Industry already has enormous capability. The problem is that much of it remains invisible.
Imagine a manufacturer trying to produce a specialised component. The engineering team knows exactly what it needs. The specifications are clear. The component is technically feasible.
Then reality intervenes. The company's own equipment cannot produce it economically. One supplier can make part of it, but not all of it. Another company has the right machine but does not advertise this capability. A third has the technology but no current capacity. Somewhere else, there may be a company that has produced something remarkably similar — but nobody in the original company knows it exists.
Eventually, the manufacturer finds someone through a personal connection. Or pays too much. Or redesigns the component. Or invests in equipment it will barely use. Or simply gives up.
This is not fundamentally a problem of insufficient industrial capability. It is often a problem of industrial capability that cannot be seen, understood and connected.
Which leads to a larger question:
What if we could build a living map of what industry can actually do — and use it to design better cooperation between enterprises?
That is the idea behind a new generation of industrial cooperation platforms.
From companies to capabilities
We traditionally describe an industrial enterprise by its products and industry. A company manufactures hydraulic cylinders. Another makes gears. Another repairs heavy machinery. Another performs heat treatment. Another has a large CNC machining centre. Another specialises in difficult welding. Another has highly experienced engineers who know how to manufacture something few others can.
But these descriptions hide something important.
A company is not only what it sells.
It is also what it can do.
Its capabilities may include:
- particular manufacturing processes;
- specialised equipment;
- engineering knowledge;
- materials expertise;
- production technologies;
- qualified personnel;
- unusual tolerances or dimensions;
- certifications;
- testing capabilities;
- repair expertise;
- available production capacity;
- experience with particular products or industries.
Much of this capability never appears clearly in a conventional company directory. And that creates an information problem.
A manufacturer looking for a supplier usually asks:
“Who makes this product?”
A more powerful question is:
“What capabilities are required to make this product, and where do those capabilities exist?”
That small change in the question changes the entire search.
Industry is a system
Industrial production has always been a system of dependencies, not a collection of isolated enterprises. A factory depends on other factories. A product depends on components. Components depend on materials. Materials depend on processing. Processing depends on specialised equipment, technology and engineering knowledge.
One enterprise can therefore become a critical node in a much larger production system. Its economic value does not come only from what it produces internally — it can come precisely from doing one thing extremely well and being connected to the enterprises that need that capability.
This is not a new insight. Different industrial traditions have recognised it in different ways, and each offers a useful lesson for what we are building.
The Soviet experience: interdependency as a discipline
The Soviet industrial system placed considerable emphasis on understanding relationships between industries and enterprises. Material balances attempted to represent how different sectors depended upon one another. Industrial specialisation was taken seriously: one enterprise might concentrate on a particular component or operation while depending on many others for the rest of its production system.
Whatever one thinks about the broader economic system, there is a significant engineering insight here:
Industrial capability has value not only inside an enterprise, but through its relationships with other capabilities.
This becomes especially visible when cooperation breaks down. After the collapse of the Soviet Union, many established production relationships were disrupted because enterprises that had previously been connected within a common industrial system suddenly found themselves separated by new borders, institutions, ownership structures and economic conditions.
The lesson is not to recreate central planning. It is almost the opposite.
The relationships were real. The problem was how they were coordinated.
A modern industrial cooperation platform can approach the same underlying reality differently. Instead of asking what should the state allocate to whom?, it asks:
- What does industry need?
- What capabilities are required?
- Where do those capabilities exist?
- Who has available capacity?
- Which enterprises could work together?
- What would the resulting production configuration cost?
- Where are the gaps?
And most importantly: could the required production system be assembled from capabilities distributed across several enterprises?
We no longer have to choose between isolated market actors and centralised industrial planning. We can create visibility and coordination without centralising ownership or decision-making.
China: capability density through ecosystems
In many regions of China, enterprises producing related products and components have developed in close proximity to one another, supported by specialised suppliers, infrastructure, engineering services, logistics, skills and institutions.
The important feature is not simply that many factories exist in the same place. It is that they form an ecosystem. One company produces a component. Another provides a specialised process. Another produces tooling. Another supplies materials. Another provides testing. Another provides logistics. Knowledge moves between companies. Workers move between companies. Customers create demand. Suppliers specialise. New companies emerge around gaps in the network.
This suggests another important principle:
Industrial strength is not only a function of individual companies. It is also a function of the connections between them.
Germany and Europe: specialisation within networks
Many highly specialised European manufacturers do not attempt to become vertically integrated giants. They concentrate on what they do exceptionally well — a particular machining technology, precision components, industrial automation, materials, testing, or a narrow class of machines.
Their competitiveness comes partly from specialisation and partly from being connected to customers, suppliers, engineering institutions, research organisations and other specialised companies.
This produces an interesting paradox.
The more specialised a company becomes, the more important its network can become. Specialisation creates efficiency. Networks make specialisation economically viable. Without connections, specialisation can become isolation. With connections, specialisation becomes a source of industrial strength.
America: the beginning of the digital manufacturing network
Modern digital manufacturing platforms have demonstrated another part of the evolution. Companies such as Xometry have shown that manufacturing capacity can be represented digitally and connected to demand at scale.
This proves something fundamental: manufacturing capability itself can become a digital object.
But there is still another step.
Much of today's digital manufacturing marketplace is oriented around a transaction: here is the part, find someone who can make it. The larger opportunity is to ask a different question:
What is the best way to organise the production of this part — or this entire product — using the capabilities available across the industrial network?
That moves from supplier search toward production-system design.
UNIDO: supplier development, not just supplier discovery
UNIDO's long-standing work on subcontracting and partnership exchanges combines enterprise profiling, supplier matchmaking, benchmarking and supplier development. The underlying idea is particularly relevant: industrial cooperation is not only about finding an existing supplier; it can also be about developing the capabilities required for stronger industrial relationships.
A platform should not only tell us that a capability does not exist. It should eventually help answer: who could develop this capability?
From finding a supplier to designing a production system
Consider a large specialised component. It may require material preparation, cutting, machining, welding, heat treatment, surface treatment, precision finishing, inspection and assembly.
There may be no single company that does all of these things economically. But perhaps there are five companies that collectively can.
The question then changes. We are no longer searching for a supplier. We are designing a production configuration. That configuration might involve two companies. Or five. Or twenty.
The platform's job is not simply to produce a list of suppliers.
It is to help answer: how could this industrial result actually be produced?
A different representation of industry
To do this, we need to describe industry in a more meaningful way.
Not simply:
Company → Product → Industry
but something closer to:
Enterprise → Capability → Operation → Resource → Capacity → Product → Relationship
A company might possess several capabilities. A capability might depend on particular equipment, materials, skills and technologies. A production operation might require several capabilities. A product might require hundreds of operations. A project might require capabilities that no single enterprise possesses.
Once these relationships are represented, an industrial network begins to emerge. And this is where the idea of ontology becomes practical.
Ontology is not about technology. It is about meaning.
The word "ontology" can sound academic. In this context, the idea is simple. Before building sophisticated software, we need to establish what things exist in the industrial world and how they relate.
What is an enterprise? What is a capability? What is capacity? What is a production operation? What is a component? What is a requirement? What is a resource?
What does it mean for a company to possess a capability? What evidence demonstrates that capability? What makes one capability suitable for a particular requirement?
These distinctions matter.
Saying "Company X does machining" is almost meaningless for serious industrial cooperation.
What kind of machining? On what equipment? For what materials? What dimensions? What tolerances? What quantities? What certifications? What quality level? What capacity? What lead time? With what demonstrated experience?
Once those questions are represented systematically, the company stops being a line in a directory. It becomes a node in an industrial capability network.
AI becomes interesting when it understands this structure
Artificial intelligence is often presented as a conversational interface. But industrial cooperation requires something deeper. An AI system needs to understand the structure of the industrial problem.
Suppose an engineer says: "We need this component manufactured at 2,000 units per year. The material is X. The critical tolerance is Y. It must survive conditions Z."
An industrial AI system could reason through the requirement. It could identify the necessary operations. It could determine the capabilities required. It could search the industrial network. It could identify enterprises with those capabilities. It could examine capacity, geography, certifications, experience and constraints. It could identify combinations of companies. It could reveal missing capabilities. It could even identify that the cheapest solution is not a single supplier but a cooperation configuration involving several specialised enterprises.
The AI does not replace the engineer. It gives the engineer something that is otherwise extremely difficult: a way to reason about an industrial system larger than any individual can see.
The platform works in both directions
The platform should not only help companies that need something. It should also help companies that can do something.
A manufacturer may have an expensive machine that is underutilised. Unused production capacity. Specialised engineering knowledge. A process few others possess. A highly experienced workforce. A certification that opens particular markets. Equipment suitable for an entirely different industry.
Today, much of this capability may remain invisible outside the company's existing network. A capability network can turn it into an opportunity. The question for an enterprise becomes: what can we do that somebody else needs?
That is potentially transformative.
Missing suppliers may mean missing capabilities
Suppose a company cannot find a suitable supplier. The obvious conclusion is: there are not enough suppliers. But that may not be the real problem.
Perhaps the required capability exists, but is poorly connected to the market. Perhaps several companies each possess part of it. Perhaps existing companies could develop the missing capability with modest investment. Perhaps a new piece of equipment would unlock it. Perhaps the product itself could be redesigned around capabilities that already exist. Perhaps a new industrial company should be created.
The platform can reveal something much larger than supplier shortages. It can reveal industrial capability gaps. And a capability gap is not only a problem. It can be an investment opportunity.
From cooperation to industrial development
Once capability data becomes sufficiently rich, another layer becomes possible. Imagine being able to see, for a particular region, what manufacturing capabilities exist, where capacity is available, which capabilities are concentrated or scarce, which technologies are aging, where specialist skills exist, which supply chains depend on single enterprises, which capabilities are missing, where logistics create constraints, and what new demand is emerging.
This becomes a living map of industrial capability.
Governments and development institutions could use such information to understand where industrial development is actually needed. Investors could identify capability gaps. Large manufacturers could identify potential partners. Smaller manufacturers could discover new markets.
The difference from traditional cluster analysis is that modern digital infrastructure can make these relationships increasingly visible and dynamic.
Clusters should emerge from reality
We should not begin by declaring: "This region should become a machinery cluster." We should first ask: what capabilities are already here? Then: how are they connected? Then: what demand exists? Then: what capabilities are missing?
Only after that should we talk about clusters. A real industrial cluster is not a geographical label. It is a network of complementary capabilities. And those networks can sometimes cross geographical boundaries. A production system may involve companies in several cities or regions because the required capabilities are distributed.
Trust will be as important as discovery
Industrial cooperation cannot be built simply by publishing a list of machines. Companies have legitimate reasons to protect information. They may not want competitors to know their customers or publish exact capacity. They may have proprietary processes or strategic relationships.
The system must distinguish between what can be public and what should remain private. A company might publicly disclose: "We provide precision large-format machining." A qualified participant might see equipment, dimensions, tolerances, materials and certifications. A trusted partner might see available capacity and lead time. Commercially sensitive information can remain confidential.
The objective is not to expose everything. It is to expose enough of the right information to make cooperation possible.
The real product is not the database
A database of factories is useful. A supplier marketplace is useful. A procurement platform is useful. But the larger vision is different.
The platform should help answer four questions:
What do we need?
Translate an industrial problem into a structured set of requirements.
What capabilities are required?
Understand the operations and capabilities necessary to produce the result.
Where do those capabilities exist?
Discover enterprises, resources, capacity and expertise across the network.
What could we build together?
Design viable cooperation configurations.
That last question is where the real opportunity lies.
A new kind of industrial infrastructure
We already have systems for managing individual companies. ERP manages enterprise resources. MES manages production. PLM manages products. CRM manages customers. Procurement systems manage purchasing. Supply-chain systems manage flows.
But the industrial ecosystem between companies is still surprisingly fragmented. Much of it lives in personal relationships, spreadsheets, trade associations, messaging applications, procurement departments, old supplier lists, engineers' memories, exhibitions and telephone calls.
In other words: a huge amount of industrial knowledge exists, but it is not organised as a system.
That is the gap we are interested in.
The Industrial Capability Network
Our vision is to build an infrastructure that represents industrial capability as a connected network. At its foundation are enterprises. But enterprises are connected through capabilities, operations, resources, products, requirements and projects.
A manufacturer does not simply appear in the system as "Machine-building company." It appears as a structured set of capabilities. A project does not simply appear as "Need component X." It becomes: required outcome → product → components → operations → required capabilities → candidate enterprises → production configuration.
This creates a fundamentally different way of navigating industry. Instead of searching through companies, we search through industrial possibility.
The first application should be real industrial problems
We do not believe such a system should begin with a giant database containing every factory in a country. The better starting point is much more practical.
Take a real industrial problem. Something that a manufacturer genuinely needs but cannot currently solve efficiently. Represent the problem. Understand its requirements. Map the necessary capabilities. Find potential participants. Test the cooperation. Measure the result. Then do it again.
Every successful project adds knowledge to the network. Every failed project teaches something. Every cooperation creates evidence. Every new capability becomes another node.
Over time, the system develops something that ordinary supplier databases do not have: industrial memory. Not simply knowing who claims to be able to manufacture something, but knowing which combinations of capabilities have actually worked.
From supplier directory to industrial memory
Imagine that, several years from now, an engineer asks: "Who can manufacture this?" The system can answer. But imagine it can go further:
- Three enterprises can manufacture the complete component independently.
- A lower-cost configuration uses four specialised enterprises.
- Two of those enterprises have successfully cooperated before.
- The heat-treatment capability is currently capacity-constrained.
- There is another qualified capability 180 kilometres away.
- A similar configuration was successfully used on three previous projects.
That is no longer a directory.
It is a memory of industrial cooperation.
A different model of industrial competitiveness
For much of industrial history, competitiveness has been associated with the strength of the individual enterprise. That remains important. But increasingly, another question matters: how effectively can an enterprise participate in a network of specialised capabilities?
A company does not necessarily need to own every capability. It needs to know what it is exceptionally good at, what it should obtain from others, which partners it can trust, where additional capacity exists, how to configure production, and how to respond when one part of the network fails.
The future industrial enterprise may therefore be simultaneously: a factory and a node in a network.
Cooperation does not eliminate competition
This is not a proposal to eliminate competition. Quite the opposite. Companies compete for customers, on quality, on cost, on technology, on speed and reliability.
But they can also cooperate where cooperation creates a stronger production system. A company can compete with another company for one project and cooperate with it on another. This is already common in industrial ecosystems. The opportunity is to make such cooperation easier to discover and organise.
The deeper idea
The deeper idea is not a marketplace. It is not a directory. It is not another procurement application. It is an attempt to create a digital representation of industrial capability and cooperation.
Once that representation exists, many applications become possible: supplier discovery, capacity utilisation, industrial procurement, production configuration, supplier development, investment analysis, industrial cluster development, regional planning, supply-chain resilience, capability-gap analysis, new product development, industrial project formation.
The same underlying industrial model can support all of them.
Why now?
The industrial world is becoming simultaneously more specialised and more interconnected. Products are becoming more complex. Supply chains are becoming more vulnerable to disruption. Companies need access to capabilities they do not own. Specialised manufacturers need new markets. Regions need stronger industrial ecosystems. Governments want greater visibility into domestic production capability.
And AI is making it possible to reason over large bodies of structured information in ways that were previously impractical.
But AI alone is not the answer. An AI system cannot understand an industrial ecosystem properly if the underlying industrial reality is poorly represented. The quality of the intelligence depends on the quality of the model of reality.
That is why we believe the foundation must be industrial knowledge and ontology.
The opportunity
The industrial world does not necessarily need more factories. It may first need a better way to understand the factories it already has.
There are enormous numbers of machines, engineers, technologies, workshops, production lines, repair facilities, specialised manufacturers and accumulated skills already distributed throughout the economy. Much of this capability is difficult to discover. Difficult to compare. Difficult to connect. And much of it remains trapped inside individual companies and personal networks.
The opportunity is to make that capability visible. Then connected. Then usable. Then increasingly intelligent.
We are not trying to centralise industry
The goal is not to create one organisation that controls industrial production. It is to make distributed industrial capability easier to discover and coordinate.
The factories remain independent. The companies remain independent. The decisions remain with the enterprises. The platform provides something different: shared visibility into industrial possibility. It becomes infrastructure for cooperation rather than an institution that dictates cooperation.
From industrial map to industrial network
The first stage is visibility. What exists?
The second is connection. Who can work with whom?
The third is configuration. How can several capabilities be combined?
The fourth is execution. Can the cooperation actually produce the required result?
The fifth is learning. What did we learn from the cooperation?
And then the network becomes progressively more intelligent.
This creates a potentially powerful cycle: Capability → Connection → Cooperation → Evidence → Knowledge → Better Cooperation. The network learns from reality.
A new question for industry
For decades, industrial managers have asked: who is our supplier? We believe another question is becoming equally important: who could become our partner?
And beyond that: what could we produce together that none of us could produce as efficiently alone?
That is the question that sits at the heart of the Industrial Cooperation Platform.
The beginning
We do not know exactly what the final industrial network will look like. And that is intentional. It should emerge from real industrial problems.
We want to start with manufacturers. With engineers. With difficult components. With production constraints. With companies that know what they can do — and companies that know what they need.
We want to take real industrial problems and see whether a better representation of capabilities can produce better cooperation. If it works once, we learn. If it works ten times, we see patterns. If it works hundreds of times, we begin to see an industrial network.
And eventually, perhaps, something much larger becomes possible: a living digital model of industrial capability — not describing industry as it appears on paper, but representing what it can actually do.
The industrial capability nobody could see
The most interesting industrial capability may not be a new machine.
It may already exist. It may be sitting in a factory that has unused capacity. It may belong to a company nobody outside its region knows. It may be an engineer's accumulated experience. It may be a specialised process developed decades ago. It may be a group of companies that could solve a problem together but have never met.
The capability is real. The problem is that the network cannot see it.
The machines are already there. The engineers are already there. The factories are already there. The capabilities are already there.
The question is: how much more could industry accomplish if those capabilities were easier to see — and easier to connect?
That is the infrastructure we want to build.