America keeps trying to rebuild the industrial capacity of the past. What we actually need is the capacity to produce the outcomes the future demands.

Years ago, I escorted a congressional staff delegation to Connecticut. Congressman John Larson joined us in his district and proudly reminded the group that Connecticut was the birthplace of Colt firearms.

It was a fitting point of pride. Samuel Colt and the manufacturing ecosystem that grew around companies like his helped define the Industrial Revolution. Interchangeable parts, precision manufacturing and mass production changed Connecticut, the country and eventually the world. For more than a century, industrial capacity was inseparable from national power.

I remember thinking at the time how much that history shaped the way Larson talked about his district.

But history and strategy aren’t the same thing.

Larson recently lost his Democratic primary after 14 terms in Congress. I’m not suggesting an anecdote about Colt explains an election result. But there is something instructive in the juxtaposition. Leaders can become so practiced at describing where power came from that they lose sight of where power is going.

We may be doing exactly that with the American defense industrial base.

Industrial Was an Era

We use the phrase constantly: rebuild the industrial base, strengthen the industrial base, invest in the industrial base.

But industrial was an era.

Since Colt, we’ve moved through electrification, telecommunications, computing, the internet, the knowledge economy, cloud computing and now artificial intelligence. Software has become a weapon system. Data has become an asset. Commercial companies launch satellites. Autonomous systems combine hardware, software, sensors and AI. Capital can move a technology from an engineer’s laptop to meaningful production before a traditional acquisition program completes its requirements process.

Yet we’re still using a term born of smokestacks and assembly lines to describe the national capability we need for the future.

Manufacturing matters. Shipyards matter. Machine tools matter. Munitions production matters.

But they aren’t the whole thing.

What we’re really talking about is the American capacity base.

Capacity includes manufacturing, but it also includes software, compute, energy, data, AI, systems integration, engineering talent, capital, supply chains, test infrastructure and logistics. Increasingly, our advantage depends not simply on possessing these things, but on our ability to combine them quickly.

Capacity is the ability to produce an outcome.

The Munitions Lesson

We don’t have to imagine what a capacity shortfall looks like. We’re living through one.

Wars in Ukraine and the Middle East have exposed something uncomfortable about the world’s most capable military: we can consume certain weapons far faster than we can replace them.

That realization has generated enormous attention around munitions production. Production lines are expanding. Facilities are being built. Congress is appropriating billions. Companies are investing in additional equipment and workers.

All of that is necessary.

But calling this an industrial-base problem understates what we’ve learned.

A missile isn’t produced by a factory. It’s produced by a network of capacities.

It requires energetics, rocket motors, guidance systems, microelectronics, castings, forgings, specialized materials, software, testing facilities, skilled labor, machine tools, suppliers and capital. Some components may have only one or two qualified domestic sources. Expanding final assembly does little good if a component buried several layers down in the supply chain remains the limiting factor.

For decades, we optimized much of defense production for efficiency rather than surge. We tolerated single-source suppliers because peacetime demand didn’t justify redundant capacity. We bought weapons in quantities that made sense against annual budgets rather than against the possibility that someday we might have to replenish them at wartime consumption rates.

Then the world changed.

The lesson shouldn’t simply be that we need more munitions factories.

The lesson is that we need to understand and invest in the entire chain of capacity required to turn raw material, knowledge, energy, capital and labor into a weapon at the speed and scale the strategic environment demands.

We Can’t Train Our Way Out of This

I recently spoke with the founders of an American company developing robotic welding systems. The conversation brought another part of the problem into focus.

We have spent enormous energy trying to rebuild the skilled trades required for shipbuilding and defense manufacturing. We’ve built schoolhouses near shipyards. We’ve created apprenticeship programs. We’ve advertised for workers. We should keep doing those things.

But we’re not going to train our way out of the problem.

The arithmetic doesn’t work.

The American welding industry estimates a shortage measured in the hundreds of thousands. At the same time, experienced welders are retiring faster than we can replace them. A child born today won’t enter welding school for roughly 18 years. Add training and then the thousands of hours of experience required to perform the most demanding work on submarines, nuclear systems and other critical applications.

No recruiting campaign changes the biology.

Meanwhile, machines can be built on an entirely different clock.

That doesn’t mean replacing the American welder. It means changing what one American worker is capable of producing.

One of the founders described something he had seen firsthand as a welding foreman. Young workers who might not be attracted to spending ten hours wearing welding leathers in the confined spaces of a ship often take naturally to programming and operating robotic welding systems. Instead of asking one worker to hold one welding gun, imagine that worker directing, monitoring and improving ten robotic welders.

Now we’ve multiplied the worker.

This company offered another statistic that should get our attention. The United States has just over 300,000 industrial robots in service. China has roughly two million. China is not merely competing with a much larger labor pool. It is aggressively multiplying that labor pool through automation.

An industrial-base solution asks how we can get 200,000 additional welders.

A capacity-base solution asks how we produce the welding output of 200,000 additional welders.

Those are two radically different questions.

We Already Have a Capacity Tool

Consider Title III of the Defense Production Act.

A decade ago, Title III was a comparatively small pot of money, often measured in tens of millions of dollars annually, intended to address specific weaknesses in the defense industrial base.

Today we’re talking about hundreds of millions – and at times – billions of dollars.

COVID changed the conversation. Suddenly, supply-chain fragility wasn’t an abstraction buried in a Pentagon report. Americans experienced it directly. The Defense Production Act became a mechanism for expanding production of critical medical supplies and addressing vulnerabilities that had little to do with our traditional image of tanks rolling off assembly lines.

More recently, critical minerals, rare-earth elements and magnets have received enormous attention because we finally recognized the national-security implications of allowing essential capacities to migrate overseas.

Think about what Title III is really doing.

It’s not preserving an industrial era.

It’s buying down national capacity risk.

A shortage of critical minerals is a capacity problem.

Insufficient munitions production is a capacity problem.

A shipyard that can’t find enough welders is a capacity problem.

Dependence on a foreign source for an essential component is a capacity problem.

And robotic welding may be a capacity solution.

So might additive manufacturing. AI-assisted engineering. Automated inspection. Digital engineering. Commercial software. Autonomous systems. New energy sources. Better supply-chain visibility. Or a technology we aren’t talking about yet.

None of these eliminate manufacturing or people.

They multiply what both can accomplish.

Stop Counting Inputs

There is an understandable instinct in Washington to measure inputs.

How many welders did we train?

How many apprentices enrolled?

How many factories did we open?

How many jobs did an investment create?

Those are useful measurements.

But national power ultimately depends on output.

How many submarines can we build?

How quickly can we replenish a depleted magazine?

How many autonomous systems can we field?

How quickly can we modify software after an adversary changes tactics?

How rapidly can we replace a critical component when a foreign supplier disappears?

Those are capacity questions.

If we need twice the output, the industrial-era instinct is to ask where we’ll find twice the workers and twice the factories.

The capacity-era question is different:

How do we produce twice the output?

Maybe the answer is another factory. Maybe it’s a second source for a critical component. Maybe we need more workers.

But it may also be automation that allows one worker to produce three times as much, software that reduces inspection time, a redesigned component that can be produced additively, AI that compresses engineering work, or a multiyear procurement that gives a supplier enough certainty to invest its own capital.

We keep diagnosing capacity problems and prescribing industrial solutions.

Sometimes the industrial solution will be exactly right. But if that’s where our thinking automatically begins, we’ve constrained the solution before we’ve properly defined the problem.

Don’t Rebuild 1955

There’s a powerful political attraction to the word rebuild. It suggests America once possessed something formidable, allowed it to deteriorate and now simply needs to put it back.

Some of that is true. There are manufacturing capabilities we should never have surrendered and others we urgently need to restore.

But rebuilding can’t become our strategy.

America isn’t competing to recreate the industrial economy of 1955. We’re competing to determine what national power looks like now and what capacities we’ll need next.

So perhaps we should stop asking:

How do we rebuild the defense industrial base?

And start asking:

What capacities must America possess that we don’t have enough of today?

Sometimes the answer will be steel, ships, machine tools or munitions.

Sometimes it will be software engineers, compute, energy, critical minerals, data, autonomous systems or capital.

Most often, it will be some combination of them.

The industrial base helped make America powerful. We shouldn’t abandon it. We should automate it, digitize it, integrate it and multiply it.

We don’t need to recreate the workforce or industrial structure of the last century.

We need the capacity to outperform it.