America's Manufacturing Hollowing Isn't a Capability Problem, It's a Financial Return Problem
Chris Power says America manages industrial policy like a hedge fund's financial statement, while China manages it like national security; tariffs are just an imperfect corrective tool, and the core issue is to stop American consumers from subsidizing rivals' industrial capacity.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
The reserve currency is the first step of decline
Chris Power gives a four-step cycle: great powers first have a generative population and the best industrial capacity, then win or prevent a war, then become extremely lazy, and finally outsource the very things that made them successful. He says this cycle repeats roughly every 150 years: the Netherlands gives way to Britain, Britain gives way to America, and each time industrial power is re-traded for financial power, the world's reserve currency changes. This judgment was the starting point for him founding Hadrian: you think you're Mike Tyson at his peak, but your opponent thinks you're the Mike Tyson right before Buster Douglas knocked him out.
— Chris Power90% of the factory floor is a software problem
Chris Power says the biggest misconception about advanced manufacturing is that seeing people moving things in a factory means it's a humanoid robot problem. In reality, 90% of advanced manufacturing is smart people sitting at computers doing digital engineering, or operating a very dumb computer on the shop floor. Welders or machinists have to stand in front of the machine because the system isn't closed-loop; they do trigonometry in their heads to judge whether the tool is about to break. So Hadrian positions itself as a software company, breaking physical tasks into high-judgment setups and then handing them to robots, CNC, or welding systems to execute.
— Chris PowerSoftware is done in Silicon Valley, hardware in Los Angeles
Tom Mueller says the aerospace talent ecosystem in Southern California is recovering; three or four years after SpaceX was founded in 2002, the aerospace employment curve in the South Bay started to tick up. He lists what's happening simultaneously in the region: nuclear energy companies, drone companies, defense companies, 3D printing companies, and Anduril in Orange County. The two companies poach from each other and compete for talent; Northrop Grumman is right across the street from Impulse. Tom's summary: software is done in Silicon Valley, hardware in Los Angeles.
— Tom MuellerVertical integration is to learn the problem, not to save money
Tom Mueller explains the real logic of SpaceX-style vertical integration: the first development engine or spacecraft must be built in-house, because you need to learn all the assembly and machining problems during the prototype phase. If you hand it to a machine shop from the start, and the drawings have problems or the parts don't fit when they come back, you learn nothing. Once the design converges, then you outsource. Chris Power adds that every company should keep 10% to 15% of in-house capacity, because there's always a "hot pot" moment when you need a genius to make something in an hour to hit a flight milestone, but scale must happen externally.
— Tom MuellerOutsourcing decisions are about trust, not finance
Chris Power says what's bought and sold in manufacturing is trust. Hadrian's vertical integration strategy is to outsource as much as possible to every reindustrialization startup and small business, because there are too many projects to run. But they will vertically integrate the links where they "don't believe you can deliver on time." This isn't a financial trade-off; they'd rather spend an extra two million dollars on equipment and hire a team than bet the risk on a traditional foundry that has been late on every submarine program for the past thirty years. His exact words: most of America is incompetent when it comes to trust and scheduling.
— Chris PowerIn WWII, America was China
Chris Power uses WWII as an analogy: German engineers crushed America on engineering, but America wore them down with McDonald's-style mass production and better logistics. He says America played the role in WWII that China plays today—not by having the best engineering, but by producing the most. Now the situation is reversed: U.S. missile stockpiles run out in seven days in war games, China built 2,600 ships while America built six. His conclusion: you can have the best Raytheon missiles, but if you can't mass-produce them, you have no options.
— Chris PowerThe robot bottleneck is the speed of training data
Chris Power says the robot competition ultimately depends on software, and software, especially AI, now depends on the velocity of training data, which in turn depends on physical-world deployment—i.e., mass manufacturing. So this isn't a friendly competition between American companies over "whether Tesla sells more Optimus units or Google wins"; it's a civilization-level confrontation. Tom Mueller adds from the space angle: building on the Moon and Mars requires robot legions rather than humans, because the environment is too dangerous and too expensive; a good robot needs three things—dexterous hands, vision (largely solved), and an AI brain.
— Chris PowerTariffs are an imperfect corrective tool
Chris Power says manufacturing has four inputs: energy, labor, cost of capital, and export tariffs. He has run the numbers: even if Hadrian's labor and energy costs were both zero, importing components from China would still be three times cheaper than buying raw materials in America and making them yourself. He believes China has operated industrial policy as national security since 1985, subsidizing energy, labor, capital, and export tariffs, while America manages industrial policy like a hedge fund's financial statement returns. The problem with tariffs is enforcement: banning only China and not Singapore means Chinese companies route through Singapore; differences in definitions of raw materials and assembly also inadvertently hurt American companies importing components from Mexico.
— Chris PowerIn their own words · checked verbatim
every great power of all time does four very simple things. they have some generative population function, they have the best industrial capacity, and then they win or prevent a war, and then they get incredibly lazy and they end up with a financial industry and they offshore what made them successful in the first place.
Chris Power2:02
what's happening in advanced manufacturing is there's an incredible amount of smart humans either at a desktop doing digital engineering or they're at the factory floor driving an incredibly dumb computer. And that's 90% of advanced manufacturing.
Chris Power10:10
I think in manufacturing, you sell and buy trust. So our vertical integration strategy is we want to outsource everything possible to every reindustrialized startup, every small business that we possibly can because we have so many programs to run. But we vertically integrate where we just don't believe you're going to get the job done.
Chris Power16:18
We were China in World War II. We just outproduced them. We had better logistics. We made more stuff because we weren't rich, but we're the industrial superpower and we became rich because of Bretton Woods.
Chris Power19:20
you can have the best engineering in the world, like you've got the best Raytheon missile, but if you literally can't just make mass, which is ultimately a production issue, there's no choice.
Chris Power20:22
if you dial our factory down to 0% labor costs, 0% energy, whatever, Chinese imports are still three. I can make a component for a customer. I can buy the raw material in the United States, and it's three times the cost of the component imported from China.
Chris Power29:40
China has always treated industrial policy like national security, and since 1985, America has treated industrial policy like financial spreadsheet returns for hedge funds.
Chris Power30:45
Figures
| Rockets SpaceX launched last year | 165 | 5:05 |
| Satellites SpaceX has in orbit | 10,000, three times all other countries combined | 5:05 |
| Satellites Impulse Space has in orbit | 3 | 8:10 |
| Closest distance between Impulse's two satellites | 200 meters | 8:10 |
| Helios payload increase over Falcon 9 | About 5x (to the Moon or Mars) | 8:10 |
| Share of Virginia-class submarines Hadrian is involved in manufacturing | 30% | 9:10 |
| Recommended in-house capacity share for companies | 10% to 15% | 15:17 |
| Time for U.S. missile stockpiles to run out in war games | 7 days | 19:20 |
| Shipbuilding comparison with China | China built 2,600 ships, America built 6 | 19:20 |
| Cost advantage of Chinese imported components | Even if U.S. factory labor and energy costs were zero, Chinese imports would still be one-third the cost of making them domestically | 29:40 |
Glossary
- fourth turning
- A historical cycle theory referring to generational phases of crisis and rebuilding in society.
- vertical integration
- A company bringing upstream and downstream supply chain links in-house rather than relying on external suppliers.
- lights out
- A factory running 24 hours unattended, usually referring to highly automated production lines.
- digital twin
- A real-time digital model of a physical object or process, used for simulation and optimization.
- CapEx
- Spending by a company to acquire long-term assets (equipment, buildings).
How to listen
Founders and investors focused on American reindustrialization, defense tech, and aerospace supply chains; manufacturing practitioners who want to understand the practical execution difficulties of tariffs and industrial policy.
The tour ad and guest introduction in the opening 0:00-1:01 can be skipped.