The Limit of Firm-Level Cooperation
The AI Powered Flexible Specialization for Manufacturing series demonstrated that AI-brokered coordination can turn a fractured regional manufacturing base into something that functions like a vertically integrated mega-factory. The key mechanism — a semantic matching engine that finds idle capacity, available expertise, and idle equipment across thousands of independent SMEs, without forcing any of them to expose sensitive intellectual property — is real, architecturally sound, and already partially prototyped.
But the unit of cooperation in every AI Flex Spec scenario was always the whole firm. Apex Milling engaged Tri-City Precision as organizations, authorized at the CEO or operations director level, transacting via commercial agreements. Veridian Systems assembled five companies. The Ontario Pocket bid was a firm-to-firm consortium, governed by smart contracts between corporate entities.
This matters because firms are not the smallest meaningful unit in manufacturing. Manufacturing is done by people, on machines, in specific spaces, using specific skills. And at that level — below the firm boundary — there is a class of coordination need that firm-to-firm commercial agreements cannot reach.
The Lumpy Asset Problem
To understand why sub-firm coordination matters, you have to understand a structural constraint that is so familiar to manufacturing people that it is rarely named out loud: productive capability in manufacturing comes in lumps.
You cannot buy 0.4 of a five-axis machining centre. You cannot lease 30% of a metrologist. Floor space is rented by the bay, certifications renewed at a fixed annual overhead regardless of how many programs they cover, and specialists hired as whole people — with full salaries, benefits, and employment relationships that carry real obligations whether their particular expertise is in demand that week or not.
The result is predictable. Manufacturing firms of every size will almost always have either too much of something or too little. The five-axis machine purchased to land a precision aerospace contract, sitting at 40% utilization after that contract’s volume tapered. The process engineer hired to stand up a new production line, fully occupied for nine months, now underutilized on routine work. The 400-square-metre annex leased at the peak of the last program cycle, half-empty since the program matured. The PLC programmer who spent two years automating a production cell, whose skills are now applied to maintaining the system he already built.
None of these situations reflect poor management. They are the structural consequence of lumpy assets meeting variable demand in organizations too small to absorb the mismatch internally. At the individual firm level, this waste is unavoidable.
At the level of the industry as a whole, however, the picture is different. Across Ontario’s manufacturing base — thousands of independent SMEs on thousands of independent demand cycles — the surplus in one firm is almost always the deficit in another, often within the same region, often in the same week.
The idle five-axis shift in Hamilton is the machine time a Kitchener robotics integrator needs for a six-week tooling job. The underutilized metrologist in Mississauga is the expertise a London medical device shop needs for a CMM qualification program. The PLC programmer with available Friday afternoons in Brantford is the resource a Guelph machinery builder needs to get through a controls retrofit without hiring a full-time controls engineer.
These surpluses and deficits exist simultaneously. What does not exist — yet — is the mechanism to connect them.
What Cooperative Specialization Enables
Cooperative Specialization is enabled by a Cooperative Specialization Support System (CSSS) — an AI-mediated cooperation marketplace built on the open Cosolvent protocol, extended to operate at the sub-firm level. A CSSS is not a new kind of company or a new legal structure. It deploys the same semantic matching, the same confidentiality architecture, and the same trust framework as the firm-level platforms described in the Ontario Roadmap — applied here to the assets and people that firms are made of.
A manufacturing company that participates in a Cooperative Specialization network can configure, precisely, which categories of resource are available for external engagement: which machines can offer idle shifts to the exchange, which roles can take on fractional external work, which domains of expertise their employees are authorized to share, and what approval process governs each category. The cooperation is person-to-person or machine-to-machine in execution; the governance remains at the firm level.
The cooperation that becomes possible operates at a scale the firm-level marketplace cannot reach:
A production engineer posts a technical query — a specific cutting parameter failure he has not seen before — through the CSSS, and is matched with a semi-retired process specialist who has solved the same problem three times. The CSSS presents an anonymized competence summary; the engineer selects, the disclosure sequence opens, and within twenty-four hours they are in a billable consultation. No purchase order. No consulting contract. A light-weight fractional engagement governed by the CSSS’s standard terms, authorized by both parties’ employers, completed in forty-eight hours.
A precision machining shop with two Makino five-axis centres running one shift instead of two authorizes its operations director to offer second-shift capacity to the network. A robotics integrator forty kilometres away, needing fifty hours of five-axis titanium machining for a tooling program, registers the requirement through the CSSS. The matching engine identifies the idle Makino as a high-confidence fit and initiates the structured disclosure protocol — neither party knows the other’s identity until both have opted in. The shop earns contribution against fixed overhead it was carrying anyway. The integrator gets precise, certified domestic capacity without a four-week search. Both employers authorized it. The CSSS logged every step.
A quality manager who spent three years achieving IATF 16949 automotive certification at a mid-size stamping plant is authorized by her employer to offer fractional guidance through the CSSS to other manufacturers navigating the same certification — two to four hours per week, at a fee the CSSS prices against comparable consulting rates, under a standard engagement framework that protects both employers’ interests.
Every one of these transactions is too small for a conventional sales agreement and too specific for a general labour market. They are exactly the transactions that a Cooperative Specialization platform is built to enable.
DeeperPoint is developing MarketForge — a sponsor-configurable application layer built on the open Cosolvent protocol — as the prototype deployment environment for testing these sub-firm coordination capabilities with real participating firms in a regional manufacturing network.
The Aggregate Effect
Individual transactions redeem individual mismatches. At scale, the effect is structural.
If a well-functioning Cooperative Specialization network operates across Ontario’s manufacturing base — even at modest adoption rates — the real-time picture of available manufacturing capability changes fundamentally. The network can see, for the first time, not just what firms own but what they are actually using. The delta between installed capacity and utilized capacity — the lump-asset mismatch waste — becomes visible, matchable, and correctable.
The metrologist who was hired for a qualification project and is now underutilized does not become a productivity loss; she becomes a resource that other firms can access on terms her employer has pre-approved. The five-axis machine running one shift does not carry overhead as pure waste; it earns contribution from organizations that need its capability and have no other way to access it at the right quantity, the right time, and the right price.
This is not utopian. It is the same economic logic that underlies every well-functioning marketplace: surplus and deficit that exist simultaneously, in the same geography, correctable if visible. The only thing Cooperative Specialization adds is the mechanism — semantic matching that works at the resolution of people and machines, not just firms.
The Stories That Follow
The scenarios in this series will make Cooperative Specialization concrete. Each one begins with a lump-asset mismatch — a surplus that one firm cannot use and a deficit that another cannot fill through conventional channels. Each shows what the platform makes possible, not just for the two parties in the transaction, but for the pattern of cooperation that accumulates across the network as matches become history, history becomes trust, and trust deepens the market.
The mismatch waste that each story resolves is not incidental. It is the structural problem that Cooperative Specialization was designed to correct. By the end of this series, the cumulative picture of those corrections — at the level of individual machines, individual people, individual problems resolved — will answer a question that the Ontario Pocket only began to address: what does Ontario’s manufacturing base actually look like when the cooperation reaches all the way down?
What makes a thin market tick? → · The MarketForge platform → · The Cosolvent open protocol →