The German DARPA

In October 2019, an agency was founded in Leipzig with a mandate to accomplish the impossible: a German counterpart to DARPA — the American agency that gave the world the internet, GPS, voice recognition and autonomous vehicles. Six years later, SPRIND has neither failed nor arrived. It is something rarer: a state experiment that is genuinely learning. The question is whether that is enough.

I. The Original

The Defense Advanced Research Projects Agency was founded in 1958 — as an immediate response to the Sputnik shock. The Soviet Union had placed a satellite in orbit, and America had not seen it coming. The Department of Defense needed an institution that was structurally compelled to think the unthinkable.

What distinguishes DARPA from every other research agency is not its budget — though four billion dollars annually is no small figure. It is its operational philosophy. DARPA has no permanent laboratories, no fixed staff of researchers, no institutional memory that needs defending. Instead there are programme managers: highly specialised individuals on fixed terms of four to five years, with their own budget and the explicit permission to fail. A DARPA programme manager who terminates a project because it is not working is not penalised. He has done his job.

The second structural feature is the client. DARPA works for the Department of Defense. This is not a footnote. Military procurement creates a customer structure that civilian funding agencies do not have: a solvent client with an existential interest in the outcome. GPS was not a basic research project. It was a navigation system for nuclear missiles — with the side effect that thirty years later it runs on every smartphone.

DARPA does not work in spite of its military orientation. It works because of it. The existential pressure of the Cold War generated a risk appetite that no civilian funding programme can ever replicate — because in civilian programmes nobody has to die if the technology fails.

II. The Founding

Germany decided in August 2018 to build its own version. The impulse was not Sputnik but the growing recognition that the existing research system produces no radical innovations. DFG, Fraunhofer, Helmholtz — all fine for incremental improvement. For genuine breakthrough innovations, for technologies that make entire industries obsolete, an institution with appetite for risk was missing.

In October 2019, SPRIND was founded as a GmbH in Leipzig, led by Rafael Laguna de la Vera — an entrepreneur who had built and sold several software companies, and who therefore brought a different background from the outset than the typical research-bureaucracy profile. DARPA as a model was explicit — but with one important constraint: purely civilian, no defence applications.

The start was rocky. The Federal Court of Auditors ensured that the public service pay ceiling — the rule that state employees may not earn more than comparable civil servants — was applied despite contrary intentions. That made it impossible to hire programme managers of DARPA calibre. Angela Merkel publicly criticised the agency's structures in 2021 as permitting too little freedom of movement. She was right.

III. The Freedom Act and What It Changed

In December 2023, the Bundestag passed the SPRIND Freedom Act — and with it began the second, operationally relevant phase of the agency. The law gave SPRIND four new instruments: first, exemptions from the pay ceiling, enabling market-rate salaries and thus the recruitment of talent from industry. Second, the right to found subsidiary companies and acquire equity stakes directly — without lengthy ministerial coordination. Third, flexible budget management: unspent funds no longer lapse at the year's end. Fourth: revenues from equity sales flow fifty percent back into SPRIND's own budget.

This is a structural breakthrough — modestly formulated, but in the German agency landscape almost revolutionary. For the first time, a state funding body can act like a venture capitalist: invest, accompany, hold stakes, reinvest returns. Not as a subsidy, but as an entrepreneurial state.

IV. The Record after Six Years

What has SPRIND actually delivered by March 2026? The publicly available figures come from a parliamentary inquiry of July 2025 and the evaluation report of January 2025, commissioned from Fraunhofer ISI by the Federal Ministry of Education and Research.

What stands out is the throughput speed: from application to funding commitment, SPRIND reports an average of 14 days — not months or years as with conventional project management agencies. This is the real core difference from the rest of the system, and it is genuine. Anyone who has ever submitted a DFG application knows what 14 days means.

In terms of content, the portfolio ranges from a new Alzheimer therapeutic to a solution for removing microplastics from waterways. SPRIND has founded a subsidiary developing high-power lasers for nuclear fusion projects — and holds stakes in the startups Marvel Fusion and Focused Energy, which are working with Siemens Energy on a next-generation power plant. This is not an incremental funding programme. This is genuine leap thinking.

V. The Evaluation and Its Problem

In January 2025, Fraunhofer ISI published the first evaluation report — positive. The agency had built flexible structures, was working purposefully towards its mission, and was closing a genuine gap in the German innovation system. Parliamentary State Secretary Claudia Müller called SPRIND a success story.

One need not contest this assessment to identify a structural weakness of the evaluation: Fraunhofer ISI was commissioned by the Federal Ministry of Education and Research to evaluate the Ministry's own funding project, SPRIND. This is not bias in the legal sense. It is the systemic tendency already described in this essay: the system evaluates itself. The metrics that characterise SPRIND as a success are largely input metrics — projects funded, companies founded, competitions conducted. Whether in ten years any of these investments has achieved the economic significance of GPS or the internet, no one knows. And that is the nature of the experiment.

The fair assessment is therefore: SPRIND is doing the right things, measured by the wrong standards, in a system that has structurally little patience for long-term experiments. Whether that suffices will become clear in a decade at the earliest.

VI. The Scale Problem

DARPA invests roughly four billion dollars annually. SPRIND has 200 million euros. That is a factor of twenty — even after accounting for purchasing power differences and the relative sizes of the two economies. The difference is not merely quantitative. It determines what kind of bet you can place.

DARPA can simultaneously fund thirty high-risk programmes, expecting twenty to fail. Three of them suffice to justify the entire investment. SPRIND cannot afford this kind of diversification. Each project carries more expectational weight because it binds a larger proportion of the total budget. This creates — despite every contrary intention — a selection pressure towards safer, more communicable projects.

Rafael Laguna has said as much, if in different words: Germany needs not only SPRIND but a digital agency with comprehensive powers and significantly more funding. Not a digital ministry — that takes an entire legislative term to assemble from departments that refuse to cede anything. A new organisation with its own cultural mandate. Whether the new federal government will deliver this is, given past patterns, doubtful.

VII. The Structurally Excluded

Here is the most uncomfortable question: does SPRIND reach the type of inventor it was ostensibly designed for?

The figures give a cautious answer. All 72 funded projects originate from universities, non-university research institutions, or — in the most favourable cases — from private individuals. The category "private individuals" sounds like openness. In practice it means: persons who already have a team, who can already point to a track record, who already speak the language of the system. The lone inventor who has spent years working without an academic network on a technology the existing system cannot categorise — he is absent from these figures.

The cartel of ignorance that the former President of the German Patent Office, Erich Häusser, described as far back as the mid-1990s has not been dissolved by SPRIND. Häusser put it precisely: inventors who do not belong to the establishment are not supported and promoted, but neglected, often treated badly, or — what is almost worse — simply ignored. Thirty years later, SPRIND is an improvement. But the five words that summarise the system have not yet disappeared: "Do you even have a company?"

As long as SPRIND is not structurally capable of identifying and accompanying the inventor without a company, without a network, without academic affiliation, it remains what it set out to overcome: an institution that supports innovators who are already inside the system — and leaves everyone else to fight alone.

VIII. What SPRIND Actually Is

SPRIND is neither a misunderstanding nor a misallocation of funds. It is a genuine experiment — smaller than its model, operating in a less risk-tolerant environment, with a budget that does not permit genuine diversification, but with an operational logic that actually challenges the German funding system. That is more than most new agencies manage.

Three things about SPRIND are structurally right: the speed (14 days instead of 14 months), the willingness to found companies rather than merely distribute grants, and the explicit right to fail. These three elements are without precedent in the German funding system — and they suffice to distinguish SPRIND from every other agency.

Three things remain structurally wrong: the budget (too small for genuine high-risk diversification), the target population (still too close to the institutionalised innovator), and the benchmarks (input figures rather than long-term impact measurement). The third problem resolves itself — if SPRIND still exists in twenty years and one of its nuclear fusion subsidiaries is feeding electricity into the grid, the evaluation will have written itself.

DARPA did not invent the internet because America was especially clever. It invented the internet because it was willing to bet on technologies that might never work — with a patience that no democratic parliament volunteers freely. With SPRIND, Germany has taken a first step in that direction. Whether that step is sufficient will not be decided by any evaluation. It will be decided by time.