To me it seems clear that the next decade produces more electronics than the last four combined. Robots, drones, data centers, grids.
Since you and I both know this wave is coming, the interesting question is whether America can actually build it. And the honest answer right now is mostly no.
America used to build its own electronics. Radios, televisions, computers. Through the 80s, the things we invented were things we made. Then, one product category at a time, we stopped making things in the US. Design stayed here; production went to Asia. Today the US invents most of the world’s electronics and manufactures almost none of them.
In 1990, the US made roughly 30% of the world’s circuit boards. Today it is about 4%. Ninety percent of global PCB production now sits in Asia. China alone makes over half.
1. The pieces of the puzzle
When people say “electronics manufacturing,” there’s actually four different layers, and America’s position in each one is very different.
At the bottom is the semiconductor. This is the most complex, most capital-intensive, and it’s where nearly all of the policy attention has gone. The CHIPS Act put over $50 billion here.
Next is the bare board. The laminated fiberglass-and-copper substrate that every chip has to sit on. No fab in Arizona changes the fact that the board itself gets etched in Shenzhen.
Then there’s assembly. The step where chips, resistors, capacitors, and connectors actually get placed and soldered onto that board. This is where a pile of components becomes a working circuit.
And finally the component supply chain feeding all of it. Most of which ship from a handful of factories concentrated in Asia.
We’ve spent tens of billions reshoring layer one and almost nothing on layers two through four.
I get why we started with the chip. It makes sense to focus on the part that is the most complex and proprietary.
Thing is - the chip is useless without the PCB that it sits on.
2. Where we are now
The domestic bare-board industry is shrinking. US bare PCB manufacturing revenue has contracted at nearly 3% annually since 2021, down to roughly $4.4 billion, with about 305 businesses remaining, most of them small shops.
The survivors didn’t make it by competing on price. They did the segments Asia couldn’t: high-mix, low-volume for defense, aerospace, and medical applications.
That change made sense for each individual shop. Collectively, it means America kept the expensive, one-off jobs and lost the factory. We can make a 30-layer board for a missile seeker. We cannot make a hundred thousand simple 4-layer boards for a drone fleet at a good price and on a fast timeline.
3. It’s hard but worth it
This is all usually greeted with a shrug. Labor is too expensive here, the supply chain is gone, China is twenty years ahead.
Giving up on making things wasn’t some conclusion America reached after a fight. It was the path of least resistance, repeated, for two decades, by people for whom it was individually rational.
But look at what’s actually stacked on top of this layer now: every drone, every robot, every rack of the AI buildout.
Currently hardware manufacturing looks like:
- Month long iteration loops for hardware teams, while their software counterparts ship daily.
- Defense and space programs that legally can’t send a board file overseas, all competing for a handful of domestic shops.
- A supply line with a single point of failure. If a crisis closes the Taiwan Strait, US hardware production stops.
The reason to fix it now is simple: most of the demand coming over the next decade hasn’t been allocated yet. The factories that will build boards for the drones, humanoids, power electronics, and defense systems don’t exist except in China. Rebuilding a mature industry from twenty years behind almost never works. Showing up early for a new one is a different problem, and a much more winnable one.
4. The comeback doesn’t look like the old factory
The good news is that America lost the volume, not the capability. On 4% of the world’s units, US shops still capture roughly a quarter of global board revenue. What survived the offshoring wasn’t the easy work, it was the hard stuff.
So the comeback isn’t getting back $2 commodity boards from Asia. That war is over, and winning it was never the goal.
The prize is the work that’s coming. Radar boards for drones, power electronics for the grid; this is complex, high-mix, regulated, expensive work where the surviving domestic capability is actually useful.
The industry that’s still here in the USA isn’t nothing. It’s a starting place. What it’s been missing is volume worth scaling for, and for the first time in thirty years, that volume is here.
5. What it takes
The fix is a few unglamorous things happening all at once.
The iteration loop has to be much faster. Quoting in minutes, software-native workflows instead of spreadsheets over email. Speed is what pulls the prototype work back. Prototypes are where production decisions get made.
The layers below assembly need attention too. Domestic assembly on imported bare boards is not a sustainable solution.
And the demand side has to commit. Defense programs, grid buildouts, robot roll outs, and the AI infrastructure wave are all publicly stressing about supply chain risk. The way to fix it is boring: place the orders domestically.
None of this requires beating China at anything. It requires being ready as the work arrives.
The next decade will produce more electronics than the last four combined. The only question left is where they get built.