When the world's at stake,
go beyond the headlines.

“Loose lips sink ships.” It was catchy, memorable, and above all, true. That wisdom feels incompatible with Silicon Valley’s culture and business model, yet today’s national security strategy depends on it. By oversharing, emerging defense companies may be undermining the very deterrence to which they hope to be contributing.

For much of the past 30 years, U.S. deterrence was straightforward: reveal to deter and conceal to win. Put simply, the United States had the biggest stick, and just brandishing it was enough. The world is no longer that clear. A peer adversary now matches or surpasses U.S. capacity in key areas. Deterrence in this environment is a choreographed dance, revealing some capabilities, concealing others, and hinting at yet more in the shadows. Yesterday, deterrence was about certainty that America would win. Today, it’s increasingly about creating uncertainty in peer adversaries.

Unfortunately, the rise of newer players in the defense industry has collided with this reality. Unlike Lockheed or Boeing, these firms are still building credibility and financial stability. That drives them to publicize not just current capabilities, but also dreamed-of ones, often in glossy decks or highly produced videos. This grants adversaries not only a stronger understanding of current U.S. capabilities but a crystal ball to both future capabilities and research and development roadmaps. While many outside observers may criticize these startups for putting marketing ahead of national security, the reality is that their bombast is about survival. As an investor in early-stage defense technology companies, I recognize that new defense startups resort to splashy marketing because they need visibility with government buyers, capital from investors, and positioning against competitors. Without marketing their capabilities, they may not survive long enough to deliver for the American people. However, their survival concerns must be weighed against the fact that every announcement erodes the fog of uncertainty critical to modern deterrence as adversaries can model, wargame, plan for, and hence mitigate future U.S. capabilities with enough accuracy to act. Resolving this tension requires a system that allows for enhanced information sharing amongst the right cohort of people and across silos while limiting public announcements. In short, the United States should dramatically expand the number and types of people with security clearances and couple that with proactively classifying cutting-edge technology.

How and Why Startups Overshare

In today’s defense-tech ecosystem, oversharing doesn’t mean leaking secrets to foreign spies —­ it means giving them away for free through marketing collateral. Spend five minutes on YouTube or LinkedIn and you’ll see glossy videos of weaponized drones, mass-produced missiles, and founders giving podcast interviews that double as detailed product roadmaps. Companies proudly display technical data, prototypes, and even core intellectual property. It’s not malicious — it’s standard operating procedure from the tech world, where transparency and hype are how you win.

Startups overshare because they have to. The defense market is opaque, slow-moving, and difficult to navigate. For a new company, being invisible is fatal. Public visibility helps them signal relevance to government buyers who might not even know their category exists. It helps them attract industry partners, since most defense contracts require teams of companies integrating different technologies into one solution, and, most pressingly, it helps them raise money.

Finding the right government buyer isn’t a trivial task. My team and I have found that many buyers and requirements writers don’t know what technology is being developed or is ready for scaling today. Similarly, most founders have no idea how to identify the right government customer — it isn’t as simple as finding a program office and knocking on the front door. Given these challenges, companies fall back on what works in the private sector: They engage in marketing. If a startup can’t pinpoint the right buyer, they know at least some of them will pass through the Pentagon Metro stop. Unfortunately, so do foreign agents.

While finding the right government stakeholders is critical, finding the right industry partners can also be vital. The Department of Defense generally prefers to buy integrated solutions, not standalone technology. That means an alternative position, navigation, and timing startup may need to join forces with a rocket motor maker and a composites firm to maximize their chance of succeeding in the defense market. Given how different each of these companies are, they are unlikely to see one another outside the confines of broad marketing efforts. Public marketing becomes the matchmaking service, even if it reveals too much along the way.

Finally, startups must get in front of the right industry partners and the right government stakeholders to secure government contracts. But before they can do either, and long before they generate material revenue from the Department of Defense, they must engage in product development, which means they need to attract investors. Once capital is raised, the pressure to keep valuations rising makes the cycle worse: oversell today, exaggerate tomorrow, and inadvertently help adversaries model capabilities.

Each of the above problems is an example of a coordination challenge brought on by information silos. In tech, visibility is oxygen. Unfortunately, this same visibility is often directly at odds with a national security mission. Publicly revealing roadmaps, performance data, or integration partners doesn’t just impress investors, it teaches adversaries how to model U.S. capability trajectories. Every slick video, every public test, every overshared deck gives near-peer rivals like China a free intelligence feed on current capabilities and research and development roadmaps that once required espionage.

Why Certainty Is Dangerous

The classic deterrence model is simple: if the United States has overwhelming force and the will to use it, adversaries won’t act. However, against a motivated peer like China, that’s too neat. A powerful rival able to model conflict outcomes with precision is more likely to call America’s bluff. In this case, deterrence depends on two levers. First, increasing capability relative to the adversary, so their models demonstrate that either they will fail to achieve their objectives or that achieving their objectives will cost more than they are willing to pay. In classic deterrence speak, this lever represents deterrence by denial and deterrence by punishment. The second lever is increasing the adversary’s uncertainty about the likely outcome so that even if their average outcome is favorable, the range of outcomes is too broad to be worth the risks. This can be thought of as deterrence by doubt.

During the Cold War, defense firms understood this instinctively. When the F-18 entered public view in 1983, the F-117 was already operational and the F-22 was in development. Adversaries saw one system deployed, suspected others were waiting in the wings, and had no idea what was coming next. Overwhelming force plus unpredictability was a stronger deterrent than brute force alone.

Contrast that with what happens when a capability becomes known to an adversary. The U-2 and the SR-71 both performed admirably as spy planes over the Soviet Union. Once their capabilities became known, the Soviet Union was able to begin missile and interceptor programs designed to counter their capabilities. Similarly, as Eye in the Sky: The Story of the Corona Spy Satellites details, the Soviet Union was able to use to information disclosed in contractors’ promotional materials to estimate the capabilities of U.S. spy satellites, which allowed the Soviets to begin hiding or obfuscating their own capabilities. Once an adversary knows what its opponents’ capabilities are, they can model them, predict them, and counter them.

A clever rival doesn’t need spies to piece together America’s playbook — YouTube and slick sheets often suffice. I worry the People’s Liberation Army can now model U.S. capabilities five years out better than it could a decade ago. While some may argue that intelligence on what the U.S. startup ecosystem is developing is of limited value, particularly given how many startups ultimately fail, I believe that even when startups end as failed prototypes or “vaporware,” there can be substantial harm in revealed intent, capability gaps, and fundamental directions. Adversaries don’t need these systems to work to benefit from the information. Product roadmaps reveal which operational problems U.S. firms are trying to solve, what performance envelopes they think are achievable, the research institutions out of which the fundamental work is being performed, the founders and engineers who bring the specific expertise to bear, where industrial capacity is being rebuilt, and untold other signals that smart and nefarious analysts can turn to their own purposes. These signals allow a peer competitor to model likely future capabilities, develop countermeasures early, and time aggression to precede U.S. capability realization. In other words, even vaporware is intelligence. It tells an adversary where the United States plans to be strong tomorrow and where it is weak today. More certainty means more confidence to act and a higher chance of success if they do. This dynamic creates a dilemma for the United States. It must figure out how to break down information silos internally while preserving substantial uncertainty in its adversaries.

A Counterintuitive Fix

My proposal: classify more and earlier, but aggressively expand who’s cleared to see it.

Simply clamping down won’t work. Neo-primes still need customers, teammates, and capital. Therefore, aggressive classification must be paired with an equally aggressive expansion of the clearance system. That means dramatically increasing the number of clearances in industry and streamlining the process for facility clearances.

Clearances for industry must expand well beyond traditional contractors to include founders, investors, bankers, and even state and local officials. Reindustrialization and renewed deterrence require buy-in from across society. As of the last report from the Office of the Director of National Intelligence, 4.1 million Americans hold a security clearance, which is about 1.2 percent of the population. However, if you exclude active duty military, reservists, National Guard, and Department of Defense civilians, well less than 0.5 percent of Americans hold a clearance of any type. To put that into perspective, over the life of the Manhattan Project, as many as 600,000 Americans contributed, which would have been almost 0.5 percent of the population on that project alone. A true whole-of-nation effort of the sort that is being contemplated demands many times that number.

Some will object that dramatically increasing the number of cleared Americans increases the chance that there will be leaks. While almost certainly true, compare it to the status quo, where sensitive commercial information is being trumpeted from the mountaintops. Just as research and development investments have migrated increasingly from the government to industry, so too has strategically valuable information.

If the United States must build a wall around its garden, and I believe it must, then the United States must also ensure the garden is stocked with the seeds of deterrence.

Inside the Garden

As more players are brought “inside,” familiar tools become far more powerful — it’s the classic network effect at work. Classified industry days, broad consortia, capability gap briefings: These forums for educating buyers, sourcing capital, and building teams all gain new life once more people are cleared to participate.

Companies will always want to market themselves. But if enough outsiders become insiders, the incentive for splashy public campaigns diminishes. That restores America’s ability to make deliberate reveal/conceal decisions, rather than letting market pressures dictate them. An additional benefit and one that should not be undersold, bringing more Americans inside the tent activates them for the national security mission, which in and of itself may serve as a further deterrent.

The Path Forward

Confronting modern national security challenges requires an all-of-nation approach, but coordination is impossible without secure spaces for trusted, vetted partners to interact. That means balancing risks and information sharing in new ways.

Deterrence today isn’t about overwhelming force alone. It’s about making sure adversaries never know quite enough to gamble. To get there, the United States needs to reverse Silicon Valley’s instinct to overshare, while preserving its ability to build the next generation of national security capabilities. A successful innovation system that maintains strategic and tactical surprise is one where those who must know, do.

Loose lips may no longer sink ships. But they could sink deterrence.

Jake Chapman is the managing director of Marque Ventures, an early-stage venture capital firm supporting companies advancing U.S. national security and the values of the Western world. Prior to founding Marque, Jake led efforts to reimagine how the Department of Defense engages in venture capital activities through the Army Venture Capital Corporation — laying the groundwork for Marque’s approach today. He is also an adjunct at RAND, where he works at the intersection of technology, policy, and national security, and serves on the editorial board of The Ledger: A Journal of Economic Statecraft.

**Please note, as a matter of house style War on the Rocks will not use a different name for the U.S. Department of Defense until and unless the name is changed by statute by the U.S. Congress.

Image: Lance Cpl. Jackson Rush via DVIDS.