. Without a guaranteed customer, breakthrough technologies remain laboratory demonstrations. Europe lacks an equivalent mechanism. There is no European procurement agency. Individual member states procure their own equipment, often preferring national champions for reasons of industrial policy and political patronage. A European DARPA could develop a transformative sensor technology, but deploying it would require twenty-seven separate procurement decisions — many influenced by factors that have nothing to do with technical merit.

5. The Talent Problem

DARPA recruits programme managers from the frontier of their fields — senior scientists and engineers who take a three-to-five-year leave of absence to serve at the agency before returning to their careers. The model works because American academic culture supports government service as a career stage. A professor at Stanford who spends four years at DARPA returns with enhanced prestige and industry contacts. Europe has no equivalent talent pipeline. EU institutions hire through formal competitions — concours — that take months or years. Staff regulations make the rapid hiring and departure of short-term technical visionaries almost impossible. National civil service traditions in France, Germany, and Italy do not encourage temporary departures for government service. Until that calculation changes, a European DARPA will struggle to attract the calibre of programme manager that makes the model work.

JEDI: The Latest Attempt and Its Limitations

In February 2025, the European Commission proposed the Joint European Defence Initiative — JEDI — as part of its ReArm Europe / Readiness 2030 plan. The name, presumably deliberate in its cultural resonance, signals ambition. The substance reveals the familiar constraints. JEDI is designed to coordinate European defence research and development, to channel investment toward critical technologies including artificial intelligence, quantum computing, hypersonics, and autonomous systems. Its proposed budget — drawn from a combination of EU funds, European Investment Bank financing, and member state contributions — is substantial, potentially reaching €150 billion in mobilised investment over the coming decade. But JEDI is not a DARPA. It is a coordination mechanism layered on top of existing institutional structures.

The Commission's proposal explicitly respects member state sovereignty over defence decisions. It does not create an empowered agency with its own programme managers and independent funding authority. It proposes to coordinate existing instruments — the European Defence Fund, the European Investment Bank, national defence budgets — through a governance structure involving the Commission, the High Representative for Foreign Affairs, and member states. JEDI may succeed in increasing the total amount of money Europe spends on defence technology. What it is unlikely to do is change how that money is spent — and the how is precisely what makes DARPA effective.

The Draghi Report on European competitiveness, published in September 2024, was unusually blunt on this point. Mario Draghi — former president of the European Central Bank, former prime minister of Italy, a figure whose institutional credentials are unimpeachable — argued that Europe's innovation gap with the United States was not primarily a funding gap but an institutional design gap. European research funding is "too fragmented, too slow, and too risk-averse," the report concluded. It recommended creating an institution modelled explicitly on DARPA, with programme managers empowered to fund high-risk research without the multi-layered approval processes that characterise existing EU instruments. The recommendation was politely received, widely discussed, and — as of early 2026 — has not resulted in any institutional change. The Commission cited the EIC and the EDF as evidence that the structures already exist. The structures exist. The model does not.

Europe does not have a spending problem. It has a spending architecture problem. The same amount of money, channelled through an institution designed like DARPA, would produce dramatically different results than the same money channelled through Horizon Europe or the European Defence Fund.

Analysis informed by the Draghi Report on European Competitiveness, September 2024

What It Would Actually Take

A genuine European DARPA — not a rebranded coordination mechanism but an institution that actually replicates the operational model — would require political decisions that no European leader has yet been willing to make. Describing them is straightforward. Implementing them would require a transformation in how Europe thinks about sovereignty, accountability, and the distribution of technological power.

First, it would need to be an agency, not a programme. The distinction matters. A programme is a pot of money distributed through existing institutions. An agency is an autonomous entity with its own legal personality, its own leadership, its own hiring authority, and its own budget. The agency would need to be located outside the Commission — not subject to Commission staff regulations, financial regulations, or procurement rules. It would need the legal authority to hire programme managers on short-term contracts at salaries competitive with the private sector, without the concours process that governs EU recruitment. It would need the authority to fund performers through contracts rather than grants — the distinction being that contracts allow the agency to specify deliverables, set milestones, and terminate for non-performance, while grants merely support research activity. DARPA does not give grants. It awards contracts. The difference in incentive structure is enormous.

Second, it would need a unified customer. Either a European procurement agency that buys defence equipment for a European force, or a binding agreement among participating member states to jointly procure the outputs of successful programmes. Without procurement pull, the agency becomes an expensive laboratory producing prototypes that sit on shelves. DARPA's technologies reach the battlefield because the Pentagon buys them. A European equivalent's technologies would reach nowhere unless someone commits, in advance, to deploying them.

Third, it would need to abandon juste retour. The agency would need to fund the best performers regardless of nationality — including, controversially, performers from non-EU countries when they possess capabilities that no EU entity can match. DARPA routinely funds research at universities and companies in allied countries. A European DARPA that could not fund a team in the UK, Switzerland, or Israel because they are not EU member states would be artificially constraining its talent pool. This is a political impossibility under current EU orthodoxy, but it is a technical necessity under the DARPA model.

• Autonomous legal personality outside the European Commission
• Programme managers hired on 4-5 year contracts at private-sector-competitive salaries
• Contract-based funding (not grants) with milestone-driven go/no-go decisions
• No geographic distribution requirements — best performers funded regardless of nationality
• A committed procurement customer: either a European procurement agency or a binding joint procurement agreement
• A unified classification framework enabling cross-border collaboration on sensitive technology
• Annual budget of at least €3-4 billion — concentrated, not distributed
• Explicit tolerance for a programme failure rate of 50% or higher

Fourth, it would need political protection from the accountability mechanisms that would normally govern an EU agency spending billions of euros. DARPA's tolerance for failure is legally and politically supported — Congress exercises oversight but does not micromanage individual programme decisions. A European equivalent would face scrutiny from the European Parliament, the European Court of Auditors, twenty-seven national parliaments, and a media environment that treats any expenditure without immediate visible results as evidence of waste. The first time a programme failed — and programmes would fail, because failure is inherent to the model — programme managers would face parliamentary inquiries, the media would compare the cost to hospitals and schools, and the political sustainability of the entire institution would be questioned. The tolerance for failure that makes DARPA work may be the hardest element to transplant into European political culture.

The Road Europe Is Actually On

The honest assessment, as of 2026, is that Europe is not building a DARPA. It is building a set of institutions that use DARPA's language while operating under fundamentally different structural constraints. The EIC funds promising startups but does not develop technologies to military specifications. The EDF develops military technologies but distributes funding through a process that prioritises industrial distribution over technical concentration. JEDI proposes to coordinate existing instruments but does not create a new institutional form. Each of these institutions serves legitimate purposes. None of them replicates the model that produced the internet, GPS, stealth technology, and mRNA vaccines.

This is not necessarily a condemnation. The DARPA model emerged from a specific American institutional context — a single sovereign government, a massive unified military, a political culture that tolerates government agencies operating with minimal bureaucratic constraint, and a revolving door between government, academia, and industry that European civil service traditions have historically resisted. Transplanting the model into the European context without addressing these structural differences would produce an institution that looked like DARPA but functioned like another Commission agency. The form without the substance would be worse than useless — it would create the illusion of transformation while delivering more of the same.

What Europe can do — and what the current generation of defence and industrial policy initiatives may, despite their limitations, eventually achieve — is create the conditions under which a genuine DARPA-like institution becomes possible. Increased defence spending creates a larger pot from which an autonomous agency could be funded. The European Defence Fund, for all its constraints, is building a community of cross-border defence research collaboration that did not exist a decade ago. The geopolitical pressure created by the war in Ukraine and the uncertainty of the transatlantic relationship is generating a political urgency around defence innovation that no amount of policy papers could have produced. The question is whether Europe's political leaders will use this moment to build an institution that genuinely operates differently — or whether they will, as they have done repeatedly, rebrand existing structures and declare the problem solved.

DARPA was born from Sputnik — from the shock of realising that an adversary had achieved a capability the United States did not possess. The agency was created not through incremental reform but through a rupture in institutional complacency. Europe may be approaching its own Sputnik moment. Russian aggression, Chinese technological competition, American strategic reorientation — the external pressures are mounting. The internal barriers remain formidable. The gap between what Europe says it wants and what its institutions are designed to deliver has never been wider. Closing that gap would require not just more money, but a different kind of institution — one that Europe has discussed for decades but never dared to build.

Sources

1. Bonvillian, W.B. & Van Atta, R. — "ARPA-E and DARPA: Applying the DARPA Model to Energy Innovation" — Journal of Technology Transfer, 2011 — https://link.springer.com/article/10.1007/s10961-011-9223-x
2. Draghi, M. — "The Future of European Competitiveness" — European Commission, September 2024 — https://commission.europa.eu/topics/strengthening-european-competitiveness/eu-competitiveness-looking-ahead_en
3. European Commission — European Defence Fund Regulation (EU) 2021/697 — https://eur-lex.europa.eu/eli/reg/2021/697/oj
4. DARPA — "Innovation at DARPA" — Defence Advanced Research Projects Agency, July 2016 — https://www.darpa.mil/attachments/DARPA_Innovation_2016.pdf
5. Hoeffler, C. & Music, G. — "European Defence Industrial Policy: Beyond the Headline Goal" — Journal of Common Market Studies, 2023 — https://onlinelibrary.wiley.com/journal/14685965
6. European Commission — "ReArm Europe / Readiness 2030" — Defence and Security Package, February 2025 — https://ec.europa.eu/commission/presscorner/detail/en/ip_25_793
7. Mazzucato, M. — "Mission Economy: A Moonshot Guide to Changing Capitalism" — Allen Lane, 2021 — https://marianamazzucato.com/books/mission-economy
8. European Innovation Council — Annual Report 2023 — https://eic.ec.europa.eu/eic-annual-report-2023_en