The Next Industrial Revolution Starts with Digital Humans and Better Fake Food

Jun 07, 2026

11 min read

2 views

68%

What if the bottleneck is not intelligence, but embodiment?

A strange pattern is emerging across two very different markets. On one side, people are spending billions to build software that can think, talk, and eventually act like us. On the other side, shoppers across Europe are buying more food that looks, cooks, and tastes like animal products, but is made from plants. At first glance, these trends have nothing to do with each other. One belongs to frontier AI labs and compute clusters. The other belongs to grocery aisles and dinner tables.

But together they point to a deeper shift: the future is not just about replacing things, it is about recreating their useful behavior at scale. We are learning that once you can reproduce the function of something convincingly enough, the original can lose its monopoly. That insight is about to reshape labor, food, manufacturing, and maybe even what we mean by a product.

The old industrial logic was simple. If you wanted more steel, you built more mills. If you wanted more milk, you raised more cows. If you wanted more expertise, you educated more experts. The new logic is different. If you can simulate the behavior of a worker, a product, or a biological process well enough, you may not need the original system in the same way at all.

The real race is not human versus machine, but biology versus replicability

The idea of digital humans sounds like a science fiction stunt until you notice how much of our economy already depends on repeatable patterns of human judgment. Customer support, tutoring, medical triage, design iteration, sales outreach, legal drafting, and scientific literature review all rely on humans performing cognitively structured tasks. If those tasks can be emulated by systems that are cheaper, faster, and more scalable, the advantage of flesh and blood becomes less absolute.

That does not mean humans become obsolete. It means the economy starts to reward replicable competence rather than scarce embodiment. The best model is not a human replacement in the narrow sense. It is a system that captures enough of the useful pattern of a human capability to make the expensive original less necessary for many jobs.

Plant-based food offers a surprisingly concrete parallel. No one buys plant-based milk because it is an entirely different category of experience. They buy it because it performs the job of milk in coffee, cereal, baking, and daily routines. Its value lies in being a functional substitute, not a philosophical statement. The same is becoming true in other parts of the food system, where cheese, seafood, and meat analogues are increasingly judged by one criterion: can they do the job well enough for the context in which people actually use them?

This matters because markets do not always reward the most authentic thing. They reward the thing that best solves the problem. When a substitute crosses the threshold from novelty to habit, the economics change fast. That is why a plant-based milk category can become a multi-billion-euro market, and why a digital human could someday absorb tasks that once required a whole profession.

The decisive question is not, “Is this the real thing?” It is, “Does this reproduce the function people were buying in the first place?”

Substitution is a ladder, not a switch

Many people think substitution happens all at once. In reality, it follows a ladder of legitimacy. First, the alternative is cheaper. Then it is acceptable. Then it is convenient. Then it becomes preferred in certain contexts. Eventually, the original becomes a special case rather than the default.

Plant-based milk is the clearest example. It did not win because it was identical to dairy. It won because it improved the user experience for a subset of buyers. It lasts longer, works for some dietary needs, and fits a different set of values. Once enough people discovered that milk is not one thing but many use cases, the category expanded rapidly. A cappuccino foam test, a cereal test, a cooking test, a lactose intolerance test. Each one opened a different door.

Digital humans will likely follow the same ladder. They will not begin by replacing the most complex, high-trust, high-stakes roles. They will start in contexts where the goal is not perfect identity, but reliable interaction. A support agent that handles routine queries. A tutor that explains algebra step by step. A sales assistant that remembers preferences and follows up instantly. A synthetic research assistant that reads a thousand papers and surfaces patterns faster than any person can.

The error is to imagine a single moment when humanity “hands over” control. The more realistic story is a thousand small substitutions. Each one seems limited. Together they rewrite the structure of work.

This is why the claim that digital humans might be built on large clusters of advanced chips in less than a decade is not just a technical prediction. It is a statement about how fast substitution can compound once the first acceptable version exists. The first version does not need to be magical. It only needs to be useful in enough places to start moving the curve.

Why food and intelligence are converging on the same economic law

The link between plant-based food growth and digital humans is not metaphorical fluff. Both are governed by the same market law: decoupling from origin.

For centuries, milk came from cows because the cow was the only practical source of milk. Human expertise came from humans because cognition came from humans. That sounds tautological, but it is really a historical limitation. Once alternative systems can generate sufficiently similar outputs, the origin stops mattering as much as the output.

Plant-based milk demonstrates a key principle of modern markets: consumers do not always want a source, they want an outcome. They want creaminess in coffee. They want protein in breakfast. They want function, consistency, and price. If a product can deliver those with fewer drawbacks, it can scale quickly even if it violates old intuitions about what the category “should” be.

Digital humans may do the same for knowledge work. Managers do not always want a person in the abstract. They want outcomes: response speed, accuracy, tone, availability, and memory. If a synthetic agent can deliver those outcomes more consistently than a person in a narrow use case, it becomes economically rational to deploy it.

This creates a powerful framework: the substitution threshold. A new technology crosses the threshold when it is good enough on the dimensions that users actually pay for, not on the dimensions enthusiasts care about. People do not pay for cows, they pay for milk behavior. People do not pay for a human face, they pay for trustworthy task execution.

The public debate often gets stuck on identity. Is it real milk? Is it a real person? That framing misses the commercial logic. Markets are less sentimental than that. They care about friction, reliability, cost, and scale.

The hidden bottleneck is trust, not technology

A useful product is not automatically a widely adopted product. The missing ingredient is often trust architecture.

Plant-based foods have grown because they fit existing shopping habits. A carton of oat milk sits where milk sits. A vegan sausage goes on the same grill as a regular one. Adoption rises when the substitute reduces the effort required to change behavior. The consumer does not need to reorganize their life. They simply swap one item for another.

Digital humans will face the same constraint, but at a higher level of complexity. People will not hand over meaningful interactions to a system they do not trust. The interface will matter. The disclosure will matter. The ability to escalate to a human will matter. The memory model will matter. So will the social signals that tell users, “This is safe enough for this task.”

In other words, the next breakthrough may not be only in model capability. It may be in trust design: how systems present identity, uncertainty, and accountability. A digital human that speaks fluently but behaves opaquely will fail. A digital human that is slightly less impressive but predictable, transparent, and context aware may win.

This is the same reason some plant-based products grow faster than others. The best products are not always the ones with the most impressive engineering. They are the ones that reduce perceived risk. If a customer can try oat milk in coffee without fear that breakfast will be ruined, the barrier falls. If a business can use an AI agent for one narrow workflow with clear audit trails, the barrier falls.

Adoption is not just a function of performance. It is a function of how safely people can experiment.

A new mental model: the mirror economy

We need a better way to think about what happens when substitutes get good enough. Call it the mirror economy.

In a mirror economy, the most valuable innovations are not necessarily new objects. They are systems that reflect the behavior of existing objects closely enough to inherit demand. Plant-based foods mirror animal products in taste, texture, and use case. Digital humans mirror human labor in conversation, memory, and decision support. The power comes from reducing dependence on the original source while preserving the customer-facing experience.

This model reveals why so many debates about authenticity miss the point. Authenticity matters emotionally, but usefulness matters economically. A mirrored product can be “less real” and still become more important if it is cheaper, cleaner, faster, or more scalable. That is not a bug. It is the mechanism of progress.

The mirror economy also explains why early stages can look trivial while the long-term consequences are enormous. At first, a digital human may only handle basic tasks. A plant-based cheese may only melt adequately in some settings. Critics dismiss both as inferior copies. But once mirrors become good enough for daily use, they change the baseline expectations of the market. The original no longer defines the category alone.

The deepest implication is this: industries do not get disrupted only when something better appears, but when something sufficiently similar appears without the old constraints. That is why scalability matters so much. If a digital human can be duplicated at near zero marginal cost, or a plant-based product can be produced without animal agriculture, the economics become hard to ignore.

What to do when the future becomes substitutable

The practical lesson is not that everything will be replaced. It is that many markets are entering a phase where function, not essence, becomes the unit of competition. That changes how companies, workers, and consumers should think.

For companies, the winning question is not, “How do we preserve the old category?” It is, “Which user jobs can be delivered differently, with fewer constraints?” For workers, the question is not simply whether AI can do your job, but which parts of your job are actually a bundle of separable tasks, some of which can be mirrored. For consumers, the question is whether loyalty should belong to the source or to the experience.

This opens room for a strategy that is more nuanced than either panic or hype:

Map the function, not the form. Ask what problem your product or role really solves. Milk is not just milk, it is convenience, nutrition, and behavior in context. Many jobs are not titles, they are bundles of repeatable outputs.
Look for the substitution threshold. Identify the point at which an alternative becomes good enough for real use, even if it is not perfect.
Design trust before scale. The best substitute will fail if people cannot safely adopt it. Interfaces, disclosure, and accountability are part of the product.
Expect category erosion, not instant replacement. Most markets will not disappear. They will split into premium originals, functional substitutes, and hybrid systems.
Invest where replication compounds. The most powerful businesses will be the ones whose improvements can be copied, distributed, and adapted at low marginal cost.

The common thread is that value increasingly comes from solving a job repeatedly and reliably, not from preserving the old material basis of the job.

Key Takeaways

Start with the function. Whether in food or labor, ask what users actually need the product or person to do.
Watch for acceptable substitutes. The biggest market shifts happen when alternatives become “good enough” in everyday use.
Trust is the adoption gate. Capability matters, but so do transparency, safety, and clear boundaries.
Expect gradual replacement, not a switch. Most disruption happens through many small substitutions that compound over time.
Think in mirror economies. The future will reward systems that reproduce outcomes while shedding the constraints of the original source.

The future belongs to the best copies that free us from the original

It is tempting to read both trends as stories about imitation. That is too small. A better way to see them is as stories about liberation from scarcity. We are building systems that can mimic the useful parts of intelligence and biology without requiring the full cost of either. That is why these markets are growing now, and why they may accelerate much faster than our institutions expect.

The real upheaval is not that digital humans will look like us or plant-based foods will taste like the originals. It is that entire industries are beginning to compete on whether they can deliver the effect without the old dependency. Once that becomes the standard, the future will not be organized around what something is made of. It will be organized around what it can do, at scale, without asking permission from the old world.