Deterrence Depends on Turning Commercial Technology Into Fielded Capability

Dan Berkenstock came to the defense-industrial question through space engineering and commercial satellite development, rather than through a career in strategy. That background informs his central concern: the United States may be closer to a great-power conflict than it has been for one or two generations, and the relevant preparation is deterrence—not a plan to welcome or merely endure such a conflict.

Berkenstock co-led Hoover’s “Providing for the Common Defense” team, which used artificial intelligence to interrogate a corpus of roughly 60 million to 65 million words: prior commission reports, think-tank work, congressional material, op-eds, interviews, Hoover research, and new interviews. He describes the exercise as an attempt to ask a broad national-security “brain trust” the same questions at once: where consensus exists, what the principal threats are, and what kind of conflict the United States should expect.

The result, he says, was striking agreement around three propositions. First, the risk that the United States will become involved in a hot war with another great power is higher than it has been in recent decades. Second, such a conflict would not resemble the rapid, initially overwhelming campaigns associated with recent American operations. It would more likely be prolonged and attritional, with difficult political objectives and an outcome shaped by the ability to produce munitions and capabilities at scale—and then improve them rapidly once fighting begins. Third, the American homeland would not remain a sanctuary. The economy, everyday life, and communications could all be directly affected; Berkenstock calls the likely consequences potentially the most devastating experience for the United States since the Civil War.

That is why the desired outcome is not simply a stronger force in the abstract. It is for China’s Xi Jinping and Russia’s Vladimir Putin, whom Berkenstock expects to remain influential for some time, to judge every day that conflict is not worth initiating.

The working theory of victory is deterrence adapted to a 21st-century competition. It requires deployable combat power, but also resilience at home: physical infrastructure, energy systems, cyber infrastructure, and the other foundations needed to withstand coercion or disruption if deterrence fails. The premise is not that a major war can be made painless. It is that its costs would be so severe, even in a nominal victory, that avoiding it must be the organizing objective.

In Mobilize the Modern Defense Industrial Base, which Berkenstock co-authored with former Space Force chief John Raymond, he identifies America’s commercial capital markets—not principally its defense budget—as the country’s greatest strategic advantage in military competition.

The argument begins with the distinction between the “high” and “low” ends of a future force mix. Over the past four or five decades, the United States has become highly capable at the high end: exquisite systems such as stealth platforms, precision munitions, and Patriot interceptors. They are capable, highly reliable, and expensive. They are also difficult to industrialize and scale quickly, and they are produced by a concentrated sector. Berkenstock estimates that the five major defense primes account for roughly 1% of total U.S. public-market capitalization.

A much larger commercial economy has separately become good at the low end. It does not usually call its output defense materiel. It calls it smartphones, computers, digital cameras, consumer drones, sensors, processors, timing chips, actuators, and the other piece parts of modern electronics. Since about 1980, Berkenstock says, commercial R&D has exceeded government R&D on an annual basis. The accumulated output is a vast catalog of components from which new systems can be assembled.

Venture-backed firms have developed a corresponding process: make an initial investment, test a configuration, iterate rapidly, and push a successful application toward low-cost production. In Berkenstock’s account, the defense task is not to abandon high-end systems for cheap ones. It is to pair the concentrated industrial base that makes the “high” with a much wider commercial capability for developing and fielding the “low.”

Berkenstock says this $80 billion is ten times the amount invested during the whole of the 2010s, and one hundred times the amount invested from 2000 through 2010. Veterans returning from Iraq and Afghanistan helped drive the shift, he says, frustrated by the disparity between the equipment available in the field and the technology they used at home. China’s rise as a strategic competitor, along with conflicts in Ukraine, Israel, and Iran, has further focused attention on the military relevance of newer technologies.

Private firms can function as a fast-moving research-and-development arm. In the AI era, Berkenstock says, four Stanford graduate students with $1 million in pre-seed financing can attempt work of a complexity once confined to small, secretive government laboratories or institutions such as Lockheed Martin’s Skunk Works. Where there was one Skunk Works, there are now hundreds or thousands of technically capable teams with access to knowledge, tools, and capital.

He calls this an “app era” for defense technology: a shift from the equivalent of mainframes toward PCs and handhelds. Far more ideas can be tested. Yet the availability of experiments is not itself a defense capability. The harder problem is identifying successful ideas, linking them to military buyers and budgets, financing their production, and sustaining them through training, maintenance, and deployment.

That is where commercial capital alone becomes insufficient. Venture financing is unusually good at quickly backing a credible technical thesis. Berkenstock has seen companies receive a $20 million to $50 million term sheet within a week and have money in the bank a few weeks later. He does not think the government should try to recreate that process.

But investors become less able to finance a company when it must commit substantial capital before it has customer backlog or signed contracts. The company may need to build a factory, establish a supply chain, complete testing, or create the operational infrastructure around its product before it can prove to investors that customers will arrive at scale. Space makes the problem particularly visible: getting a system into orbit can cost $20 million, $30 million, or $50 million before the company knows whether the system has meaningful customer value.

The issue is not whether the government should replace venture capital. It is whether it can make its demand credible early enough to allow private capital to finance the transition from prototype to production.

Dan Berkenstock calls the gap between a validated idea and a fielded, supportable capability the “valley of death.” A company may receive enough research-and-development funding to establish that an idea works, but lack the capital to build a factory, complete testing, establish production, and deliver something ready for a warfighter.

The structure of defense appropriations has helped create that gap. There is a longstanding divide between RDT&E—research, development, test, and evaluation funding—and procurement and sustainment funding. Procurement dollars have conventionally been meant for finished products: systems built, tested, validated, and ready to leave the factory. But young companies often need procurement-scale financing before they can afford the production capacity required to become finished-product suppliers.

For Berkenstock, the government’s role is therefore not to select every technical bet or directly manage startups. It is to provide a legible path from an early technical success to an actual order. A limited amount of government revenue can have disproportionate leverage, he argues, because it gives private investors evidence that a product will have a buyer.

The Defense Innovation Unit’s use of Other Transaction AgreementsA flexible federal contracting arrangement that can be used outside many standard procurement rules., or OTAs, is one such tool. Berkenstock says OTAs, previously used by NASA but not widely understood within the Department of Defense, have helped companies move from development work toward procurement. Their importance is not merely procedural flexibility. They can create a transition path for a company that has outgrown a small R&D award but does not yet look like an established prime contractor.

The Small Business Innovation Research program, or SBIR, addresses a related stage. Berkenstock says its recent reauthorization significantly expanded incentives for private investors and government program managers to bring matching dollars to larger Phase III awards. The relevant company has tested a technical premise, but is not yet able to manufacture or deploy at scale. A larger award supplies capital and signals a customer’s interest.

Title III of the Defense Production Act is still more directly oriented to production. Berkenstock describes it as one of the few mechanisms through which the department can make a flexible, multi-year “if you build it, we will buy it” commitment outside ordinary congressional authorities for multi-year contracts or appropriations. A company might receive a $50 million award for a product expected to be ready in 12 months, 24 months, or three years. The award can then help it raise the additional capital required to deliver.

The size of that signal matters. Berkenstock says Title III caps have not been adjusted for inflation since the early 1990s. If Congress adjusted them, a potential award could rise from roughly $50 million to roughly $125 million. For a company considering a factory or production line before substantial customer orders are in hand, he argues, that difference could change the investment case.

The reforms he describes are technical rather than dramatic: flexible contract vehicles, matching dollars, transition pathways, and bounded purchase commitments. Their common purpose is to turn a military requirement into demand that commercial markets can understand and finance.

China’s military-civil fusion strategy links state priorities to civilian industry, research, and economic planning. The United States cannot simply reproduce that system, Dan Berkenstock says. China has a legal mechanism for ensuring that private enterprise does what the government wants. The United States has influence, particularly economic influence, but it has to pitch: persuade firms, make contracts, and build incentives that bring private actors along.

That difference makes mutual literacy a practical requirement. For much of the 2000s and 2010s, Berkenstock says, one group spoke Silicon Valley venture capital while another spoke government acquisition. Entrepreneurs and investors need to understand the opportunities, motivations, constraints, and limitations facing even forward-looking program managers and congressional appropriators. Military leaders, meanwhile, need to understand the lived reality of startups: raising the next round, increasing valuation, building a profitable going concern, and eventually returning enough capital to keep investors interested in financing the next cohort.

The historical separation was not inevitable. Silicon Valley began as a defense-research hub. Early U.S. spy satellites were built nearby, and companies such as Fairchild and National Semiconductor had early growth driven in substantial part by defense applications. Berkenstock identifies several sources of the later split: the legacy of the Vietnam War in universities and the broader scientific and engineering establishment; the post-Cold War peace dividend; consolidation from roughly 80 to 100 prime contractors to five or six; weak public-market comparables for defense primes; and a contracting system that was difficult to enter without what he calls an army of lobbyists.

Organizations including In-Q-Tel and the Defense Innovation Unit have spent years trying to reverse that separation. The resulting community of founders and investors is larger and more motivated than it was, but the investment cycle is still unproven. Berkenstock describes the sector as being in its second or third inning. The capital already invested must generate exits and returns strong enough to establish a durable commercial model. Investors can develop conviction about dual-use and national-security technology, but they also follow evidence that a sector can produce returns.

Public and private companies serve different purposes in that model. Public companies provide scale, substantial pools of capital, and longer planning horizons. Berkenstock estimates that U.S. public markets offer roughly 50 times the capital available to venture capital: public companies collectively have market capitalization in the mid-$50 trillion range, compared with a little over $1 trillion in venture-capital assets under management. Private companies provide rapid experimentation and early development. The task is to connect those functions, rather than treating one as a substitute for the other.

Berkenstock’s counterpart to military-civil fusion is therefore not a command structure. It is a durable alignment of defense requirements, scientific capacity, private incentives, and public demand over a generational competition. That alignment serves deterrence by expanding combat power, but it also supports resilience at home—and, in his view, can create jobs, careers, industries, and investment opportunities during a period of broad anxiety about globalization, offshoring, and AI-driven changes to white-collar work.

Dan Berkenstock sees space as both a domain of great-power competition and a particularly clear case of the industrial-finance problem. A spacecraft is not a useful defense capability merely because a startup can design it from commercial components. It must be manufactured, launched, operated, maintained, connected to customers, and supported through an architecture capable of providing service at scale. Each step raises the capital required before the company can necessarily demonstrate durable demand.

For most of the past 50 years, he says, the United States lacked a true strategic competitor in space. Its program often had ambitious visions but limited follow-through, constrained by appropriations that did not match the scale of its stated objectives. That began to change in the late 2000s and early 2010s, as reports indicated that space might no longer be the relatively safe haven it had been when norms and treaty obligations constrained weaponization. Russia and China increasingly appeared to view space as part of a potential theory of victory against the United States.

Space has also become more congested and contested as commercial firms place growing numbers of systems in orbit. Berkenstock, who cautions that he is not a China expert, says China’s effort to build orbital capability at scale can serve several purposes at once: demonstrating technical parity or excellence, providing infrastructure and services to other countries, and shaping operating norms in a domain where formal governance remains comparatively light.

He characterizes the prospective infrastructure role as “Huawei in space.” A country that can get to orbit and place systems there at scale can provide global services, including to developing countries, in ways that complement initiatives such as China’s Belt and Road effort. Scale develops industrial competence, expands geopolitical influence, and may affect how a more crowded orbital environment is governed.

The Moon is a longer-horizon expression of the same competition. Berkenstock recalls the cohesion of the Apollo era: people could disagree about the path to the goal without disputing the goal itself. He sees a comparable need for strategic clarity now, both in the broader competition with China and in determining how far and how permanently the United States intends to extend its presence in orbit and beyond it.

A lunar presence would create incentives to remain and expand rather than settle a one-time symbolic race, he argues. The relevant interests could be exploratory, scientific, economic, and potentially military. He notes, with explicit caution, a claim by the one geologist who walked on the Moon that lunar regolith might contain enough helium-3 to provide clean nuclear power on Earth for 100 to 1,000 years. Realizing even a small part of that possibility would require substantial time, effort, and technological development. The point for Berkenstock is not that the resource is readily available, but that the next space race may generate enduring competition over access, infrastructure, and economic activity rather than end at a landing.