Defense Tech Needs the State, Not Less of It

If you’ve spent any time around the new “defense tech” crowd, you’ve heard the story.

It goes like this: In 1993, Pentagon leaders hosted a famous dinner with defense-industry executives—the “Last Supper.” The Cold War was over, budgets would shrink, and the industry would have to consolidate. After that moment, the tale continues, U.S. defense procurement became a slow, risk-averse, paperwork-choked machine. Innovation fled. Big primes hardened into immovable monopolies. If only the government would stop meddling—if only it would deregulate harder—Silicon Valley could save national security.

It’s a neat narrative. It’s also the wrong lesson.

The uncomfortable truth is that the U.S. defense industrial base didn’t become fragile because government intervened too much. It became fragile because government retreated—from production capacity, from oversight, from in-house expertise, and from the boring but essential work of managing a market that will never behave like a normal market.

And defense is not a normal market. It never has been. It never will be.

Contents

The “Last Supper” is a convenient myth

The dinner in 1993 makes a great origin story because it has all the ingredients: a dramatic name, a clean turning point, a villain (the bureaucracy), and a moral (deregulate, move fast, break things).

But by the time that dinner happened, the industry had already been changing for years—driven by forces that had little to do with one Pentagon event:

Globalization was pulling supply chains overseas.
Financialization was reshaping corporate behavior—debt loads, short-term shareholder logic, “financial engineering.”
Defense budgets had already started drawing down after the mid-1980s.
And Washington itself was being pushed to “run government like a business,” which often meant: shrink internal capacity, outsource, and assume markets will handle the rest.

So when the new defense-tech leaders point to 1993 as the moment everything broke, they’re treating a late-stage symptom as the cause.

The real erosion started earlier: hollowing out, not over-regulation

In the early Cold War era, the American model was closer to a partnership with leverage: the government invested heavily, maintained public arsenals and shipyards, and could steer industry because it retained real in-house capability. That mix produced not just weapons, but spillovers—technologies that reshaped civilian life.

Then the world changed. Foreign manufacturing recovered and surged, often backed by state support. American firms chased lower costs abroad. Meanwhile, Wall Street logic seeped deeper into industrial strategy.

The result wasn’t just “less competition.” It was structural thinning:

fewer suppliers in key subsectors,
extreme concentration in strategically vital niches,
and big diversified companies deciding defense wasn’t worth the hassle, walking away.

This is the part that rarely makes it into Silicon Valley’s version: the defense industrial base became brittle long before the famous dinner.

When the government outsourced, capability didn’t “move”—it disappeared

Here’s the hard thing about national security production: you can’t treat it like a logistics problem. If you “shift” capacity out of government without keeping the knowledge and infrastructure alive, you don’t get an efficient private substitute. You often get a capability gap.

Take shipbuilding as the archetype. As public capacity receded and private yards dominated, some private builders stopped bidding. Instead of rebuilding a public option, policy choices narrowed government’s role further. Over time, the ecosystem didn’t become more efficient—it became less resilient, more dependent on fewer actors, and harder to surge in a crisis.

The same pattern showed up in services and systems: outsourcing often came with contracts that were too narrow, poorly structured, and written under the illusion that you could specify reality in advance. When reality changed—as it always does—the government paid more and learned less. And once in-house talent was gone, “bringing it back” became incredibly expensive, slow, and politically difficult.

Outsourcing wasn’t just a cost strategy. It was state capacity liquidation.

Deregulation didn’t unleash innovation. It weakened the buyer.

In the 1990s, Washington tried something very close to what today’s defense-tech evangelists recommend: cut military-unique requirements, favor commercial products, streamline acquisition rules, treat emerging tech as something the private sector would naturally supply.

But here’s what happened in practice:

Contracts became more complicated.
Oversight became harder.
The acquisition workforce shrank dramatically.
The government increasingly lacked the internal technical depth to judge what it was buying, integrate it properly, or manage vendors effectively.

That last point is crucial: a “faster” procurement system is meaningless if the buyer becomes incompetent. When the state can’t evaluate claims, can’t negotiate well, and can’t enforce performance, the system doesn’t become innovative—it becomes extractive.

And in defense, the downside of extraction isn’t just wasted money. It’s strategic vulnerability.

Silicon Valley’s “First Breakfast” risks repeating the same failure

The new defense-tech argument is seductive: reduce reporting requirements, loosen compliance, rely more on flexible contracts, shift more risk to contractors with fixed-price deals, and let a new generation of firms outcompete the old primes.

Some of that contains real frustration with real problems. Procurement is often slow. Requirements can be bloated. Legacy programs can become self-protecting.

But the proposed cure can easily worsen the disease.

1) Less oversight means less truth

If you reduce reporting and regulatory burdens too far, you don’t just make it easier for startups—you make it harder for the Pentagon to answer basic questions:

Is the product real or a demo?
Can it scale or only impress in a pilot?
Is the price fair?
What does it actually cost to sustain for ten years?

When the government loses visibility, it doesn’t become “agile.” It becomes blind—and blind buyers get fleeced.

2) Fixed-price doesn’t magically shift risk away from the government

In theory, fixed-price contracts put risk on contractors. In reality, when programs slip, costs balloon, or strategic needs change, the government is still the one that can’t walk away. National defense isn’t a casual subscription you cancel. The state ends up absorbing risk anyway—just later, and often at higher cost.

3) Consolidation will happen again—just with new logos

It’s easy to denounce the consolidation wave of the 1990s while quietly practicing the same playbook today. Tech-enabled defense firms are already acquiring aggressively, building mini-conglomerates. That may be rational from a business perspective. But it doesn’t automatically produce a resilient industrial base. In many cases, it produces a familiar outcome: fewer independent suppliers and more dependency on a small set of players.

4) Software won’t solve “unglamorous” capacity

Even if Silicon Valley perfects faster software procurement, it won’t magically fix the ability to build ships, produce munitions, manufacture propulsion systems, or maintain depots at surge levels. The modernization crisis is real—but so is the production crisis. And production crises are solved with facilities, tooling, workforce pipelines, and long-term investment—not just better interfaces and quicker pilots.

The real fix: rebuild the state as an intelligent buyer and producer

If you want defense innovation that actually scales and lasts, the priority isn’t “unleash the market.” It’s rebuild the government’s ability to shape and manage the market.

That means embracing a fact many people hate saying out loud:

Defense innovation requires government intervention—not as bureaucratic meddling, but as industrial strategy and statecraft.

Here’s what that looks like in practice.

Invest in surge-ready capacity—on purpose

Resilience costs money. If you want suppliers who can ramp production quickly in a crisis, you have to pay for that readiness. It won’t emerge from quarterly incentives. It has to be designed into contracts and supported with steady investment.

The cheapest time to build capacity is before you need it.

Restore and modernize the “organic” industrial base

Public arsenals, shipyards, depots, labs—these aren’t nostalgic relics. They are strategic infrastructure. You don’t need government to do everything, but you do need it to retain enough capability to:

understand production realities,
keep critical know-how alive,
provide a backstop when private capacity collapses or consolidates,
and sustain competition by preventing total dependence on a few firms.

Rebuild the acquisition workforce as a strategic asset

You cannot run a trillion-dollar defense apparatus with a hollowed-out corps of overworked acquisition officers and limited technical staff.

If Washington wants better outcomes, it needs more people who can actually do the work:

acquisition professionals trained to negotiate and manage complex contracts,
systems engineers and product managers who can translate needs into buildable requirements,
technologists who can evaluate claims and integrate new tools responsibly,
and market researchers who can prevent “vendor capture” before it happens.

This is not bureaucratic bloat. It is basic competence.

Build a serious “build vs buy” discipline

Some things should be purchased. Some things should be built in-house. The point is not ideology—it’s strategy.

The Pentagon needs a repeatable way to decide:

What must remain sovereign capability?
What can be safely sourced from commercial markets?
What requires a hybrid model?
What must be maintainable and surge-ready even when profits aren’t?

The core idea Silicon Valley misses

A lot of defense-tech rhetoric treats the government like a slow customer and the private sector like the true engine of progress. But in national security, the government isn’t just a customer. It’s the steward of long-term public risk.

Markets optimize for profit. Defense optimizes for survival under uncertainty.

That’s why the first questions in procurement shouldn’t be “How fast can we buy?” but: