Peter Thiel once wrote, “We wanted flying cars; instead we got 140 characters.”
It is a line that has stuck around because it captures something real about the last few decades. Technology has reshaped how we communicate, consume information and interact with one another, but much of the physical world still looks surprisingly familiar. Commercial aircraft do not move materially faster than they did in the 1970s. Major infrastructure projects take decades. New power plants, factories and transportation systems are difficult to finance, permit and build. Even when a genuinely new physical technology works, getting it into the hands of customers can take far longer than inventing it in the first place.
The common explanation is that we have stopped taking risks or run out of important ideas. I do not think either is true.
There are more people working on ambitious technical problems today than at almost any point in history. Founders are building new aircraft, robots, satellites, energy systems, medical devices, defense platforms, and entirely new categories of machines. Capitalism has not stopped producing invention. In many ways, it has become extremely good at funding and producing proof of concept.
We know how to assemble a talented group of engineers, give them capital and ask them to build something that has never existed before. We know how to finance research, prototypes, pilot programs, and initial demonstrations. A small, highly capable team can take an idea from a slide deck to a working physical system in a surprisingly short period of time.
The problem is that a working prototype is not the same thing as a product that can be produced reliably and economically.
A prototype can be built by the same engineers who designed it. Those engineers can make adjustments as they go, work around incomplete drawings, replace parts that do not fit, and solve problems in real time. Cost is often secondary because the purpose of the prototype is to prove that the underlying technology works.
The production version has to survive a very different environment. It needs to be built by people who were not involved in the original design. It needs to use parts sourced from a network of suppliers, arrive on a predictable schedule, meet quality requirements, and perform consistently in the field. It also needs to be made at a cost that allows the customer to buy it and the manufacturer to remain in business.
This is where we continue to struggle.
We have become very good at going from zero to one. We are much less effective at going from one to one thousand, particularly when the thing being built is large, complex, regulated, or dependent on a fragmented industrial supply chain.
That transition is often described as a scaling problem, but the word “scaling” makes it sound more straightforward than it is. In software, scaling often means distributing something that has already been built. Once the software exists, another customer can usually access it at very little additional cost.
Hardware does not work that way. Every additional product requires more material, machine time, labor, inspection, transportation, and working capital. Producing the thousandth unit is not simply a matter of reproducing the first. It requires a manufacturing system that can absorb variation, identify problems and improve over time.
That makes manufacturing a different problem set from invention: it is an engineering, economic, and operational problem.
You have to understand how a product is actually built, which processes create unnecessary cost, where defects originate, how long each operation should take, which suppliers can be trusted, and how production should be scheduled. You need to decide when automation makes sense and when a skilled operator is more effective. You need enough visibility into the factory to know whether a late order is caused by a machine, a material shortage, a quality issue, or a planning mistake.
Most importantly, you have to make the product repeatedly at a cost that makes sense.
This work is less glamorous than the first successful demonstration, but it determines whether an invention becomes part of the world or remains a promising prototype. A new technology does not change society because it worked once. It changes society when it can be produced in sufficient quantities, at a sufficient level of quality, for a price that people or institutions are willing to pay.
The technologies that defined previous industrial eras followed this pattern. Cars mattered because factories learned how to produce them at scale. Aircraft changed the world because an entire industrial base developed around their production, maintenance and operation. Computers moved from laboratories into homes and businesses because manufacturers learned how to make them smaller, cheaper and more reliable.
In each case, the breakthrough was important, but the manufacturing system around it was what allowed the breakthrough to spread.
A great deal of the future we were promised already exists in some form. It exists inside laboratories, test facilities, prototype shops, and the early production lines of deep-tech companies. The question is whether we can build the industrial capacity required to move those technologies out of controlled environments and into widespread use.
That is the problem we are building Ironstead to work on.
We believe the next generation of hardware companies should be able to focus more of their attention on the technology they are creating and less on reconstructing a manufacturing base around themselves. Today, many of these companies are forced to navigate a supply chain that is fragmented, difficult to understand, and often operating with limited software, limited production data, and limited capacity for rapid improvement.
A founder may have designed a genuinely important product and secured demand from customers, but still find that the parts required to build it are late, inconsistent or far more expensive than expected. The supplier may be capable of making the part, but lack the systems required to schedule it properly, capture what happened during production, or improve the process from one order to the next.
Ironstead is being built to address the manufacturing side of that equation. We are building and operating manufacturing capacity, developing software around real production environments, and using the resulting data to improve how advanced parts and products are made.
We are starting from inside factories because manufacturing is difficult to understand from the outside. The important information is often found in the gap between what the process is supposed to look like and what actually happens on the shop floor. It is found in setup times, inspection results, machine utilization, scrap, rework, material movement, and the decisions experienced operators make every day.
By operating in that environment, we can build tools that reflect how manufacturing actually works rather than how it appears in a planning document. Over time, that should allow us to help factories increase output, reduce cost, and provide customers with a more reliable path from design to production.
The broader goal is to make it easier for ambitious hardware companies to become enduring industrial businesses. The founders building new aircraft, defense systems, energy infrastructure and autonomous machines should not have to solve every part of the manufacturing problem independently.
They should be able to rely on an industrial system that can take what they have designed and build it repeatedly, reliably, and economically.
We have not run out of ideas, and we have not lost the ability to invent. What we have allowed to weaken is the machinery that turns invention into abundance. Rebuilding that machinery will require factories, skilled people, capital, software, and a much deeper understanding of production economics.
The future we were promised is still available to us, but getting there will require more than imagining it or proving that it can work. We also have to learn how to build it.
George Payne 