How New Factories and Supply Chains Could Build the Next SpaceX

Elena Burger opens the discussion of the AD50 by asking why the American industrial base has become such a prominent theme. Erin Price-Wright and Connor Love argue that defense production depends on the broader industrial economy: decades of lost capacity have consequences for soldiers and allies, and recent conflicts have made that vulnerability harder to ignore.

The guests see several forces converging: geopolitical pressure, capital seeking opportunities in physical industries, and founders willing to build in them. They frame the AD50 as an end-to-end system, from materials extracted from the ground through factories to the finished products those factories make.

Burger asks whether defense demand can provide a tailwind for energy, robotics and AI. Historically, the guests say, government programs often initiated technologies such as spaceflight, the internet and nuclear energy, with commercial applications following later. Today private companies and investors are increasingly willing to initiate development themselves.

A startup still needs focus: trying to serve every market at once can be dangerous. But a factory, energy source or supply chain built for a defense customer can eventually serve civilian demand too. The same production capability may support vehicles, consumer products and other systems.

Price-Wright points to DJI as an illustration of transferable capability. The company started with consumer drones and expanded into public safety and defense; the experience accumulated in drone motors and production can also inform robotic actuators. The guests see the movement of process knowledge between military and civilian applications as a recurring pattern.

Government demand can help pay for early production and technology, while demanding consumers push products toward greater capability and lower prices. That relationship is not confined to drones: the speakers see similar links in energy, electronics and other layers of the industrial stack.

First-of-a-kind hardware is expensive, especially at small volumes. A defense buyer willing to pay more than an early commercial customer can support investment in machines and production capacity. As volume grows and that capital cost is absorbed, products that were once uneconomic for consumers may become viable.

A new printed circuit board factory illustrates the argument. Its equipment is costly, but a defense order can help finance it. Once the capacity exists, boards can also go into robots, hearing aids and other products. The benefit extends beyond the initial customer.

Asked how startups can reduce costs and move faster, the guests turn to SpaceX and Tesla. Both built substantial capabilities internally when the supply chains they needed were missing. Vertical integration is not a universal rule, however: when capable, reliable suppliers exist, using them can be the better choice.

Building the final system while also solving for valves, bearings and other components is a formidable burden. Tesla had to develop a battery supply chain, power electronics and drivetrain capabilities that were not available at the required scale. The lesson is to integrate where necessary, not to assume every component must be made in-house.

Burger raises Mariana Minerals, which acquired a copper mine. Price-Wright says the company wanted to apply technology throughout the value chain and capture the resulting value. Its work spans acquiring and developing assets, mining, operating and refining materials before selling the commodity.

The guests argue that changing a physical process sometimes requires owning the asset rather than merely selling software to its operator. Mariana's approach aims to rethink planning, operations, extraction and refining together. Data collected at each stage can feed back into the rest of the process.

Hadrian offers another example of improvement across many steps rather than one magic technology. Quoting a part, making it, inspecting it and moving it through a factory each present opportunities for redesign. A new factory or an acquired operation gives a team a chance to ask whether long-established processes still make sense.

Love expects AI to be useful in physical industries, but cautions that it cannot yet turn a wrench. A software model alone is not a substitute for understanding equipment and the physical process being changed. The most effective improvements will depend on that practical knowledge.

Price-Wright contrasts high-volume production of a single item with low-volume, high-mix manufacturing, where frequent changes make conventional automation costly. She says new software tools can help technicians who know the process reconfigure a line without necessarily having a software engineering degree.

As an illustration, she contrasts a traditional integration project taking months and hundreds of thousands of dollars with a manufacturing engineer using Claude to reconfigure a line in minutes. The point is not that every job can be compressed that way, but that flexible tools can change the economics of switching production.

Faster, adjustable production also changes how systems companies prototype. If a design can be iterated on quickly and its components produced on flexible lines, the journey from prototype to production can be shorter. Love connects this to designing for manufacture from the outset, a discipline associated with Tesla and SpaceX.

The guests point to new facilities being built by defense and maritime companies as evidence that the discussion is moving beyond designs on paper. They expect these sites to produce substantial numbers of systems, although the production volumes they discuss remain ambitions rather than completed outcomes.

Burger asks what distinguishes founders tackling these large problems. Love describes defense procurement as particularly difficult because the user, the buyer and the people appropriating money are separated. In the older model, contractors often waited for a funded government specification before developing a product.

Love argues that private capital can allow founders to respond to a perceived need before a formal government order exists. His hope is that this produces systems aligned with current technology and battlefield requirements sooner. He cites a hypersonic missile startup that spent years building with venture funding before receiving government contracts.

Price-Wright stresses the personal difficulty behind such ambition. A mining founder may travel to remote places without reliable electricity or phone service, spend extended periods away from family, and negotiate with suppliers whose culture and pace differ from those of startups. The mission has to justify that demanding work.

The speakers distinguish between industries that are important and companies that can become enormously valuable. They say examples such as SpaceX, Anduril and Palantir have made it easier for ambitious founders to imagine building large businesses in defense, energy and other physical sectors.

Their argument combines commercial opportunity with public purpose. Investors still seek returns, Love says, but companies can build valuable businesses while working on energy, security and production capacity. The guests view the arrival of founders willing to pursue these hard categories as a meaningful change.

On visits to mines and factories, the investors see the physical scale and practical hazards behind the story. Love recalls advice from his military service that the physical world does not care about you: moving earth and operating large equipment demand care. Both guests also describe the enthusiasm of teams at the sites they visit.

Price-Wright likes to revisit a factory every few months. An almost empty warehouse can become a crowded operation with technicians, machines and a sudden need for more power. Watching the pace of that transformation gives investors information that is difficult to obtain from a presentation alone.

Burger asks whether making one component locally can accelerate the creation of others. Love says the effects can include cost and quality as well as time. Defense systems require motors, actuators, valves and switches that may also be useful to future robotics manufacturers.

The guests expect those overlapping components to support a broader robotics supply chain, though they emphasize that the ecosystem is still developing. Price-Wright says robotics companies have had reasons to seek domestic suppliers beyond geopolitics, including quality control and the speed of iteration.

She describes one company that previously ordered twice as many printed circuit boards as it needed because half failed quality tests. New suppliers using software and modern processes, she says, have improved the proportion that passes. The example illustrates why proximity and co-development between customers and suppliers can matter.

Love sees an opportunity in what he calls the manufacturing middle. New aircraft, vehicles and defense systems create demand for components, but established suppliers may continue producing designs that satisfy an existing specification rather than rethinking them. A component that technically works can still embody decades-old assumptions.

He expects some suppliers to become co-engineers with original equipment manufacturers, redesigning components alongside the overall system. New materials and new requirements can justify a fresh design. In his view, this differs from simply buying a manufacturing asset and cutting its costs before selling it again.

The electricity grid offers another example. Price-Wright describes a shift from older static and mechanical equipment toward power electronics and battery storage that are more flexible and software-driven. A change in architecture creates a chance to reconsider the supply chain and engineering choices behind it.

The guests argue that even if future demand is smaller than the most optimistic forecasts, improvements in cost, efficiency, manufacturability and serviceability can persist. Once customers see an improved component work, they may have little reason to return to the older design.

Burger asks about founders leaving companies such as Anduril, Palantir and SpaceX. Price-Wright says people who have struggled with a production or supply-chain problem can recognize that it affects others too. They may leave to solve it more broadly, bringing practical knowledge and relationships with potential first customers.

Those relationships help create an ecosystem: early customers may be willing to tolerate an imperfect product because they know the founders and want a difficult problem solved. Love adds that prominent hardware companies took many years to reach their current scale; the apparent overnight success was a long process.

The guests emphasize focus and responsibility. A team building a difficult physical system may need to concentrate on one product instead of spreading itself across many. Once it delivers, it can reuse manufacturing knowledge, supplier relationships and customer trust when developing the next product.

On capital markets, Love argues that the United States has a large pool of private money able to follow technological and production success. Venture capital, private equity, banks and credit providers are showing greater interest in physical industries, he says, which matters because factories and hardware production lines require substantial capital.

The speakers see an advantage in investors being able to fund a company before demand is fully obvious, while acknowledging the expense and long timelines involved. Price-Wright says greater availability of risk capital may help companies bridge the period in which they must invest in capacity before a major government contract arrives.

Burger asks about the gap between US and Chinese growth in power production. Price-Wright argues that the central question is not simply technological but whether Americans are willing to build projects and grow. Love adds that successful companies could attract more capital, but physical construction and supply-chain development take years.

Price-Wright says the effects of physical infrastructure are visible to the communities around it. A mine, power plant or data center changes a place directly, so residents need a stake in the economic benefits. She argues that companies will have to find ways to bring communities along, even when that is difficult in the short term.

Love points to companies partnering with small machine shops to grow their businesses and develop products rather than simply cutting staff. The guests close by saying that production can bring jobs and prosperity beyond established coastal technology centers, but that realizing those benefits will require time and cooperation with local communities.