The Startups Rebuilding American Industry | The AD50

1. Key Themes

The Industrial Base Is the Prerequisite for the Defense Base

The year's central lesson is that defense capability cannot exist without manufacturing capability, and that geopolitics has made this tangible. Erin Price-Wright framed it: "We've been saying for a while, you don't have a defense industrial base without an industrial base. And I think this year really put that into perspective with conflicts going on in the world and realizing that the hollowing out of our industrial base that's happened over the last several decades actually has real consequences on the battlefield for soldiers" [00:02:23]. Connor Love added that this is a convergence of three forces, geopolitics, capital markets, and an entrepreneurial base ready to build: "it's just this conglomeration of geopolitical factors of capital markets and then entrepreneurs that are kind of all just coming together" [00:02:48].

AD50 Is Organized End-to-End: Inputs, Factories, Systems

The a16z framework this year treats American Dynamism as a connected chain rather than isolated categories. Love explained: "you really need to start looking at American Dynamism companies as a totality from end-to-end, from the input that comes out of the ground to the factory that makes the thing, to the end system" [00:03:28]. Most investors and founders focus on end systems (missiles, ships, planes), but the bottlenecks sit upstream. Love: "a lot of times what we focus on... is just the systems companies... But the problem is, if you can't iterate on that design and iterate on that prototype, you... move at a slower pace" [00:16:03].

Defense as the Anchor Customer That Subsidizes Commercial Industrial Capacity

Defense spending can pay the CapEx for first-of-a-kind factories, which then serve commercial markets. Price-Wright: "if you have a defense buyer as the first buyer to help you kind of pay for, subsidize a lot of that CapEx investment, that's great because you know what else has PCB boards? Robots. You know what else has PCB boards? Hearing aids" [00:08:29]. Love: "the Department of War is saying... we're going to try to actively rebuild some of this industrial base and by doing that, they're actually changing that unit economic decision from the start for these companies" [00:07:43]. The DJI arc is the proof of concept: "the folks behind DJI are now the largest producers of robots in the world because what is very similar to a drone motor is a robot actuator, which is the hardest part of building a robot" [00:06:02].

Vertically Integrate Only Where No Industrial Base Exists

Both speakers pushed back on blanket vertical integration. Love: "I take the ideology that you shouldn't vertically integrate everything" [00:09:04]. Price-Wright: "SpaceX and Tesla were doing first of their kind. Like, there didn't exist an industrial base in the US to support them. And so they had to build it themselves. I think where there are high-quality suppliers, like, you should use suppliers" [00:09:35]. The exception is when owning the asset is required to prove technology change, as with Mariana Minerals: "they also recognized that they wanted to actually own the entire value chain to be able to really put their money where their mouth was and capture the end value of the new technology they were bringing" [00:10:46].

AI in the Physical World Is a Process Tool, Not a Silver Bullet

Love was direct about the limits: "AI will be a phenomenal tool that can be used in the physical world. The reality is AI cannot turn a wrench yet. It will at some point, but we're far off that" [00:00:06]. The value comes from rethinking 15 to 20 discrete process steps, not one AI product. Price-Wright described the concrete use case: manufacturing technicians reconfiguring lines with LLMs, "instead of having to call a system integrator... it takes 3 months and hundreds of thousands of dollars to kind of reconfigure a line for a new setup. You can have a manufacturing engineer use Claude and do that in, you know, 20 minutes" [00:15:22].

The "Manufacturing Middle": Co-Engineering Replaces Receive-Mode Suppliers

Love's most original framework: tier-2 and tier-3 suppliers have spent 40 years in "receive mode," and new end systems will force them to co-engineer. "Boeing... just pick up the phone and they call Honeywell... Honeywell says, I don't want to build a new landing gear again. We have this landing gear that meets the spec" [00:32:41]. His prediction: "you're going to have this middle layer of supplier companies that... will effectively be co-engineers with the OEM of rethinking... the individual subcomponents" [00:34:14]. He called it "the American manufacturing asset class," distinguishing it from private equity roll-ups: "That's not fundamentally changing the things that we build" [00:35:18].

Private Capital Now Builds Ahead of Government Requirements

The historic defense model was government-defined requirements funded after the fact. Love: "instead of sitting and waiting for the government to tell you what they want, the government signals and you see it and you build in advance by using private capital, and that just has not happened in the history of defense" [00:19:30]. Castelion is the example: "if you go look at the last 3 years of that business, I mean, it was private venture capital... entrepreneurs like taking the risk and building the thing in advance" [00:20:03].

The Economics Changed: These Can Now Be the Most Valuable Companies

Love: "those companies weren't just the most important thing to us. Now they can be the most valuable companies" [00:22:23]. Price-Wright added that the talent supply followed the economics: "it hasn't been economically worth it for a young, smart person... But I think companies like Palantir, Anduril, SpaceX have proved that you can build a really big business." Her anchoring stat: "the top 10 most valuable companies in the world... the majority of them are hardware companies" [00:00:14].

Capital Markets Are the American Advantage Over China's Top-Down Model

Love: "we do not have a top-down, directed industrial policy... the money will follow the entrepreneur. The money will follow the success" [00:41:33]. He cited J.P. Morgan, private equity, and credit firms all moving into the physical economy. Price-Wright gave the datapoint: "if you look at just the first half of 2026, more risk dollars went into the defense ecosystem than all of last year combined, which was many times what went into the defense ecosystem in all of the previous year combined" [00:43:46]. She said the "Valley of Death" narrative is starting to disappear.

Founder Lineage: Second-Generation SpaceX/Anduril/Palantir Founders Solve Their Own Supply Chain Pain

Price-Wright: "a great way to light a fire under an entrepreneur is for them to be banging their head against the wall on a sort of engineering or supply chain problem... I'm going to go solve that more generally" [00:38:11]. Those founders have "automatic access to a customer base that's captive," and she compared it to YC companies selling to each other: "that's how Silicon Valley was built. And what makes me really optimistic is the extent to which we're starting to see that in these kind of physical world ecosystems" [00:38:58].

2. Contrarian Perspectives

"Dual-Use" Is Partly a Folly, but the Dual-Use Outcome Is Inevitable

Most investors pitch "dual-use" as a product strategy. Love argues it's a distraction for startups while being true of the industrial base itself: "the term dual-use is talked about a lot in investing, and to some extent, I think it's a little bit of a folly, like focusing on too many things at once is not good for a startup. But it is 100% true of what's happening with any factory that's being built today" [00:04:58]. The Castelion team proved the point by refusing multi-product pressure: "Be multi-product, go build this small drone here.' And they're like, 'No, we need to do this one thing really well'" [00:39:52].

The Thing Holding Back US Energy Is Will, Not Technology

Asked how to close the US-China power-build gap, Price-Wright rejected the technology framing: "I don't think that is a technology problem. I think that's a we need to do some soul-searching about our will as an American people. And do we want conforming projects to be built? Do we want to grow? And if we do, I think the technology and the capital will follow" [00:00:47]. Love echoed it as "an everything problem" and noted "it's ours as a government and a nation to... mess up and to get in the way of this" [00:46:04].

Even Half-Delivered Demand Forecasts Permanently Change the Supply Chain

The usual skeptic's response to new energy and military demand is "not all of it will materialize." Love argues it doesn't matter: "once you start the snowball of the demand cycle moving in one way, even if you deliver half of that requirement... you're not going back... when you rethink something first principles... a lot of times that comes with better cost, better efficiency, you're just going to buy that thing now going forward" [00:36:54]. Price-Wright: "better serviceability, easier to manufacture. Yeah, you're never going back" [00:37:35].

Rational Incumbent Behavior Is What Freezes Innovation

Love defends Boeing and Honeywell while explaining why they stagnate: "It's not Honeywell's fault... They're going to build more 737s. They're going to get paid the same way if it's a 40, 50-year-old landing gear or if it's a new landing gear" [00:32:41]. The incentive is flat between old and new components, so innovation has to come from outsiders who co-engineer.

Physical Industry Will Reduce Wealth Concentration, but Only If Communities Share the Upside

Counter to the view that tech concentrates gains on the coasts, Price-Wright argues physical-world companies spread them: "Less concentration of wealth in these pockets, in these elite pockets on the coasts. It's a good thing. But it will take time" [00:49:02]. She warned that unlike software, "someone's building a data center, a power plant, a mine in your backyard... if you don't feel like you have participation in that economic upside, it's hard" [00:47:16].

3. Companies Identified

Mariana Minerals

Copper mining company that bought a mine and is rebuilding the entire value chain with software. Cited as the exemplar of full-stack ownership to change a legacy industry. Elena Burger introduced it: "Mariana Minerals, which bought a copper mine and is kind of reinventing things end to end." Price-Wright: "they do everything from acquiring the asset, developing the asset, mining the asset, operating the asset, and refining the material themselves across multiple different minerals. And then they sell a final commodity product in the market" [00:11:10]. Love: "It's not just saying, hey, we're going to be better at extracting copper. It's rethinking everything" [00:12:08]. Price-Wright noted the data flywheel: "once you own that process end-to-end and are using software at every stage... you actually capture that data and then it feeds back" [00:12:20].

Hadrian

Advanced manufacturing company (CEO Chris) building factories with software at every node. Love: "it's not one silver bullet that all of a sudden makes a factory or a mine perfect. It's actually 15 or 20 different things... what it takes to quote and even take an order of a part... how you inspect it and transfer it" [00:12:35]. Also noted for visible site culture: "Hadrian's [flag] was pretty big last time I saw" [00:26:16].

Castelion

Hypersonic missile company; example of building ahead of government demand and of single-product focus. Love: "like Castelion building a hypersonic missile... it's really hard to build a missile that goes that fast and is reliable... the last 3 years of that business, I mean, it was private venture capital" [00:20:03]. On focus: "they focused for 2 and 1/2, 3 years on the first hypersonic missile that was really hard to build" [00:39:52].

Anduril

Defense systems company, a "neoprime." Cited for its production facilities and as a source of founder lineage. Love: "go look at what Anduril's built in Ohio and now on the East Coast in Baltimore... at some point, in a very soon timeline, going to be producing thousands, tens of thousands, hopefully hundreds of thousands of systems. We haven't done that in America in a long time" [00:17:09]. Price-Wright named it with Palantir and SpaceX as proof that big businesses can be built in these categories [00:23:18].

Saronic

Autonomous maritime systems company building large shipbuilding capacity. Love: "go look at what Saronic has built with the ship. I mean, these are massive, huge facilities" [00:17:09], and "Saronic's making a lot of boats" [00:29:04].

SpaceX

Cited as the template for vertical integration, design for manufacturing, and engineer accountability. Love: "the few modern day companies that have built anything at scale are probably SpaceX and Tesla" [00:09:04]. On culture: "at SpaceX, the concept of being a responsible engineer and like, if this thing doesn't work, like it's my ass. I'm going to make sure... I'm going to obsess over this thing" [00:40:23].

Tesla

Built a power electronics and drivetrain supply chain from scratch. Price-Wright: "the whole power electronics, the whole drivetrain, you know, they had to really build completely from scratch based off papers that were published in the '70s that had been collecting dust for 40 years" [00:10:22].

Palantir

Named as proof that large businesses can be built in defense and government categories, and as the origin of a founder lineage (Price-Wright worked there). Price-Wright: "companies like Palantir, Anduril, SpaceX have proved that you can build a really big business operating in these categories" [00:23:18].

DJI

Chinese drone maker used as the case study for the consumer-to-defense-to-robotics arc. Price-Wright: "Started out building consumer drones and recognized over the late 2010s that there was a large market for drones in both public safety and national defense... now the largest producers of robots in the world" [00:06:02].

Amca

Company that partners with small machine shops. Love: "they show up to these really small machine shops that exist. They partner with them. They don't cut the management and fire people. They say, how can we build new products and grow the kind of business that you have today?" [00:48:03]. He also listed it among large facilities being built: "what Amca will do" [00:26:00].

Valor

Mentioned for having a very large American flag at its site. Price-Wright: "I think Valor's is pretty big" [00:26:22].

Joby and Archer

New-aircraft makers cited as examples of new end systems driving subsystem redesign. Love: "Joby and Archer, we're building new planes, too. Like, we're building new end systems" [00:32:20].

Boeing

Cited as a good company whose 737 supply chain illustrates "receive mode." Love: "Boeing's a great company, important to the fabric of America. But they just pick up the phone, and they call Honeywell" [00:32:41].

Honeywell

Tier-1 supplier that reuses decades-old components such as landing gear that meet spec. Love: "It's not Honeywell's fault, but what Honeywell does is they say, I don't want to build a new landing gear again" [00:32:41].

Rio Tinto

Mentioned as a long-standing, good mining company, in the context of first-principles rebuilds. Love: "Rio Tinto's been around for a long time" [00:23:44].

J.P. Morgan

Cited as evidence of capital markets moving into the physical world. Love: "J.P. Morgan has committed, you know, trillions of dollars towards both investing in the physical world, not just defense, but factories and other things" [00:42:08].

Waymo

Mentioned in passing as a place where these systems will end up. Love: "these systems will be used in the Waymo that we took today" [00:05:30].

Y Combinator

Used as the model for ecosystem-building through customers who tolerate each other's early products. Price-Wright: "people make fun of YC for all YC companies kind of selling to each other, but like that's how it works... That's how Silicon Valley was built" [00:38:58].

Claude (Anthropic)

Named as the tool a manufacturing engineer can use to reconfigure a line in 20 minutes instead of 3 months with an integrator [00:15:22].

4. People Identified

Chris (Hadrian founder)

Founder of Hadrian. Love credited him with fresh-slate factory design: "when you look at what Chris has done at Hadrian... when you kind of start from a fresh slate like Chris has in building his factories... you can look at every single node from a first principles perspective and say, hey, was this process that's been done the last 100 years the right process?" [00:12:35].

Turner (CEO, Mariana Minerals)

Cited by Price-Wright as the archetype of the missionary founder. "when I talk to Turner, who's the CEO of Mariana Minerals, and he's flying all over the world to build his company. He's like out where there's no cell phone signal, where there's no electricity, touching dirt, spending like weeks and months away from his family. Like, it has to be worth it" [00:20:51].

Elon Musk

Referenced as the model for vertical integration and first-principles building. Love: "we can look to Elon for a lot of answers here" [00:09:04]. Elena Burger: "they make any problem an Elon-sized problem" [00:18:01].

Brian Schimpf (founder, Anduril)

Cited by Love for the point on trillions of dollars of private capital following technology and production success: "Brian Schimpf was talking about it, the founder of Anduril... this concept that we have literally trillions of dollars of private capital in our capital markets that inherently follow success" [00:41:33].

Isaiah (El Segundo founder)

Founder Love visited years ago, memorable for the on-site bench press: "when I first visited many years ago, Isaiah in El Segundo, he made me do a rep on his bench press" [00:28:29].

The Castelion founders

Praised for single-product discipline rooted in SpaceX culture: "I've seen this with the Castelion guys... they focused for 2 and 1/2, 3 years on the first hypersonic missile" [00:40:23].

5. Operating Insights

Visit Portfolio Factories Every Three Months to Measure Build-Out Velocity

Price-Wright's diligence tactic: "One of my favorite things to do is visit the same company sort of multiple times within a 12-month period, whether they're in the portfolio or not, because you learn a lot by how fast a factory build-out happens... you go once and then you go 3 months later, you go 3 months later, and you go 3 months later" [00:27:01]. Pace from empty warehouse to power-constrained, chaotic expansion ("we need to get the city out to give us more power because we have too many machines plugged in") is itself a signal of execution quality.

Design for Manufacturing From Day One, and Make Lines Reconfigurable

Love: "if you have a more transferable line, a more adjustable line, number one, you can also build the prototype faster, but then two, you can then set up each one of these different components in a way where it actually can get to rate production quicker" [00:16:53]. Price-Wright: "you're designing for manufacturing from the outset" [00:17:05]. Prototype iteration speed and production scalability should be treated as one design problem.

Put Software Tools in the Hands of Technicians, Not Just Integrators

The low-mix/high-volume unlock: technicians with process intuition plus LLM tooling can reprogram lines on the fly. Price-Wright: "this is not someone with a college degree, maybe they are, but not necessarily, who's studied software engineering, is now able to use these tools to reconfigure their manufacturing lines... using the fact that they understand the manufacturing process" [00:15:22]. Operators should shift line-change authority from outside integrators to internal process experts.

Co-Locate Suppliers and Customers to Move Quality Metrics

Concrete data: one portfolio company's PCB yield went from needing to over-order 2x because "50% of them wouldn't pass a quality test" to "70, 80, 90% passing quality tests" after switching to new US suppliers. Price-Wright: "we underestimate how much the close co-location and co-development of some of this technology with the end customer and their suppliers really matters to drive quality higher" [00:31:43].

Do One Hard Thing Until It Works, Then Reuse the Process Knowledge

Price-Wright: "if you can do that one thing really well, that sets you up to do other things really well... You've developed a supply chain. You have your relationships. You've developed manufacturing... You've learned how to build a factory to a high-quality, high-spec. You've learned how to build that trust with customers. But it doesn't work unless you deliver on that one core thing" [00:40:50].

6. Overlooked Insights

Zero-Defect Tolerance Behind "Throw Away 50%" Is a Hidden Cost Layer Across the Hardware Economy

Buried in an anecdote is a major economic signal: a portfolio company "3 years ago would order two times the number of PCBs that they needed because they assumed that they would have to throw 50% of them away for quality issues" [00:31:10]. Hardware companies have been silently paying a 2x procurement tax, and cash-flow and working-capital models across robotics and defense are built on assumptions that domestic quality improvements are now dissolving. Any hardware investment model that embeds legacy yield assumptions is mispricing the upside from onshored suppliers.

The "Soft" Constraint Is Community Consent, and the Winners Will Be Whoever Solves Economic Participation

Price-Wright slipped in the most structural point of the episode near the end: unlike software, "someone's building a data center, a power plant, a mine in your backyard... You physically see it. It affects your life. And if you don't feel like you have participation in that economic upside, it's hard" [00:47:16]. Combined with her conclusion that "capital markets long-term will solve [this] really well because companies will figure out how to bring communities along," it implies the next durable moat in physical industry is a repeatable community-ownership and benefit-sharing model, which most of the AD50 pipeline has not yet productized. Amca's partner-don't-cut approach with small machine shops [00:48:44] is an early example worth watching.