Meta-production

A new way to think about the future of manufacturing

There’s no shortage of material arguing for the reindustrialisation of Western economies. Whether it’s about supply chain resilience, achieving net zero, defence preparedness, or simply reversing the hollowing-out of manufacturing jobs, the consensus is clear: we need to build things again. And as the cost of intelligence goes to zero with AI, the return on physical industries is only likely to go up.

Most of the approaches that are promoted in this context focus on using software to enhance efficiency, streamline operations, and optimise existing production systems. A recent(ish) article by Palantir’s CTO about ‘software defined production’ captures this sentiment:

"It allows managers to regain control of sprawling bureaucracies. It helps workers understand how to act amid complexity. It connects strategy with operations. It acts as a digital aid to human agency, allowing them to execute and win."

There are many fantastic businesses being built to do just this and overlay software onto existing manufacturing models, making current systems more efficient, more flexible, and more data-driven. They are important, but it’s not enough.

Because China is also doing this—and they’re doing it better.

Catching up is a losing game

It’s tempting to believe that if we just automate more of our factories and use software to optimise processes, we can catch up with China. But this ignores well known realities:

• China is already years ahead in manufacturing efficiency.

• It’s not just about cheap labour anymore—it’s about scale, supply chain dominance, and raw material control.

• Even if we optimise our factories to the nth degree, we are simply iterating on a model they’ve already perfected.

Even a fully automated Tesla Gigafactory in Germany cannot outcompete China’s vertically integrated end-to-end electric vehicle supply chains, where materials, batteries, and assembly all happen within an ecosystem of hyper-efficient, government-backed megafactories.

Some have made the argument that the ‘cost physics’ of manufacturing have changed, which makes a shift back to western (more specifically US-led) manufacturing dominance possible. The argument goes that four key factors—labour, capital expenditure, raw materials, and energy—define the "cost physics" of manufacturing, and technological advancements are rapidly reshaping these elements. With intelligence, automation, and affordable energy converging, the economics of production are shifting, making American manufacturing increasingly competitive.

While the US leads in cutting-edge AI, China now dominates AI-related patent filings and is increasingly competitive (see DeepSeek for details). Its manufacturing investment also vastly outpaces the US, with over $2.3 trillion in fixed-asset investment in 2023, compared to America’s record $236bn in manufacturing construction. And China holds a near-monopoly on key raw materials—producing 70% of the world’s rare earths, refining over 85% of them, and controlling much of the global supply chain for lithium and cobalt, while the US remains heavily import-dependent. Even in energy, where US manufacturers benefit from cheap shale gas and industrial electricity rates of $0.08/kWh, China’s government stabilises costs through coal subsidies, keeping industrial rates similarly low at ~$0.09/kWh.

Taken together, it doesn’t feel like the US (let alone Europe, which scores less well on all these metrics) has sufficiently low cost physics for meaningful catch-up.

But if cost efficiency is a losing battle, intelligence and adaptability might be our edge. Unlike traditional manufacturing, which scales through volume and supply chain dominance, what i’m calling ‘meta-production’ scales through recursion—where each innovation generates new tools that fuel the next. The US and Europe still lead in advanced materials and biotech—domains that could underpin a radically new production paradigm. Rather than competing on cost, we could compete on the ability to create entirely new classes of factories, materials, and products.

Leapfrogging instead

We happily offshored manufacturing capacity to China in the last few decades and let them not just catch-up, but dominate (see chart).

Given this lead, and the argument that a strategy based on catching up isn’t likely to work, the US and Europe might need to think more in terms of leapfrogging not just by improving production, but by reimagining it entirely.

Meta-production is a system that doesn’t just make products but also recursively generates new tools, processes, and capabilities. It means redefining what a factory is, how it works, and what it produces.

Aaron Slodov set out a useful framework for thinking about manufacturing in his Techno-Industrialist Manifesto, which I’m going to shamelessly borrow with a minor tweak.

1. Capacity → Capacity refers to the fundamental layer of processing power in terms of efficiency, throughput, and raw materials. It addresses questions like: How many factories and machines do you have? What are their capabilities? What do they process or produce?

2. Execution → Execution refers to the layer above capacity, focusing on the tacit knowledge and processes required to operate and produce goods. This includes skilled trade knowledge in manufacturing, operations, engineering, and materials.

3. Product → Product refers to the most abstracted layer—what goods are actually produced. This encompasses everything from semiconductors and submarines to cars, clothes, food, and drugs.

The version of the future that we need to build isn’t just making these elements tech-enabled or tech-improved, but fundamentally rethought. So let’s break down capacity (the raw ability to produce) and execution (the methods by which production happens) in a hypothetical neo-factory world.

This is purposefully a little out there .. it’s not a prediction, feasibility study or necessarily remotely viable .. but might be a useful way to think about what leapfrogging might look like, if it is a necessary ambition.

Impossible products

If we had factories that can reconfigure themselves, materials that can assemble on demand, and AI that designs every product in real-time, what kinds of things could we create? The possibilities extend far beyond mere efficiency—this is about creating entirely new categories of products that challenge what we consider possible.

1. Evolving cars

• Vehicles that physically change over time, optimising weight, aerodynamics, and durability.

• Self-healing materials mean no more rust or structural damage.

• Electromagnetic shape-memory alloys could allow cars to adjust their structure dynamically, making them safer and more adaptable in different driving conditions.

2. Self assembling aerospace

• Aircraft that reconfigure their aerodynamics mid-flight for fuel efficiency.

• Self-assembling drones that build themselves from modular components.

• Dynamic fuselage structures that expand or contract based on passenger load or cargo needs.

• Solar-powered atmospheric aircraft that can stay airborne for months at a time without refuelling, using regenerative energy storage.

• Smart wings inspired by bird flight, where the material shifts to optimise lift and drag on demand.

3. Adapting consumer goods

• Clothing that dynamically adjusts for climate and body shape.

• Furniture that reshapes itself, adjusting height, form, and function based on usage.

• Sports gear that enhances performance in real-time, adjusting grip, stiffness, or cushioning based on the activity.

Some of this sounds a little too sci-fi, but not all of it is that out of reach. For instance, Neri Oxman’s lab sits between synthetic biology, architecture and design, with projects related to 3d printed glass, guiding silkworms to construct domes, and generative design. You can see a documentary about Oxman’s lab on Netflix, watch the video below, or listen to her (fantastic) interview with Lex Friedman here.

Clone Robotics released a video of their Protoclone embodiment—a bio-inspired bipedal, musculoskeletal ‘humanoid’, which is pretty dystopian, but a signal of where things might be going in terms of what we might need to manufacture in future.

And Machina Labs have developed mobile factories:

Where does this leave us?

The future of manufacturing isn’t just efficiency or optimisation—though that of course should and will remain an important objective for many years to come—it’s rethinking production and creating things we’ve never even imagined before. Meta-production may then become more than a futuristic vision, but a necessary shift if we want to lead rather than follow in the next industrial era.

I’m hoping to see lots more European startups exploring this space—get in touch if you are!

Thanks for reading!