“Basic research is the pacemaker of technological progress.” — Vannevar Bush, Science, the Endless Frontier, 1945[^bush1945]

Overarching epistemic status: Probable. This essay applies the Orbis Futura Epistemic Labelling Protocol throughout. Assertions are tagged inline as [Confirmed] (source-traceable to primary documents, named researchers, or documented programmes), [Probable] (strong supporting evidence, not fully determinative), [Possible] (suggestive evidence, significant uncertainty), or [Speculative] (hypothesis with limited direct evidence). The re-reading of Bush in Parts I–III is argument, not prediction. The three case studies in Part V are drawn from the open record. Every predictive claim attached to the framework in Part VI is flagged as the author’s original analysis and rated accordingly. All sources are referenced in the Oxford documentary-note system, with footnotes and a closing bibliography.

From the Review

The most cited document in the history of American science policy is also the one we have most reliably misread. Vannevar Bush’s Science, the Endless Frontier is invoked in nearly every hearing on research funding, claimed by university presidents and venture capitalists alike, and so canonical by the 2020s that Congress simply borrowed its name for a flagship competition bill. Yet the thing we remember about the report is almost exactly wrong. We remember a funding memo — a 1945 argument that Washington should pay for basic science — and treat everything else as preamble. Bush’s actual achievement was architectural. He did not propose a programme. He described a living system: an arrangement of talent, institutions, incentives, and flows of knowledge whose health could never be reduced to a single line in a budget. He was, in the vocabulary I have used for years to describe what makes nations indispensable, describing an ecosystem — and the word matters, because ecosystems fail in ways budgets cannot see.

This is not an antiquarian quarrel. It bears directly on the problem that should occupy a publication like this one: how breakthrough laboratory science becomes operational military capability, and why, today, it so often does not. Bush built his architecture for a world in which the government was the dominant customer for advanced technology and discoveries flowed outward from the laboratory to industry to the field. That world is gone. The most consequential dual-use technologies of the 2020s — drones, autonomy, artificial intelligence, microelectronics — are now commercial-first, and the binding constraint is no longer the supply of fundamental knowledge but the architecture that moves it across the gap between the lab and the battlefield. To read Bush correctly is to see that he gave us a manifesto for ecosystem health, and that the task now is not to recite it but to edit it forward into a century he did not live to see.

I. The Memo That Was a Manifesto

In July 1945, Vannevar Bush — engineer, former vice-president of MIT, and the man who had run American scientific mobilisation through the war as director of the Office of Scientific Research and Development — delivered a slim report to the President. Franklin Roosevelt had commissioned it the previous November; by the time it arrived, Roosevelt was dead and Harry Truman had inherited it.[^bush1945] The title was a small masterstroke. Science, the Endless Frontierlifted the central myth of American self-understanding — the frontier as the engine of national renewal, declared closed by Frederick Jackson Turner in 1893 — and reopened it. The geographic frontier was gone; here was a new one, inexhaustible, that a nation could keep settling forever. [Confirmed]

Eighty years later, the title has outrun the text. When Senators Chuck Schumer and Todd Young introduced their flagship technology bill in 2020, they called it the Endless Frontier Act, a deliberate echo meant to summon Bush’s authority for an era of competition in artificial intelligence, quantum computing, and semiconductors. The bill was folded into the broader CHIPS and Science Act, signed in 2022.[^fas2022] The lineage was explicit and proud. And yet, as the Federation of American Scientists observed in reviewing the final law, almost the entire public conversation around the bill fixated on a single number — how much money it authorised — while the parts of Bush’s original report that had nothing to do with topline funding, and everything to do with the cultivation of scientific talent, went largely unmentioned.[^fas2022] [Confirmed] The homage reproduced the misreading. We named a law after a manifesto and then argued about its price tag.

This is the standard fate of the report. It is treated as the founding charter of federal research funding — the document that, by intellectual shorthand, “invented” the idea that government should pay for university science — and on that reading its content collapses into one sentence: fund basic research; the applications will follow. The shorthand is not baseless. Richard Atkinson, a former director of the National Science Foundation, rightly identified the report’s most far-reaching recommendation as the proposition that it was in the national interest for the federal government to fund university research, and credited it with moving research universities from the periphery of the American system to its core.[^atkinson1997] [Confirmed] But Atkinson also described something the shorthand omits: that Bush’s model was a national enterprise in which basic research, federally supported and conducted in universities under norms set by scientists, would then be implemented by private industry — a structured division of labour, not a cheque.[^atkinson1997]

The cost of forgetting the structure is not merely interpretive. By the report’s 75th anniversary in 2020, even admiring reassessments were emphasising what the funding reading omits: writing in Science, Emily Gibson called the report “the Magna Carta of American science” while arguing that the apolitical, autonomous ideal it enshrined had always been partly a myth.[^gibson2020] [Confirmed] That is the predictable result of remembering only the budget line: if you believe Bush handed you a subsidy, you will judge him by whether the subsidy still works, and you will miss the part of the document that was never about money at all. To recover it, you have to do the unfashionable thing and read the report’s structure rather than its slogan. The structure tells a different story.

II. The Architecture Bush Actually Drew

Roosevelt’s commissioning letter of 17 November 1944 did not ask Bush one question. It asked four. The first concerned the diffusion of knowledge: what could be done, consistent with military security, to release the scientific contributions made during the war? The second concerned mission: with particular reference to the war of science against disease, how should the country organise to continue medical research in peacetime? The third was the famous one, about means: what could government do to aid research by public and private organisations? And the fourth concerned people: could a programme discover and develop scientific talent in American youth, so that the nation’s research capacity would be secure for generations?[^bush1945] [Confirmed]

These four questions are the bones of the report, and they are not four versions of how much should we spend. Bush answered each through a dedicated committee — on the diffusion of wartime knowledge, on medical research and the public welfare, on the support of research, and on the discovery and development of scientific talent — and bound their findings into a single argument.[^bush1945] [Confirmed] Read them in sequence and the shape of his thinking is unmistakable. Knowledge must keep moving and not stay locked up; research must be oriented toward human welfare and not only toward curiosity; the work must be supported broadly and stably; and the supply of the people who do the work must be continuously renewed. Funding is one of these four concerns. The other three are the conditions under which funding does anything at all.

It helps to remember who was answering. Bush was not a theorist of science policy. He was the man who had just run the largest applied-science effort in human history. Under his Office of Scientific Research and Development, university and industrial laboratories had produced the radar that decided the air war, the proximity fuze, the mass production of penicillin, and the early organisation of the atomic project. [Confirmed] He had spent four years watching fundamental physics become a weapon and fundamental chemistry become a cure, on timelines compressed by emergency. When such a man writes that basic research is the nation’s strategic reservoir, he is not romanticising the laboratory; he is reporting what he learned converting it into matériel. This biographical fact is what makes the most durable charge against him — that he saddled American science with a naïve “linear model,” in which pure research flows in one direction to applications like water down a pipe — so implausible.

The careful scholarship bears this out. Dennis Leyden and Matthias Menter, examining Bush’s writings and life, conclude that he believed in a more sophisticated model than the one attributed to him: one in which basic and applied research cross-fertilise each other, and in which the government’s job is less to stimulate basic research in isolation than to facilitate the interaction between basic and applied work for the benefit of both.[^leyden2018] [Probable] The point is consonant with Donald Stokes’s well-known reframing in Pasteur’s Quadrant: that the basic-versus-applied dichotomy is false, that the most transformative research is use-inspired — fundamental questions pursued because they bear on urgent practical problems — and that Bush’s own text sits far closer to Pasteur’s quadrant than to the caricature.[^stokes1997] [Confirmed] A field that spent decades refuting a model Bush never held was, in effect, arguing with a slogan rather than the document.

Underneath all of it ran the metaphor that does the real work — not the frontier, but the reservoir. Bush described basic research as “scientific capital,” a fund of fundamental knowledge from which all practical applications must ultimately be drawn, and warned that the United States had been living off borrowed capital — the basic science of a Europe the war had shattered — that would have to be replenished at home.[^bush1945] [Confirmed] This is an ecological way of thinking, not an accounting one. It treats knowledge as a renewable resource with a stock and a flow, and it asks the questions a forester asks, not the questions a cashier asks.

Consider, too, the organ Bush treated as a prerequisite rather than an afterthought: talent. He argued, in effect, that money spent on research is wasted without enough trained minds to absorb it — that the binding constraint on a scientific enterprise is rarely dollars and almost always people. The war had interrupted the education of a generation and hollowed out the universities, and Bush wanted scholarships and fellowships not as a charitable supplement to the research budget but as the headwaters of the whole system. A nation that funds laboratories without growing scientists is pouring water into a reservoir with no inflow. This is precisely the part of the report the Federation of American Scientists identified as forgotten when Congress revived Bush’s title in 2022, and its omission from the modern conversation is not a small editorial loss.[^fas2022] It is the loss of the load-bearing idea. [Confirmed]

Bush did not, in the end, get the system he drew. The single coordinating National Research Foundation he proposed became the object of a five-year fight with Senator Harley Kilgore, who wanted a science agency accountable to elected officials, oriented toward explicit national goals, and distributed across the country rather than concentrated in the elite universities. Truman vetoed the first bill in 1947, largely over whether the agency’s director would answer to the President or to a board of scientists. By the time the National Science Foundation emerged in 1950, it was a fraction of Bush’s conception — smaller, later, and stripped of the medical research and sweeping authority he had imagined.[^itif2025] [Confirmed] The quarrel was not a petty turf war; it was a genuine disagreement about ecosystem design, and both men were partly right. Kilgore feared an unaccountable guild serving its own prestige; Bush feared an enterprise whose priorities would be reset every election and whose autonomy would slowly be strangled. History has vindicated both anxieties at different moments, and the tension between them — insulation versus accountability, curiosity versus mission, concentration versus distribution — was never resolved so much as distributed across the agencies that filled the vacuum: the Office of Naval Research from 1946, the expanding National Institutes of Health, the Atomic Energy Commission, and, after Sputnik, the agencies that became DARPA and NASA. [Confirmed] The ecosystem assembled itself out of parts Bush had sketched as one organism. That it worked at all is a testament to how right he was about the components.

III. The Unraveling

By the end of the century, the people closest to the system were warning that it was coming apart at exactly the joints Bush had identified. In 1999 the federal government convened a formal interagency review, Renewing the Federal Government–University Research Partnership for the 21st Century, precisely because the postwar compact’s principles had never been written down and were fraying in practice.[^nstc1999] [Confirmed] The sharper alarm came from within the laboratories. In 2014, four of the most senior figures in American biomedical science — Bruce Alberts, Marc Kirschner, Shirley Tilghman, and Harold Varmus, among them two former NIH directors and a university president — argued in the Proceedings of the National Academy of Sciences that US research had become trapped in a set of systemic flaws: hypercompetition, perverse incentives, an unstable reliance on soft money, and a training pipeline producing far more scientists than the system could absorb.[^alberts2014] [Confirmed] The complaint, in both cases, was not that the money had stopped. It was that the architecture — the implicit understanding about autonomy, stability, talent, and the division of labour between government, university, and industry — was eroding even as appropriations held.

Paula Stephan gave the same erosion its most rigorous economic anatomy. In How Economics Shapes Science, she showed a system whose incentives had turned against its own purpose: stagnant budgets and expanding universities had made funders and scientists alike risk-averse, pushing investigators toward safe projects rather than the uncertain, potentially path-breaking work the system exists to support.[^stephan2012] [Confirmed] Read against Bush, the diagnosis is damning: the damage falls on his four organs — the talent watershed, the stability of institutions, the freedom to take long bets — and none of it shows up in the topline number the public conversation treats as the whole of science policy.

Then came the contraction that turned a chronic condition into an acute one. Through 2025 and into 2026, the United States entered the most severe disruption of its research-funding architecture in the system’s history. Roughly three billion dollars in active grants were terminated; the indirect-cost rates that sustain the institutional infrastructure of university research were cut toward fifteen per cent; federal science agencies shed staff; and successive budget proposals sought to cut the National Science Foundation by something on the order of half its appropriation, with deep reductions at the National Institutes of Health and the proposed elimination of entire institutes.[^cen2025][^cen2027] [Confirmed] What is striking, through the lens Bush gives us, is not the dollar figures but which organs took the damage. Indirect-cost cuts strike institutional stability — the banks that hold the reservoir. Topic-targeted terminations strike the insulation of scientific judgment from political priority. The redirection of funding toward favoured fields strikes the diversity of the system. And the chill on careers — departures, and international researchers reconsidering the United States while the European Union and China expand their own fundamental-research investment — strikes the talent watershed, the slowest organ to fill and the slowest to recover.[^penn2026] [Probable] The erosion diagnosed in 1999, and again in 2014, did not reverse. It accelerated.

IV. The Gap Bush Left: Defence Technology in the Twenty-First Century

Here the historical argument must give way to a contemporary one, and I want to mark the transition honestly. The analysis in this part and the next is my own. It extends Bush’s ecosystem logic into a domain his report did not address and the existing literature does not adequately analyse — the dynamics of twenty-first-century defence technology — and its conclusions should be read at the epistemic level indicated, not as established findings. A survey of the available scholarship confirms the gap: the sources offer moderate evidence on Bush’s structural vision and on the contemporary critiques of it, but they do not specifically analyse how that framework fails under modern defence-technology dynamics. [Speculative — author’s original analysis]

Bush’s architecture rested on three assumptions that were true in 1945 and are largely false today. The first was that the government was the dominant customer for advanced technology, so that public funding decisions effectively steered the frontier. The second was that knowledge flowed outward from the laboratory — basic research first, then applied, then industrial production, then, eventually, deployment in the field. The third was that the cycle was slow enough for institutions to manage deliberately. In the dual-use technologies that now decide military advantage — autonomy, artificial intelligence, commercial space, advanced microelectronics, small drones — none of these assumptions holds. The Pentagon is a minority buyer in markets it once commanded; the most important capabilities are developed commercially and adapted to military use rather than the reverse; and the cycle is fast enough that a procurement system built for deliberation is structurally too slow to keep pace.

Consider how completely the flow has inverted. The small quadcopters that now define the tactical edge of the modern battlefield descend from a consumer-electronics supply chain the Pentagon did not build and does not control; for years the dominant manufacturer was a Chinese commercial firm. The reusable rockets that collapsed the cost of reaching orbit were developed by a private company selling launches on the open market, with the government as one customer among many. The foundation models that anchor military interest in artificial intelligence were trained in commercial laboratories chasing commercial returns. In each case the capability matured outside the defence system and had to be drawn into it — the precise reverse of the outward flow Bush assumed. A 1945 architecture optimised to push knowledge from the laboratory toward the field is, in these markets, pointed the wrong way. [Confirmed]

This inversion does not make government irrelevant; it relocates its leverage. In a commercial-first ecosystem the state’s decisive lever is its conduct as a customer — whether it can buy quickly, commit credibly, and integrate what the market already produces before the technology turns over. The centre of gravity of defence-innovation policy has shifted from the laboratory toward the procurement office, and most procurement offices were designed for a world that no longer exists. [Speculative — author’s original analysis]

The binding constraint, in short, has moved. The scarce resource is no longer fundamental knowledge but transition — the capacity to move a capability across the gap between a working prototype and a fielded system, the chasm that defence practitioners have long called the “valley of death.” A framework that measures ecosystem health by how much basic science it funds is measuring the wrong organ. The relevant question is no longer only how full is the reservoir but how patent are the channels that carry its contents to the field, and how fast.

The mismatch is most visible in the clocks. A major defence acquisition programme is measured in years, and often decades, from requirement to fielding; a commercial software or drone-design cycle is measured in weeks or months. When the underlying technology turns over several times within a single procurement cycle, a system optimised to specify a requirement precisely and then build to it will field, at great expense, a capability already a generation behind by the time it arrives. That gap is the valley of death in operation. It is not, at root, a funding problem. It is an architecture problem, and no increase in the topline can repair an architecture pointed the wrong way and running on the wrong clock. [Probable]

This is the editorial change the title of this essay proposes: not a new clause appended to a 1945 document, but a shift in what we count as ecosystem health when the customer, the direction of flow, and the clock have all changed. [Speculative — author’s original analysis] The three cases that follow are offered as evidence that the shift is real.

The paradox, in numbers

Strip the argument to its measurements and the paradox is stark: capital has never been more abundant, and transition has rarely been slower. The financial gravity has shifted decisively toward the commercial sector. By PitchBook’s reckoning, private venture capital poured roughly US$135.3 billion into defence technology across 4,744 deals between 2016 and 2022, a market it projected to reach US$184.7 billion by 2027 at a compound annual growth rate near sixteen per cent.[^pitchbook2023] [Confirmed] American defence-startup volume has run well ahead of Europe’s — by a factor of about 2.4 over 2021–2024, in the baseline compiled for this essay.[^baseline] [Probable] Tier-one firms that once shunned the sector — Sequoia, Andreessen Horowitz, Bessemer, Y Combinator — are now fixtures in early-stage defence investing. The reservoir, in Bush’s terms, is now filled largely by private money chasing dual-use returns; the government is no longer the financier of first resort but one customer among many.

And yet the public side spends heavily and fields slowly. The United States appropriated nearly US$146 billion for defence research, development, test, and evaluation in fiscal year 2026 — about fifteen per cent of the defence budget — while transition rates into formal programmes of record remained stubbornly low.[^baseline] [Probable] The single most telling figure is a clock running backwards. In its 2025 annual assessment of weapon systems, the Government Accountability Office found that the expected time for a Major Defense Acquisition Program to deliver even an initial capability had lengthened by eighteen months in a single year, to almost twelve years from a programme’s start — roughly 124 to 142 months.[^gao2025] [Confirmed] In the very period when every official speech invoked speed, the system added a year and a half to its delivery time. That is not a funding shortfall. It is an architecture moving in the wrong direction.

The structural causes are documentable. Promising technologies stall at low readiness levels for want of clear demand signals; entrenched prime contractors capture procurement to such a degree that, as of 2024, the primary-contractor rosters of the United Kingdom, Germany, and Poland reportedly held only two startups between them; and requirements systems such as the Joint Capabilities Integration and Development System grew so rigid that a serious push emerged in late 2025 to disestablish them outright in the name of speed.[^baseline] [Possible] The empirical record says precisely what the cases will show: the binding constraint is neither capital nor science. It is the channel.

V. Three Readings from the Open Record

None of the following is a controlled experiment. Each is an episode from the open, unclassified record, chosen because it isolates a different relationship between ecosystem architecture and the transition of laboratory capability into operational use. The facts are source-traceable; the interpretive frame I lay over them is mine.

The F-35, or architecture defeating resources

The F-35 Joint Strike Fighter is the most expensive weapons programme in American history — its projected lifetime cost, acquisition plus decades of sustainment, runs past US$1.7 trillion — and the standard account treats its troubles as a story of technical ambition outrunning engineering. [Confirmed] The more instructive reading, drawn from the case literature, is that its failures were architectural rather than technological. The programme’s difficulties stemmed less from the physics of a single airframe serving three services and allied air forces than from misaligned organisational incentives, procurement inflexibility, and ineffective governance among Lockheed Martin, the Department of Defense, and partner nations.[^hughes2014] [Confirmed] The ecosystem, in other words, had abundant resources and abundant talent and still could not reliably convert them into capability, because its structure worked against it.

The architecture worked against it in specific, documentable ways. The decision to build a single airframe to satisfy the Air Force, Navy, and Marine Corps at once produced a jet that compromised each service’s requirements rather than meeting any of them cleanly. The choice to manufacture aircraft concurrently with flight testing — “concurrency,” in the programme’s own vocabulary — meant that defects found in testing had to be retrofitted across jets already built, turning every surprise into a cascade of rework. And the software, on which a modern combat aircraft depends as much as on its engines, grew into tens of millions of lines of code whose logistics and maintenance systems became troubled programmes in their own right. None of these is a failure of physics or of money. Each is a failure at the joints — of governance, of incentives, of the refusal to trade a zero-error ideal for iterative learning.[^patil2020] [Probable]

The deeper diagnosis concerns a cultural choice. Defence procurement has tended to prize what one analysis calls a “zero-error” standard over a “good-enough” one — an institutional preference for exhaustive specification and the elimination of risk over iterative learning and rapid feedback.[^patil2020] [Probable] A zero-error culture produces rigid processes that suppress exactly the rapid prototyping and real-world feedback loops on which fast-moving technology depends. The contrast with what the research literature identifies as the enabling conditions for defence innovation — flexible governance, prototyping authority, and testing infrastructure that lets scientific progress transition into sustained operational advantage — is sharp.[^mulyanyuma2026] [Possible] The F-35 is the clearest demonstration in the modern record that an innovation ecosystem can possess every input Bush prized — money, science, industrial capacity — and still fail at the joint, because the architecture that governs transition is rigid where it needs to be adaptive.

Ukraine, or adaptation under constraint

The opposite case is being written, at speed and under fire, in Ukraine. From 2014 and intensively after 2022, the Ukrainian defence sector underwent a digital transformation that reshaped its innovation architecture deliberately to compress the distance between laboratory, startup, and front line.[^korendovych2026] [Probable] Systems such as the DELTA situational-awareness platform, the Kropyva artillery-fire software, and the broader “Army of Drones” initiative worked not because Ukraine commanded greater resources — it commanded far fewer — but because the architecture aligned technology, people, and process through federated governance that connected academic research, commercial startups, and military operators directly.[^korendovych2026] [Probable] The valley of death was narrowed by building tight feedback loops between the warfighter’s requirements and the developer’s iterations, abandoning the patchwork automation that elsewhere keeps prototype and field apart.[^lee2025] [Possible]

The architecture was built deliberately, not merely improvised under pressure. Ukraine stood up a dedicated defence-technology cluster, Brave1, to connect developers, the military, and the state; it reformed acquisition so that units could obtain drones far more directly than a conventional procurement chain allows; and it scaled domestic production of first-person-view attack drones from a cottage effort into an output measured in the millions per year. The decisive feature was cadence. In the electronic-warfare contest over the front, a drone design effective one month could be jammed the next, which forced developers into iteration cycles measured in weeks — hardware and software revised continuously against live battlefield feedback. That is feedback fidelity in its most extreme, war-forced form, and it is the mechanism behind the speed, more than any single platform. [Probable]

The structural shift left a measurable economic trace. By open-source accounts, Ukrainian defence technology grew from near-zero market share in 2022 to roughly thirteen per cent — on the order of fifty-nine million US dollars — by 2024, even as the broader startup sector recovered to around four hundred and sixty-two million dollars.[^butenko2026] [Possible] I record those figures at the epistemic level they deserve: they rest on a single recent source and should be treated as suggestive rather than settled. But the qualitative lesson is robust and matters more than the precise numbers. A dual-use ecosystem can transition capability with extraordinary speed when institutional barriers are removed and private capital is aligned with national priority — and, soberingly, the catalyst here was existential necessity, which is not a policy instrument any peacetime system can deliberately reach for. The Ukrainian case shows what an unobstructed channel looks like. It does not show how to build one without a war.

The American military–civil interface, or institutional diversity as resilience

The third case is the United States’ own attempt to engineer transition deliberately, through a layered set of institutions: DARPA, the Defense Innovation Unit, and venture mechanisms such as In-Q-Tel. Each was created to solve a different facet of the transition problem: DARPA, since 1958, to fund high-risk bets toward a mission; In-Q-Tel, founded by the Central Intelligence Agency in 1999, to invest as a venture capitalist in startups building intelligence-relevant technology; and the Defense Innovation Unit, established in Silicon Valley in 2015, to court commercial firms on their own terms and timelines rather than forcing them through traditional procurement. [Confirmed] The literature reads this not as a single fix but as a case of structural adaptation succeeding despite persistent friction — differing organisational cultures and timelines between government and private firms, regulatory complexity, intellectual-property conflicts, and genuine ethical concerns about military application.[^stepanov2025] [Probable] What is instructive is the response to that friction. Rather than forcing every technology through one rigid pathway, the American system created several: traditional procurement, research grants, fast-track innovation programmes, and dual-use venture vehicles. That redundancy lets a capability follow the route suited to its maturity and context.

DARPA remains the sharpest illustration. By systematically asking which commercial innovations can be adapted for defence, and pairing multi-year funding with organisational flexibility, mission relevance, and genuine public–private partnership, it functions as an effective transition mechanism rather than merely a funder of science.[^mulyanyuma2026] [Possible] This is the part of the story that complicates any simple reading of Bush. DARPA was not the autonomous, curiosity-driven, peer-reviewed institution he designed; it is mission-driven, manager-led, and aggressive — a different organ of the same ecosystem, supplying the directed integration his architecture lacked. Its persistence suggests the central design lesson: institutional diversity is not waste. It is resilience. The accidental disassembly of Bush’s single foundation into a scatter of agencies produced an ecosystem with multiple transition pathways and multiple definitions of fundable work — and that very redundancy is what lets the system route around its own blockages. The friction never disappears; tensions over IP, ethics, and great-power competition in AI and microelectronics persist.[^stepanov2025]

The most rigorous evidence that this bottom-up, multi-pathway approach actually works comes from the economists Sabrina Howell, Jason Rathje, John Van Reenen, and Jun Wong, who studied the Air Force’s shift from “conventional” research solicitations — which specify the technology wanted — to “open” topics, which invite firms to propose anything useful. In a regression-discontinuity design, the first causal evaluation of a defence R&D programme of its kind, they found that winning an open, bottom-up award measurably increased a firm’s subsequent venture-capital investment, non-SBIR defence contracting, and patenting, while conventional awards did none of these and instead reinforced incumbency.[^howell2021] [Confirmed] A subsequent Government Accountability Office review found the same pattern in the Air Force data — open-topic awards drew in firms with no prior federal contracts at three times the rate of conventional awards, and were issued far faster — while noting that only about two per cent of the near-miss non-awardees went on to secure venture capital, against a markedly higher rate for those selected.[^gao2022] [Confirmed] In the baseline compiled for this essay, that relative venture uplift is put at 178 per cent.[^baseline] [Probable] This is the empirical heart of the channel argument: a credible government demand signal is itself a transition mechanism, and the scarce resource is less capital than a buyer who can commit.

The same logic now drives the allied response. NATO’s Defence Innovation Accelerator for the North Atlantic (DIANA), established in 2022, and the NATO Innovation Fund, a deep-tech venture vehicle backed by twenty-four allied nations, exist precisely to carry dual-use technology across the valley of death alliance-wide.[^mcsorley2025] At the Hague summit in June 2025, NATO heads of state endorsed a Rapid Adoption Action Plan that sets a hard target — adopt and field a new technology within twenty-four months of identifying the need — while explicitly committing to greater “acquisition risk” and a “fail fast” posture; the European Defence Fund, in parallel, earmarked roughly €7.3 billion for 2021–2027 to accelerate such capabilities.[^nato2025][^baseline] [Confirmed] Whether a twenty-four-month target survives contact with thirty-two national bureaucracies is, fairly, an open question. But the direction is unmistakable: the West has begun, belatedly, to treat the channel as the thing to engineer rather than the thing to assume. [Probable] A system with many channels survives the closure of any one of them.

VI. An Updated Framework: Six Vital Signs of a Defence-Innovation Ecosystem

The framework below is my original analysis. It synthesises the Bush reading with the three cases and is offered as a diagnostic instrument, not a description of consensus. Each predictive claim is flagged and rated; readers should weight them accordingly. [Speculative — author’s original analysis]

If a defence-technology enterprise is an ecosystem rather than a budget line, the question is what its vital signs are — what we should measure to know whether it is healthy, and where the early symptoms of failure appear. Bush gave us four organs; the twenty-first-century cases add two more and change the weighting of all of them. The organising intuition is that these ecosystems fail at the joints rather than at the budget line, and that each joint fails on its own clock. Money can be cut and restored within a single appropriations cycle; a drained reservoir of fundamental knowledge refills over a decade; a discouraged talent cohort may never return at all. A framework worth having must respect those different clocks, because the gravest errors in this field are the ones whose consequences arrive long after the officials who made them have moved on.

Reservoir depth. This is Bush’s scientific capital: the stock of fundamental knowledge and the rate of its replenishment against its drawdown. Its defining property is delay; investments in basic research pay out over years to decades, which makes reservoir depth the vital sign least legible to short-term observation. Prediction: the deep reservoirs accumulated since 1945 are deep enough that the 2025–26 contraction will not visibly reduce scientific or operational output for somewhere between five and fifteen years — and that very invisibility is the danger, because the political system will read the absence of immediate harm as proof the cut was harmless, and repeat it. [Speculative — author’s original analysis]

Channel patency. A reservoir is useless if its outflows are blocked. In Bush’s world this was a secondary concern, because the government was the customer and the flow was outward; in the defence ecosystem of the 2020s it is the primary one, because the reservoir is largely commercial and the scarce resource is transition across the valley of death. The F-35 shows a blocked channel with abundant resources; Ukraine shows an open one with scarce resources. Prediction: among ecosystems of comparable funding and talent, channel patency — measured by the speed and reliability with which a working prototype becomes a fielded capability — will predict operational advantage better than reservoir depth does. [Speculative — author’s original analysis] The gauge is already moving the wrong way: a Major Defense Acquisition Program clock that lengthened by eighteen months in a single year, toward almost twelve years to first capability, is channel patency in measurable decline, whatever the topline shows.[^gao2025] [Confirmed]

Niche diversity. The American case demonstrates that multiple institutional forms and funding philosophies — autonomous grants beside mission-driven programme management, procurement beside venture capital — make a system resilient because a capability that stalls in one pathway can migrate to another. A monoculture is efficient in the short run and fragile in the long run. Prediction: ecosystems with diverse transition pathways will absorb shocks — a budget cut, a political reprioritisation, an institutional collapse — markedly better than concentrated ones; the long erosion of corporate basic research and the current concentration of stress on the federal niche have already reduced this diversity and raised systemic fragility. [Speculative — author’s original analysis] The warning is already legible in Europe, where as late as 2024 the primary-contractor rosters of the United Kingdom, Germany, and Poland reportedly held only two startups between them — a procurement monoculture that DIANA and the NATO Innovation Fund were, in effect, created to break.[^baseline] [Possible]

The talent watershed. Bush’s fourth and most neglected question — where the scientists and engineers come from, and who keeps the pipeline flowing — is the system’s inflow, slow to fill and slow to recover. Stephan’s anatomy of risk aversion, PhD oversupply, and ever-later independence shows the watershed was already stressed before the recent cuts.[^stephan2012] Prediction: the most consequential and least reversible damage an ecosystem can suffer is to its talent watershed, because a discouraged cohort does not return — a researcher who leaves, an international scientist who chooses another country, a student who never enters represents capacity that cannot be restored by restoring next year’s budget. Funding toggles; careers do not. [Speculative — author’s original analysis]

Governance insulation with accountability. This is the Bush–Kilgore tension, unresolved by design: how to spend public money on a public good without letting the political cycle micromanage it into mediocrity, while keeping it accountable to public purpose. Too much insulation yields an unaccountable guild; too little yields an arm of whichever priority currently holds power. Prediction: the failure mode to watch in the current period is not under-funding but the politicisation of selection — the redirection of what counts as fundable or fieldable work according to the priorities of the moment rather than technical merit — which damages reservoir depth and talent retention simultaneously and is far harder to detect in budget figures than a simple cut. [Speculative — author’s original analysis]

Feedback fidelity. The one vital sign Bush could not have named, because the tools did not exist: the system’s capacity to observe its own state and correct course before consequences arrive on the reservoir’s decade-long delay. Ukraine’s compressed warfighter-to-developer loops are feedback fidelity in its most extreme, war-forced form; the F-35’s suppressed feedback is its absence. Prediction: of the six vital signs, feedback fidelity is the highest-leverage place to invest, because its improvements compound — a system that can see itself can repair the other five, while a system blind to its own state will keep cutting the reservoir and the watershed and calling it prudence, precisely because their damage is invisible on the timescale at which decisions are made. [Speculative — author’s original analysis]

The diagnostic value of the framework is that it disaggregates one misleading number into six independent organs, each with its own failure mode and its own clock. A system can have a rising budget and a collapsing watershed; deep reservoirs and blocked channels; generous funding and politicised selection. The funding-memo reading sees one dial. The ecosystem reading sees a dashboard — and on a dashboard you can watch the gauge that is failing while the others still look fine.

VII. Edited Forward

Return, then, to the irony that gives this essay its title. In 2022 the United States passed into law a bill descended directly from one named after Bush’s report, authorising historic sums for the kind of research he championed.[^fas2022] The homage was sincere, even as it reproduced the old fixation on the topline. And then, within three years, the system that law was meant to strengthen entered the deepest contraction in its history — and, read through the six vital signs, the contraction has fallen first on the invisible organs: on institutional stability, on the insulation of judgment, on the diversity of pathways, and on the talent watershed.[^cen2025][^cen2027][^penn2026]

A Bush-faithful reading counsels something specific here, and it is not the reflexive defence of every existing dollar. It is a reordering of what we worry about. The funding-memo frame fights over the topline and treats the rest as detail. The ecosystem frame insists the topline matters less than the joints — and the joints are visibly stiffening, in a major-program clock that lengthened rather than shortened even as speed became the official watchword. My forecast for the current period is that its most serious damage will be invisible for a decade and will fall not on the budget but on the watershed and the reservoir — and that, in the defence domain specifically, the decisive vulnerability is not how much basic science we fund but whether our channels can move commercial capability to the field faster than our rivals move theirs.[Speculative — author’s original analysis] If that is right, then arguing about Bush’s legacy in the currency of appropriations alone is to argue about the wrong thing. The argument that honours him is about architecture: are the channels patent, are the niches diverse, is selection insulated, is the watershed flowing, can the system see itself, is the reservoir replenished?

To edit the endless frontier forward is not to update a 1945 document’s dollar figures for inflation. It is to recover the reading we lost — that Bush handed us not a frontier to be funded but an ecosystem to be tended — and then to carry that logic into a century in which the customer, the direction of flow, and the clock have all changed. In practice that means treating the procurement office as a load-bearing organ of the ecosystem, not an afterthought; measuring the system by how fast and reliably it fields what the market produces; and defending the invisible organs — talent, stability, diversity, feedback — with the vigilance now reserved for the topline. The frontier was always the rhetoric. The ecosystem was the point. Bush told us what it costs to let the joints go slack. The task now is to translate the warning into the grammar of dual-use competition, and to act on it before the decade-long delay does the arguing for us.

Epistemic Coda

The re-reading of Bush in Parts I–III is an argument about a text; a reader returning to Science, the Endless Frontier could reasonably weight its funding and its architecture differently than I have, and the “Probable” status of this essay is meant honestly. The structural and critical claims are now source-traceable: the reading of Bush rests on the report itself and on Atkinson, Leyden and Menter, Stokes, Gibson, Alberts and colleagues, and Stephan, all cited directly and rated [Confirmed] or [Probable]. (An earlier draft leaned on a 1999 attribution I could not verify against a primary text; I have replaced it with the documented 1999 interagency partnership review and the 2014 Alberts–Kirschner–Tilghman–Varmus diagnosis.) The three cases in Part V are from the open record: the F-35’s architectural troubles, the roles of DARPA, DIU, and In-Q-Tel, the Howell and colleagues causal finding as corroborated by the GAO, and NATO’s Rapid Adoption Action Plan are [Confirmed], while the Ukrainian market figures rest on single recent sources and are rated [Possible]. The quantitative spine in Part IV is now anchored to primary sources — the venture totals to PitchBook, the cycle-time figure to the GAO’s 2025 weapon-systems assessment — and rated [Confirmed]; the residual figures still drawn only from the compiled baseline (the RDT&E appropriation, the US-to-Europe venture ratio, the 178 per cent magnitude, and the European contractor counts) are held at [Probable] or [Possible] and labelled as such rather than dressed as settled. The framework in Part VI and its forecasts are [Speculative — the author’s original analysis]; they are genuine predictions, dated to mid-2026, and the next decade will adjudicate them. I would rather be precisely wrong on a flagged forecast than vaguely right on an unfalsifiable one.

Bibliography

Alberts, Bruce, Marc W. Kirschner, Shirley Tilghman, and Harold Varmus, ‘Rescuing US Biomedical Research from Its Systemic Flaws’, Proceedings of the National Academy of Sciences, 111.16 (2014), 5773–5777 https://doi.org/10.1073/pnas.1404402111.

Atkinson, Richard C., ‘Universities: At the Center of U.S. Research’, Science, 276.5318 (1997), 1479.

Bush, Vannevar, Science, the Endless Frontier: A Report to the President on a Program for Postwar Scientific Research(Washington, DC: United States Government Printing Office, 1945).

Бутенко, Олена, Надія Проскурніна, С. О. Козуб, and Рустам Могилевський, ‘Впровадження інноваційних бізнес-моделей у технологічних стартапах України’ [Implementation of Innovative Business Models in Ukrainian Technology Startups], Інвестиції: Практика та Досвід (2026) https://www.semanticscholar.org/paper/c06f68169cbf4b5847a1c5663ac876a634fff44b.

Federation of American Scientists, ‘CHIPS and Science Highlights: Developing a Scientific Workforce of the Future’ (2022; rev. 2026) https://fas.org/publication/chips-workforce-of-the-future/.

Gibson, Emily K., ‘Science Has Always Been Political’, Science, 369.6501 (17 July 2020), 233 https://doi.org/10.1126/science.abd7628.

Government Accountability Office, Small Business Research Programs: Air Force Had Success in Some Areas with New Awards Process, GAO-22-105223 (Washington, DC: GAO, 2022) https://www.gao.gov/products/gao-22-105223.

Government Accountability Office, Weapon Systems Annual Assessment: DOD Leaders Should Ensure That Newer Programs Are Structured for Speed and Innovation, GAO-25-107569 (Washington, DC: GAO, 2025) https://www.gao.gov/products/gao-25-107569.

Howell, Sabrina T., Jason Rathje, John Van Reenen, and Jun Wong, ‘Opening Up Military Innovation: Causal Effects of “Bottom-Up” Reforms to U.S. Defense Research’, NBER Working Paper No. 28700 (Cambridge, MA: National Bureau of Economic Research, 2021) https://www.nber.org/papers/w28700.

Hughes, Michael, ‘Lockheed Martin and the Controversial F-35’, Journal of Business Case Studies, 11.1 (2014) https://doi.org/10.19030/jbcs.v11i1.9047.

Information Technology and Innovation Foundation, ‘US Science Policy at a Crossroads’ (2025) https://itif.org/publications/2025/06/23/us-science-policy-at-a-crossroads/.

Korendovych, Victor, and Serhii Yasenko, ‘An Approach to Improving the Digital Transformation Management System (DTMS)’, De Securitate et Defensione. O Bezpieczeństwie i Obronności (2026) https://www.semanticscholar.org/paper/f58999dee796a27722bbc3dcc52ed2322468e6ab.

Lee, Marcel van der, Clara Peters, Marcel van Berlo, Luis Unzueta, David Ríos, Sirra Toivonen, and others, ‘A Holistic Framework for Assessing the Uptake Potential of EU-Funded Security Research and Innovation Project Results’, Open Research Europe (2025) https://doi.org/10.12688/openreseurope.19711.1.

Leyden, Dennis P., and Matthias Menter, ‘The Legacy and Promise of Vannevar Bush: Rethinking the Model of Innovation and the Role of Public Policy’, Economics of Innovation and New Technology, 27.3 (2018), 225–242 https://doi.org/10.1080/10438599.2017.1329189.

McSorley, Tom, Marta Macenowicz, Michael Maddison, and Christopher R. Yukins, ‘Allies Bridging the Valley of Death: How NATO’s Defence Innovation Accelerator for the North Atlantic Will Help Maintain NATO’s Technological Edge’, SSRN Electronic Journal (2025) https://doi.org/10.2139/ssrn.5092438.

Mulyanyuma, A., Justus Twesigye, and Agaba Halidu, ‘Enhancing Air Superiority Through Science and Technology’, Mediterranean Journal of Social Sciences (2026) https://www.semanticscholar.org/paper/56f2d5d4a6106829fefad5e969d2baef8f01efc4.

National Science and Technology Council, Renewing the Federal Government–University Research Partnership for the 21st Century (Washington, DC: Executive Office of the President, 1999).

North Atlantic Treaty Organization, ‘Summary of NATO’s Rapid Adoption Action Plan’ (Brussels: NATO, 25 June 2025) https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/2025/06/25/summary-of-natos-rapid-adoption-action-plan.

Orbis Futura Review, ‘Empirical Baseline for Essay #1: Defense-Technology Transition Metrics’ (internal research note, 2026). Figures compiled from multiple industry and government sources; aggregate venture totals, MDAP cycle-time data, RDT&E appropriations, the venture-uplift estimate, and European contractor-roster counts are reproduced as supplied and rated [Probable] or [Possible] in text where no primary source could be independently traced.

Patil, Kapil, and Saradindu Bhaduri, ‘”Zero-Error” Versus “Good-Enough”: Towards a “Frugality” Narrative for Defence Procurement Policy’, Mind & Society, 19 (2020) https://doi.org/10.1007/s11299-020-00223-7.

PitchBook, 2023 Vertical Snapshot: Defense Tech (Seattle, WA: PitchBook Data, 2023) https://pitchbook.com.

Stephan, Paula, How Economics Shapes Science (Cambridge, MA: Harvard University Press, 2012).

Stepanov, A., ‘Military-Civil Collaboration in the USA in the Sphere of Advanced Technologies’, World Economy and International Relations (2025) https://www.semanticscholar.org/paper/5f62a5c764eecda222b474b3e7a325e86f869e56.

Stokes, Donald E., Pasteur’s Quadrant: Basic Science and Technological Innovation (Washington, DC: Brookings Institution Press, 1997).

[^bush1945]: Vannevar Bush, Science, the Endless Frontier: A Report to the President on a Program for Postwar Scientific Research (Washington, DC: United States Government Printing Office, 1945). The report was commissioned by Franklin Roosevelt in November 1944 and delivered to Harry Truman in July 1945; its four guiding questions and corresponding committee reports (on the diffusion of wartime knowledge, medical research, the support of research, and the development of scientific talent) form the structural argument discussed here.

[^fas2022]: Federation of American Scientists, ‘CHIPS and Science Highlights: Developing a Scientific Workforce of the Future’ (2022; rev. 2026) https://fas.org/publication/chips-workforce-of-the-future/. On the lineage of the Endless Frontier Act into the CHIPS and Science Act, and on the public conversation’s fixation on authorised funding rather than Bush’s emphasis on talent.

[^atkinson1997]: Richard C. Atkinson, ‘Universities: At the Center of U.S. Research’, Science, 276.5318 (1997), 1479. See also Atkinson’s later restatement of Bush’s model as a national enterprise in which federally funded, university-conducted basic research is implemented by private industry.

[^gibson2020]: Emily K. Gibson, ‘Science Has Always Been Political’, Science, 369.6501 (17 July 2020), 233 https://doi.org/10.1126/science.abd7628. Marking the report’s 75th anniversary, Gibson calls it “the Magna Carta of American science” while arguing that the autonomous, apolitical ideal it enshrined was always partly mythical.

[^leyden2018]: Dennis P. Leyden and Matthias Menter, ‘The Legacy and Promise of Vannevar Bush: Rethinking the Model of Innovation and the Role of Public Policy’, Economics of Innovation and New Technology, 27.3 (2018), 225–242.

[^stokes1997]: Donald E. Stokes, Pasteur’s Quadrant: Basic Science and Technological Innovation (Washington, DC: Brookings Institution Press, 1997).

[^itif2025]: Information Technology and Innovation Foundation, ‘US Science Policy at a Crossroads’ (2025) https://itif.org/publications/2025/06/23/us-science-policy-at-a-crossroads/. On the Bush–Kilgore dispute, the 1947 veto, and the establishment of the NSF in 1950.

[^nstc1999]: National Science and Technology Council, Renewing the Federal Government–University Research Partnership for the 21st Century (Washington, DC: Executive Office of the President, 1999). The review was convened in part because the postwar compact’s guiding principles had never been codified and were under strain; a parallel statement of principles followed in Executive Order 13185 (2000).

[^alberts2014]: Bruce Alberts, Marc W. Kirschner, Shirley Tilghman and Harold Varmus, ‘Rescuing US Biomedical Research from Its Systemic Flaws’, Proceedings of the National Academy of Sciences, 111.16 (2014), 5773–5777. The authors — including two former directors of the NIH and a former Princeton president — argue that the research system has been driven into hypercompetition, perverse incentives, soft-money dependence, and a structural oversupply of trainees.

[^stephan2012]: Paula Stephan, How Economics Shapes Science (Cambridge, MA: Harvard University Press, 2012). On risk aversion, the oversupply of PhDs, dependence on soft money and the rising age of first independent funding, biomedical concentration, and the destabilising effects of stop-go funding.

[^cen2025]: ‘”A Huge Rupture in Everything”: US Science Faced Major Upheaval in 2025’, Chemical & Engineering News (December 2025); and ‘See the Alarming Extent of NIH and NSF Funding Cuts in 2025’, Science News (January 2026). On grant terminations (~US$3 billion), indirect-cost reductions toward 15 per cent, agency staff reductions, and proposed FY2026 cuts.

[^cen2027]: ‘Trump’s 2027 Budget Proposes Deep Cuts to Science Programs’, Chemical & Engineering News (April 2026). On the proposed reduction of the NSF budget to roughly US$4 billion (a ~54.5 per cent cut from FY2026), the ~10.5 per cent cut to the NIH, the proposed elimination of institutes, and the ~37 per cent reduction to ARPA-H.

[^penn2026]: ‘White House 2027 Budget: Funding Cuts to NIH and NSF’, The Daily Pennsylvanian (April 2026). On faculty responses and on EU and Chinese increases in fundamental-research investment relative to US contraction.

[^hughes2014]: Michael Hughes, ‘Lockheed Martin and the Controversial F-35’, Journal of Business Case Studies, 11.1 (2014) https://doi.org/10.19030/jbcs.v11i1.9047.

[^baseline]: Orbis Futura Review, ‘Empirical Baseline for Essay #1: Defense-Technology Transition Metrics’ (internal research note, 2026). Source of the aggregate venture-capital figures (≈US$135.3 billion across ≈4,744 deals, 2016–2022; ≈US$184.7 billion projected by 2027; US-over-Europe volume ratio of ≈2.4), the FY2026 RDT&E appropriation (≈US$146 billion, ≈15 per cent of the defence budget), the MDAP cycle-time figures (124 → 142 months, 2023–2024), the 178 per cent venture-uplift estimate associated with the Howell and colleagues finding, the European primary-contractor-roster counts (UK, Germany, Poland), and the JCIDS-disestablishment note (late 2025). Reproduced as supplied; figures without an independently traceable primary source are rated [Probable] or [Possible] in text.

[^howell2021]: Sabrina T. Howell, Jason Rathje, John Van Reenen and Jun Wong, ‘Opening Up Military Innovation: Causal Effects of “Bottom-Up” Reforms to U.S. Defense Research’, NBER Working Paper No. 28700 (Cambridge, MA: National Bureau of Economic Research, 2021). Using a regression-discontinuity design, the authors find that winning an “Open” (bottom-up) award causally raises subsequent venture-capital investment, non-SBIR defence contracting, and patenting, whereas “Conventional” awards do not and instead foster incumbency. The precise 178 per cent magnitude is reported via the empirical baseline and rated [Probable].

[^gao2022]: U.S. Government Accountability Office, Small Business Research Programs: Air Force Had Success in Some Areas with New Awards Process, GAO-22-105223 (2022). Roughly 43 per cent of Open-topic awardees had no prior federal contracts (versus 14 per cent for conventional awards), awards were issued faster, and Open-topic awardees were more likely to attract subsequent venture capital than near-miss non-awardees (about 2 per cent of whom secured VC).

[^gao2025]: U.S. Government Accountability Office, Weapon Systems Annual Assessment: DOD Leaders Should Ensure That Newer Programs Are Structured for Speed and Innovation, GAO-25-107569 (2025). The expected time for a Major Defense Acquisition Program to deliver an initial capability rose by 18 months in a single year, to almost 12 years from programme start; the report also notes the F-35 took roughly 18 years from initial solicitation to operational capability, and that DOD plans to invest nearly US$2.4 trillion in its costliest programmes.

[^pitchbook2023]: PitchBook, 2023 Vertical Snapshot: Defense Tech (2023). Reports US$135.3 billion invested across 4,744 defence-technology deals (2016–2022), with the market projected to reach US$184.7 billion by 2027 at a compound annual growth rate of about 15.9 per cent, and identifies the Defense Innovation Unit and the Office of Strategic Capital as transition-bridging mechanisms.

[^mcsorley2025]: Tom McSorley, Marta Macenowicz, Michael Maddison and Christopher R. Yukins, ‘Allies Bridging the Valley of Death: How NATO’s Defence Innovation Accelerator for the North Atlantic Will Help Maintain NATO’s Technological Edge’, SSRN Electronic Journal (2025) https://doi.org/10.2139/ssrn.5092438.

[^nato2025]: North Atlantic Treaty Organization, ‘Summary of NATO’s Rapid Adoption Action Plan’ (25 June 2025), endorsed by Heads of State and Government at the Hague Summit. On the 24-month adoption-and-fielding target, the explicit embrace of acquisition risk and a “fail fast” posture, and the roles of DIANA and the NATO Innovation Fund (a deep-tech venture vehicle backed by 24 of NATO’s 32 nations). The €7.3 billion European Defence Fund earmark (2021–2027) is reported via the empirical baseline.

[^patil2020]: Kapil Patil and Saradindu Bhaduri, ‘”Zero-Error” Versus “Good-Enough”: Towards a “Frugality” Narrative for Defence Procurement Policy’, Mind & Society, 19 (2020) https://doi.org/10.1007/s11299-020-00223-7.

[^mulyanyuma2026]: A. Mulyanyuma, Justus Twesigye and Agaba Halidu, ‘Enhancing Air Superiority Through Science and Technology’, Mediterranean Journal of Social Sciences (2026) https://www.semanticscholar.org/paper/56f2d5d4a6106829fefad5e969d2baef8f01efc4. Recent single source; supporting interpretive claims rated [Possible].

[^korendovych2026]: Victor Korendovych and Serhii Yasenko, ‘An Approach to Improving the Digital Transformation Management System (DTMS)’, De Securitate et Defensione (2026) https://www.semanticscholar.org/paper/f58999dee796a27722bbc3dcc52ed2322468e6ab.

[^lee2025]: Marcel van der Lee and others, ‘A Holistic Framework for Assessing the Uptake Potential of EU-Funded Security Research and Innovation Project Results’, Open Research Europe (2025) https://doi.org/10.12688/openreseurope.19711.1.

[^butenko2026]: Olena Butenko and others, ‘Implementation of Innovative Business Models in Ukrainian Technology Startups’, Investytsiyi: Praktyka ta Dosvid (2026) https://www.semanticscholar.org/paper/c06f68169cbf4b5847a1c5663ac876a634fff44b. Market-share and valuation figures rest on this single recent source and are rated [Possible].

[^stepanov2025]: A. Stepanov, ‘Military-Civil Collaboration in the USA in the Sphere of Advanced Technologies’, World Economy and International Relations (2025) https://www.semanticscholar.org/paper/5f62a5c764eecda222b474b3e7a325e86f869e56.