The most underpriced asset in the current AI cycle is not a model, a chip, or a data center. It is a portfolio company owned by a private equity sponsor in the European mid-market. The argument that follows is the longer version of why I believe that, and what KYI was built to do about it.
KYI is the AI engineering partner for European private equity. We deploy senior engineering teams inside portfolio companies and build AI infrastructure that expands EBITDA inside a single hold period. The thesis behind the work rests on a simple economic claim: coordination has been the dominant cost in business for two centuries, and that cost is now falling faster than at any point in modern history. The question is who captures the surplus.
Coordination as the constraint
Ronald Coase asked the right question in 1937. Why are there firms at all? If markets allocate resources efficiently, why do most people earn their living inside organizations rather than as independent contractors trading with each other? His answer was transaction cost: the cost of finding counterparties, negotiating terms, monitoring performance, enforcing contracts. When those costs are high, firms grow. When they fall, firms contract and markets expand.
Oliver Williamson sharpened the framing in the 1970s. The constraint was not transaction cost in the abstract. It was a particular flavor of it: the cost of coordinating activity under uncertainty, with information asymmetry, when the parties involved cannot fully specify in advance what they want from each other. That kind of coordination is expensive, hard to outsource, and historically only possible inside a firm.
For most of the post-Coase era, the variable that has moved is information cost. The internet flattened search. Cloud flattened infrastructure. APIs flattened integration. Each step expanded what markets could do efficiently and contracted what firms had to do internally. Software ate the parts of coordination that were legible enough to specify.
What it did not eat is the rest. The parts of coordination that require judgment, context, negotiation, and trust stayed inside the firm. The cost of doing them at scale stayed high. That is where the wage bill of the modern services economy lives. That is what rolls up into the operating cost line of every mid-market industrial, distribution, and services business in Europe.
The variable that is moving now
Cheap reasoning is finally moving the next layer. Not search, not infrastructure, not integration. Coordination itself. The cost of translating intent across systems, monitoring quality, catching mistakes, reconciling outputs, drafting and negotiating, mediating between humans and machines. The kinds of work that used to require a trained operator with judgment.
When coordination becomes cheap, the boundary of the firm moves again. The question for a private equity sponsor is which side of that boundary their portfolio companies end up on.
The answer is not symmetric. Some kinds of work get pulled deeper into the firm. The kinds that depend on tacit knowledge, customer trust, regulatory accountability, or physical presence become more valuable, not less, because the firm can do them better when coordination cost drops. The relationship-driven services business with cheap coordination is materially better than the same business with expensive coordination. Margin expands. Throughput per employee expands. Customer retention expands.
Other kinds of work leave the firm entirely. The kinds that depend on routine information processing get priced down to the marginal cost of the underlying intelligence. They become market services. Companies that sell those services as a product get compressed. Companies that consume them as an input get cheaper.
The interesting case is the third one. The kinds of work that used to be impossible inside a firm because coordination cost made them prohibitive: dynamic pricing across thousands of SKUs, route optimization with real-time demand, predictive maintenance across a heterogeneous asset base, customer-level profitability tied to operational decisions made hours earlier. These are now possible. They are not yet ubiquitous. The firms that build them first will compound on the firms that do not.
Why private equity portfolio companies
Let me make the asset class argument explicit. The European mid-market is the largest concentration of asset-heavy, recurring-revenue, undermanaged operations in the world. The major sponsors collectively own thousands of these businesses. Manufacturing, distribution, logistics, business services, food production, regulated services, industrial services. They share a few traits.
They have meaningful scale, typically between 50 million and 1 billion euros of revenue. They have recurring or semi-recurring demand. They serve customers who value continuity over novelty. They have rarely been instrumented at the operating layer. They have been managed by operators who knew the business in their heads. They have been bought and sold several times by sponsors who optimized governance, capital structure, and commercial discipline, but did not change how the work itself was done.
That last point is the one that matters. The standard private equity playbook over the last twenty years has been governance and capital, plus a layer of commercial transformation. Procurement, pricing discipline, salesforce effectiveness, working capital. All of it valuable. None of it touched the operating layer where coordination cost lives. The operating layer was assumed to be fixed.
It is not fixed anymore. Cheap reasoning makes the operating layer addressable for the first time. A sponsor that takes a mid-market distributor with a 6 percent EBITDA margin and re-engineers its operating layer using AI does not produce a 7 percent business. They produce a 10 to 14 percent business. That is not an incremental improvement. It is a different asset.
The coordination surplus
Where does the value go? This is the part that deserves more attention than it has had. As coordination becomes cheap, a surplus is released. The question of who captures it is decided by where the surplus shows up first.
The surplus does not show up at the model layer. The model layer is competitive and getting more competitive. Margins there are compressing toward the marginal cost of inference. The surplus does not show up at the application layer either, at least not durably. Most application companies sell coordination intelligence back to incumbents who could increasingly run it themselves. Surplus there is real but transient.
The surplus shows up at the operating layer of established businesses. It shows up as expanded EBITDA in companies that already have customers, already have regulatory presence, already have the physical asset base, already have the trust. It shows up as a structural margin uplift that did not exist a year ago. And it shows up in private hands first, because public companies cannot move fast enough to capture it without disclosure consequences.
Private equity is the natural collector of this surplus. The sponsor has the asset. The sponsor has the capital. The sponsor has the hold period. The sponsor has the operating function. What the sponsor does not, in general, have is the AI engineering capability to convert these conditions into actual operating change at the asset level.
What KYI does
That last gap is the one we built KYI to close. We are not a consulting firm. We are not a software vendor. We are not a fund. We are an engineering partner that integrates with the sponsor's existing operating function and delivers AI infrastructure inside portfolio companies on a measured, sequenced framework.
We build three things inside portfolio companies. Decision systems that turn data into pricing, forecasting, and operating decisions the company can act on. Operational automation that compresses cost across functions. Commercial intelligence that surfaces margin and growth opportunities invisible to traditional reporting. The infrastructure is built alongside management and the operating partner, with EBITDA captured over the hold period and the methodology replicating across the sponsor's portfolio.
The work is not glamorous. It is engineering inside operating businesses. It looks like instrumenting a packaging line. It looks like building a pricing engine for a 12,000 SKU distribution catalog. It looks like a triage model that routes customer exceptions to the right operator at the right time. It is the boring substrate on which the next decade of mid-market value creation in Europe will be built.
The window
Coordination cost will keep falling. The firms that move first will compound the advantage. The firms that wait will discover, as they underwrite their next vintage, that the operating layer of the comparable assets they bought five years ago has moved without them. Multiple expansion at exit will favor the assets that were re-engineered. Multiple compression will punish the ones that were not.
That is the case. Coordination is the variable that is moving. Private equity portfolio companies in the European mid-market are the asset class best positioned to capture the surplus. The capability gap is engineering. The window is now.
Frequently asked questions
What is the coordination economy?
The coordination economy is the framing in which coordination cost — the cost of negotiating, monitoring, reconciling, and translating intent across systems — is the dominant variable cost of running a firm. As cheap reasoning lowers that cost, the surplus released shows up first at the operating layer of established businesses that have customers, regulatory presence, and physical assets.
Why does the coordination surplus accrue to incumbents rather than to AI startups?
The model layer is competitive and compressing toward the marginal cost of inference. The application layer sells coordination intelligence back to incumbents who can increasingly run it themselves. The durable surplus accrues at the operating layer of established businesses, because that is where customer trust, regulatory accountability, and physical scale live. Those conditions are slow and expensive to build and they compound when coordination cost falls.
Why is European mid-market private equity well positioned to capture the coordination surplus?
European mid-market private equity holds the largest concentration of asset-heavy, recurring-revenue, undermanaged operations in the world. Sponsors have the asset, the capital, the hold period, and the operating function. What they do not, in general, have is the AI engineering capability to convert those conditions into operating change at the asset level. KYI Capital exists to close that gap.
What does KYI Capital build inside portfolio companies?
KYI builds three things inside portfolio companies: decision systems that turn data into pricing, forecasting, and operating decisions; operational automation that compresses cost across functions; and commercial intelligence that surfaces margin and growth opportunities invisible to traditional reporting. The infrastructure is built alongside management and the operating partner, with EBITDA captured over the hold period and the methodology replicating across the sponsor's portfolio.
How does the coordination economy change exit multiples?
Multiple expansion at exit will favor assets that were re-engineered to the new operating layer. Multiple compression will punish assets that were not. A mid-market distributor with a 6 percent EBITDA margin re-engineered with AI is not a 7 percent business — it is a 10 to 14 percent business. That is a different asset, and the next sponsor pays for it.