Hidden Champions Between Premium and the AI Algorithm
If you read the U.S. business press, you see it every week. Restaurant chains, software vendors, mid-tier consulting firms. A particular kind of company is quietly disappearing, and the obituaries arrive on schedule. Howard Yu has the cleanest description of the mechanism in Coase vs. Claude. He calls it the death of the middle. Markets turn into a barbell, and mid-tier players vanish between premium and the new logic of AI algorithms.
What Yu does not write about is what happens when this same mechanism reaches a market that was built to hide it.
I am writing this for U.S. investors and operators looking at the German and Swiss mid-market. The structural erosion is the same as in the U.S. The visibility is not. German insolvency law, with instruments like Schutzschirmverfahren and Eigenverwaltung (pre-insolvency moratorium and self-administered restructuring under court protection), is engineered to preserve jobs, not to clear inviable business models. A German company with a broken model can hold its position two or three years longer than a comparable U.S. firm. Add political bailout packages and short-time work subsidies, and you get the result: market erosion is delayed, not removed. Adjustment arrives later, and it arrives harder.
That delay creates an investor-relevant gap. Companies that look stable on the surface are bleeding margin underneath. The headline insolvency wave the U.S. press is reporting in the U.S. mid-market never quite shows up in DACH the same way. The erosion is real. The signal is muted.
Machinery, specialty chemicals, plant service, B2B standard advisory. The German and Swiss mid-market sits exactly where the mechanism strikes. Trumpf, the German laser-systems and machine-tool maker, is a recognizable anchor for an entire class. Family-owned, broad product range, solid margins, decades-long customer relationships, service as the bridge, a position in the middle of the market that held for forty years.
These companies now compete with two pressures at once. On one side, highly specialized providers whose margin runs on identity, not scale. On the other, Asian competitors whose unit costs have fallen so far that even a tenfold increase in warranty exposure remains profitable. Process power without differentiation, which is what most of these middle players actually had, no longer pays.
In Switzerland the middle never functioned the way it did in Germany, because the home market was too small for it. Anyone selling industrial equipment had to choose: push into premium or scale globally. In Germany the choice stayed optional, because a broad domestic middle class and a job-protective insolvency regime allowed it to stay optional. What looked like strength now operates as drag.
It’s not about restaurants. It’s about position.
Premium and Operator Pole, Defined
The two ends of the barbell are not two versions of the same business. They are two different architectures.
Yu calls the second pole Algorithmic-Cheap. The phrase reads like a discount play, which is wrong. The mechanism is an operator architecture, where AI executes rather than coordinates, and where the value claim is delivered in outcomes, not hours. I will call it the operator pole, because that names what is actually happening.
Whoever builds the premium pole defends a position that cannot be reproduced. Whoever builds the operator pole controls scale that no one else can hold. Both architectures last for the next decade. But they require different capital, different organizational forms, and different owner identities. A company that confuses the two ends up building a third path, which collapses exactly where the middle is now disappearing.
The premium pole has two layers. Inside sits implicit IP. Knowledge that does not fit into patents because it is too specific and too procedural. Knowledge a firm has built over decades, knowledge that lives in heads, routines, and unspoken tolerances. Knowledge a competitor cannot reverse-engineer, because the result is visible but the path to it is not.
Outside sits brand. Not in the marketing sense, but as a filter that reduces uncertainty for the buyer. In markets where AI output becomes functionally interchangeable, the value of this filter rises. When a customer must choose among ten similar providers, they choose by recognizability. Recognizability is brand.
Implicit IP inside, brand outside. That is the architecture of the premium pole. It is not romantic. It is expensive. Building implicit IP requires time that a private equity holding period does not accommodate. A company that took thirty years to build workshop knowledge can lose it in two if the wrong CEO arrives. Family-owned businesses are structurally better positioned in this architecture than financial investors. The holding period is the architecture, not the strategy.
The other pole runs on a different logic entirely. The operator pole is not discount. It means unit costs have fallen so far that competition runs on scale, not quality. A company selling industrial equipment that delegates a substantial share of service orders to an AI operator centralized across ten sites cuts its unit costs in half. Whoever achieves this halving prices every middle competitor out of the market.
The vehicle is not the cheap product. It is software that executes rather than coordinates. Outcome responsibility sits with the provider, not the customer. I described the mechanism in my AI Service Operator piece. For the barbell question, only the consequence matters. Whoever wants to win the operator pole takes on skin in the game. Selling tools means no liability for application. Selling outcomes means liability for delivery. That changes margin model, insurance structure, balance sheet logic, and owner accountability.
Margin in both architectures comes from position, not from doing. I made that point in February in Value No Longer Comes From Doing.
The premium pole takes decades to build, the operator pole can be scaled in five to seven years if the architecture works. Premium is robust against business cycles and exposed to cultural shifts. The operator pole is robust against business cycles and exposed to technological shifts. Whoever builds premium is the keeper of a story. Whoever builds the operator pole is the architect of a machine. Both are legitimate owner roles. They rarely combine in one person.
Premium requires patience and identity discipline. The operator pole requires capital and architecture competence. Patience and capital can coexist. Identity discipline and architecture competence rarely coexist in one owner.
The third path, riding the middle out with a bit more efficiency, a bit more brand, and a bit more service, does not work. Yu shows it in restaurants. The February 2026 SaaS repricing showed it in software. The next proof will arrive in German machinery balance sheets in 2027 or 2028.
Hermann Simon’s Quiet Falsification
Hermann Simon discovered a class whose foundation is now disappearing. That is not polemic. It is the sober consequence of his own assumptions.
What Simon called Hidden Champions are mid-sized firms, often family-led, mostly invisible to the public, that achieve world market leadership in narrowly defined niches. Thirty to sixty percent global market share, specialization rather than diversification, direct sales, low R&D ratios with high innovation rates, long employee retention. The description was analytically precise. It gave a generation of German and Swiss owner-operators an identity through which to read their own position in the market.
What Simon never made explicit is the precondition under which this mechanism works. The precondition: there are mid-sized, stable markets with sufficient volume in which specialization, sustained over decades, leads to inimitable depth. The markets are segmented enough that a provider with two- or three-digit headcounts can hold thirty or fifty percent of a global segment. Process power carries weight without strong brand or operator architecture. Thirty years of workshop knowledge translates into competitive advantage.
That precondition held until roughly 2020. It has been failing since.
Three mechanisms dissolve it, each sharp on its own, structurally lethal in combination.
Order shrinkage from above. What Choudary calls the atomic unit of work in Reshuffle does not only apply to fashion or software. Whoever previously sold a complete module, a service package, or a configured installation now increasingly sells single components, single maintenance outcomes, single specification steps. This is not loss of volume but loss of bundling. A business built on bundled orders with three-year lead times loses margin precisely where customers begin to buy in smaller units, because their AI-supported procurement now allows it.
Customer erosion from the side. Hidden Champions sell mostly to mid-sized industrial buyers who themselves sit in the middle of their own markets. When those buyers are split apart by the same barbell logic, the Hidden Champion customer base shrinks from both directions. The premium customer migrates to the highly specialized boutique offering more identity. The operator customer migrates to the Asian direct supplier with AI-driven quality control. The middle as customer disappears at the same time as the middle as provider.
Margin pressure from below. Operator providers push unit costs in service verticals so low that the typical Hidden Champion premium margin no longer holds. A machinery firm with service tradition charges twelve to eighteen percent on standard maintenance because it pays for personnel, tools, travel time, and tradition. An AI operator centralized across ten sites delivers the same maintenance at five percent margin and remains profitable, because its unit costs sit below the Hidden Champion’s. What is happening is not price pressure. What is happening is structural margin erosion in the field many Hidden Champions built as their bread-and-butter layer.
Together, the three mechanisms interlock. Simon described a class optimized for stable mid-markets. The mid-markets are no longer stable.
Germany shows the shift more slowly than the U.S. The home market was broad enough to absorb the early waves. Industrial culture rewards tradition more than efficiency, which slows adjustment. In Switzerland the Hidden Champion layer was thinner from the start because the home market never carried it, and successful Swiss machinery owners always had to choose earlier between premium and scale.
Simon’s recent writing on platforms and digital transformation is not wrong, but it does not address the deeper problem. The problem is not that Hidden Champions need to reposition for the platform economy. The problem is that the market structure that made Hidden Champions possible in the first place is dissolving. The class is not under-modernized. It is mis-positioned.
That diagnosis demands a different answer.
Continuing on the Hidden Champion path means running in a market that is losing its foundation. Switching to the premium pole or the operator pole means leaving the Hidden Champion identity behind. Both are difficult, but only the second carries forward. The doctrine carried a class for forty years. It will not carry the next twenty.
Two objections come up often. First: Hidden Champions were never the middle, they were always specialists in niches. Nominally true, structurally wrong. A machinery niche worth two billion euros globally is mid-market, not premium boutique. Whoever dominates it is the largest middle provider, not a premium keeper. That position is exactly what is eroding. Second: Hidden Champions always had implicit IP, so they are automatically premium. Implicit IP alone does not make a premium pole. Premium needs brand as an external filter. Most Hidden Champions have implicit IP without brand. They lead B2B niches whose end customers do not know their names. The implicit IP layer holds in the core. It does not protect against unit-cost erosion in the service periphery.
What remains is a split of the class itself. One part will make the jump into premium. Another will tip into the operator pole. A third, probably the largest, stays stuck and will erode over ten years in slow margin loss, without a dramatic collapse, without realistic recovery.
Patek After Quartz, the Real Lesson
Whoever wants to make the premium jump should not start with a strategy paper. They should start with a vault holding two dozen watches from the 1970s.
The Swiss watch industry in 1970 had about 90,000 employees in over 1,500 firms. It was the industrial backbone of the Jura. By 1985, the industry had lost almost two-thirds of its jobs. More than a thousand firms closed. The trigger was Japanese quartz technology, which entered mass production in the early 1970s and became the dominant price class within a decade. Mechanical movement turned into a remainder. The mid-class providers disappeared first. The middle of the market vanished long before anyone had a name for it.
The survivors were a small group of houses. Patek Philippe, Rolex, Audemars Piguet, Vacheron Constantin, a handful of others. They did the opposite of what the industry’s majority tried. They did not lower prices, they raised them. They did not automate, they kept the workshop principle. They did not open up to quartz, they sharpened mechanical identity.
In economic language, these houses defended two layers in parallel. Inside, implicit IP, in workshop tradition, apprenticeship, and unspoken tolerances. Outside, brand as an economic filter, actively rebuilt during the 1970s. Patek shifted its advertising to the generational argument that everyone now knows: “You never actually own a Patek Philippe. You merely look after it for the next generation.” The slogan was created in 1996. The logic behind it was built twenty years earlier. Identity as the product, not mechanics as the product.
The premium jump is possible even when the structural crisis is acute. It works not through better products but through a shift in the value claim. It costs time, which feels uncomfortable in the middle of the crisis, but it produces identity that holds through the next one.
So far the encouraging picture. The honest lesson arrives only when the story is told to its end.
Patek saved the premium pole. The Swiss watch industry saved the premium pole. But the other pole was built by someone else. Apple now sells more watches on wrists than the entire Swiss industry combined. The operator pole of the watch market emerged in Cupertino, not in the Jura. Whoever holds premium holds the premium pole. The other pole goes to someone with a different architecture, often someone who was not in the industry before.
That is the more important lesson for U.S. investors thinking about DACH machinery and specialty chemicals. The owner who succeeds at the premium jump owns the premium pole. The owner who wants the operator pole has to build it. If they do not, it goes to a competitor not yet visible in the industry. Probably from the U.S. or China.
For the mid-market, the implication is direct. The third-generation machinery owner with workshop tradition, the specialty chemicals operator with decades of process knowledge, the equipment builder with multi-generational service relationships, all of them have the implicit IP layer. What most of them do not have is the brand layer in the strict economic sense. That is the operational gap where the premium jump typically fails.
The pragmatic path runs through owner visibility, standard-setting in industry associations and norm-setting bodies, reputation in investor and customer circles, content with substance. It is the operational translation of what Patek built over twenty years, applied to a B2B context.
What Patek did not do is equally instructive. They did not try to recapture the middle by gently lowering prices. They did not enter the quartz market to stay current. They did not diversify their way through. Each of these moves would have been defensible in 1975. Each would have so diluted the identity that the premium position thirty years later would not have existed.
Premium is a discipline that, in the middle of a crisis, requires the opposite of what feels short-term reasonable. Whoever lowers the premium level during the crisis exits the crisis as a middle provider, in a market where the middle no longer exists.
The risk in the strategy is obvious. Holding premium loses near-term volume. Lower volume means less cash for brand investment. A narrow path between starvation and identity preservation. Several Swiss watch houses held that path. Others took the seemingly reasonable middle road and achieved neither depth nor height. They sold to larger groups, often below book value. The Swiss watch consolidation of the late 1980s, which produced the Swatch Group and Richemont, is the direct consequence.
The Operator Pole, the Path No One Wants to Choose
To owner-operators, the operator pole sounds like a loss of pride. That reaction is understandable but wrong. The operator pole is not a price position. It is an architecture.
Anyone who cannot or will not build premium still has a viable path, provided architecture competence and capital are present. The path is not “become a discount provider.” It is rebuilding a business on an operator architecture, where AI executes rather than coordinates. Unit costs fall, because human and machine have a different division of labor. Margin holds or rises, because the model bills for outcome, not hours. The owner shifts from keeper of a story to architect of a machine.
A service business rebuilt on the operator approach gains three things. A margin base independent of unit-volume pressure, because unit costs are lower. Scaling capability across sites that a traditional workshop model cannot match. A data layer that grows more valuable with every job, eventually becoming an asset no competitor can replicate without the same architecture.
In DACH practice, three obstacles cannot be ignored.
Codetermination. Restructuring two hundred employees through an operator architecture means engaging works councils, the IG Metall industrial union, and the supervisory board at a depth unimaginable in the U.S. It is not a deal-breaker, but it adds two to three years to the rollout and consumes negotiation capital the owner needs for the strategy itself. Switzerland has lower codetermination thresholds, which makes the path relatively faster, but its informal severance conventions are stricter, which reduces the practical room.
Patriarchal structures. Many DACH mid-sized firms have been shaped over decades by an owner-operator who sits deeply in operations. An operator architecture moves the operational center from the owner into a machine the owner no longer controls in detail. That is not a tool change, it is an identity change. As I argued in an April piece on AI not making companies more valuable, AI adoption without organizational redesign fails on exactly this identity question. A third-generation managing director walking through the production hall in the morning knows what runs in every shift. An owner running an operator architecture reads dashboards. Both are legitimate owner practices. The second one is, for most, an underestimated psychological break.
Fiduciary duty in the German GmbH. A managing director restructuring a business model carries personal liability for proper care. An operator architecture producing losses in its ramp-up phase can create liability questions that make a non-owner managing director cautious. In family-run firms this is muted. In holding structures with employed managing directors, it is a real brake.
These obstacles do not mean the operator pole is impossible in DACH. They mean naive transfer of U.S. or Chinese models fails. Haier is often cited as a model. Its Chinese owner Zhang Ruimin reorganized 80,000 employees into 4,000 self-managing micro-enterprises with their own profit and loss responsibility. Under Haier, GE Appliances tenfolded its U.S. market share in two years. The model works in a legal framework, capital market structure, and labor culture that DACH does not have. Anyone trying to copy Haier in Pforzheim or St. Gallen fails within twelve months.
Three transition patterns are realistic for DACH mid-market.
Deliberate shrinkage. The owner reduces the business to the core where operator architecture clearly carries, and exits or closes the periphery. Two hundred and fifty employees become one hundred and twenty, margin rises, complexity falls, the owner becomes the architect of a smaller, sharper business. The risk is that shrinkage gets read as defeat, which it is not.
Modularization without dissolution. The owner partitions the operational chain into modules, some automated, some run as an operator service layer, some preserved as a boutique premium layer. The holding stays intact, the modules run internally like micro-enterprises with their own profit and loss, but without formal dissolution. A cautious approach to Haier without breaking the legal framework.
Platforming through investment. The owner becomes a holding investor, founding or acquiring small operator businesses in adjacent verticals while keeping the legacy business in its current position. This shifts the owner identity gradually from workshop master to platform architect, without rebuilding the legacy business. The path with the lowest internal friction. It demands investor competence which is rarely present in owner families.
All three patterns are expensive. All three require an owner decision that must be strategic rather than operational. All three need at least three years before the architecture carries. Faster, and the move fails on codetermination, patriarchy, or fiduciary duty. Slower, and the market overtakes.
In Switzerland, smaller family firms have historically modularized earlier, because the home market was too small for an integrated business. Modularization discipline is culturally embedded. In Germany, the operator path will, over the next five years, be forced more by investors than chosen by owners.
The Diagnostic Lens
An owner who wants to know where they stand on the barbell does not need a consulting firm. They need six observations, readable on their own balance sheet, in their own order book, in their own personnel structure.
Margin in the service layer. If a B2B business has historically achieved twelve to eighteen percent margin on standard service, and that margin has fallen by two to four percentage points over the last two years without a visible competitor, that is a signal. Margin erosion without identifiable cause is typically structural pressure from operator providers below the radar. Recognizing it early creates a two-year head start.
Order size. If average order size has been falling for two years while order count is stable or rising, that is the Choudary mechanism on the firm’s own balance sheet. Customers are partitioning bundles into atomic units, because their procurement, AI-supported, now allows it. Setup costs hold, revenue per order falls. An owner not seeing it because the top line stays flat is missing the margin erosion.
Talent retention. When new hires negotiate for optionality, asking for flexible hours, shorter notice periods, external side activities, instead of long-term commitment and security, that is an early signal. Talent reads market structure faster than owners do. Candidates keep options open because they read the segment as unstable. That is not declining loyalty. It is market analysis from the candidate’s perspective.
Customer concentration in the middle. If the top twenty customers sit disproportionately in the middle layer of their own markets, and have been stagnating or shrinking slightly over two years, that is demand erosion at the second derivative. The customer base shrinks without any single visible loss.
Investment policy of competitors. When traditional competitors have invested in AI software over the last eighteen months without visible product improvements, that is most likely tool adoption without architectural redesign. Those competitors will stagnate over the next two years. When a competitor builds an operator architecture, often visible as a new service subsidiary with its own pricing model, that is the warning signal worth taking seriously.
Own brand investment. If the firm has not made a deliberate brand investment in five years, no owner visibility, no industry standard-setting, no reputation maintenance beyond standard advertising, then the brand layer for the premium jump is not in place. That means the premium path is closed without significant additional investment. The operator pole remains a viable option. So does staying in the middle, with the corresponding risk.
Six observations, each readable from the owner’s own desk, together a picture more sober than any consulting analysis. Whoever sees five or six of these symptoms simultaneously sits in a Hidden Champion segment that will erode over the next five years. Whoever sees three or four has time for an orderly adaptation. Whoever sees none should test the diagnosis with someone looking from outside, because absence of symptoms often correlates with industry self-deception.
In Switzerland, mid-sized firms have tended to choose between premium and scale earlier, because the home market never carried a middle. That has produced clearer business identities, which is an advantage in the current shift. In Germany, the picture is more diffuse. A broader middle class allowed staying in the middle as a legitimate strategy. That breadth is now collapsing.
Who Switches, Who Stays Stuck
The empirical record of the next five years will confirm or refute the structural diagnosis. From today’s view, three DACH verticals serve as observation fields.
Specialty machinery in southern Germany and eastern Switzerland. Family-owned firms with one hundred fifty to four hundred employees, multi-generational workshop tradition, international customer base, market leadership in narrowly defined equipment segments. Since around 2024, the response has been bimodal. Houses that have invested in brand building and premium positioning hold margin and gain ground with demanding customers. Houses that have invested in operator architecture and service consolidation gain volume with unit-cost-sensitive customers. Houses that continue the traditional model with marginally more efficiency lose share in both directions, without any single visible crisis. The first acquisitions of that third group by the first two are observable since 2025, often below book value.
Mid-sized IT services in DACH. The shift moves faster and more visibly here, because customers themselves sit in the software sector and adapt their procurement earlier. Traditional managed services providers offering patching, monitoring, and standard support face direct pressure from operator models delivering the same outcome at substantially lower unit costs. The first successful pivots are visible. They share one trait: the owner personality has actively moved the architecture question to the center, rather than delegating it to IT leadership. Those who do not see ARR base erode faster than the sales team can replace it.
Mid-sized advisory services, tax, audit, and IT consulting in the narrow sense, excluding the large firms. Margin erosion bites most directly here, because the billable hour is the atomic unit of work and AI replaces it most visibly. Practices with outcome pricing rather than hourly billing, paired with AI-supported standard processes, hold margin. Practices continuing the classical model with somewhat more software lose two or three percentage points of margin per engagement per year. Consolidation will move fastest here.
A common pattern emerges across the three verticals. Successful pivots have an owner with active strategic mandate, not a hired managing director with operational mandate. Business identity does not delegate. Successful pivots take at least three years before the new architecture carries. Faster attempts collapse mid-way.
The honest estimate is that two of three DACH mid-market firms will not make the pole switch, either because owner identity cannot be released, or because capital is missing, or because fiduciary duty does not allow the necessary courage.
Three scenarios apply to that third.
Slow margin erosion over five to ten years, without dramatic collapse. The firm shrinks gradually, the owner runs it to the end, succession either does not happen or happens at a value well below earlier expectations. The most inconspicuous form of failure.
Sale to an investor who carries out the pole switch in place of the owner. Often the most rational option commercially, but it collides with the owner tradition of many family businesses. Private equity investors with standard holding periods of three to five years are structurally unsuited, because the pole switch takes longer than their patience holds. Investors with permanent capital and longer horizons are positioned differently here. Acquisitions will happen anyway, often below book value, by both investor types, depending on who sees first.
Acquisition by a successful pivot competitor, meaning intra-industry consolidation. The most likely form of cleanup over the next decade. Fastest in advisory, slowest in machinery.
There is no hidden fourth option in which staying in the middle becomes profitable again because the market settles. The barbell is not a pendulum. It is a structural shift, pushed outward by Death of Normal demand and the shrinking atomic unit of work.
Structure Architect or Workshop Master
There is an owner decision that is rarely as explicit as it deserves. It sits behind every choice between premium pole and operator pole, behind every architecture, behind every brand investment, behind every codetermination negotiation. It is the decision between Structure Architect and Workshop Master.
Workshop Master is the owner identity that carried the DACH mid-market ideal for forty years. The owner who walks the floor in the morning, who knows every shift, who personally addresses every major customer, who carries the implicit IP in their own person. Direct, physical, day-near access. Responsibility because they see. Decisions because they understand. The embodiment of a tradition that is fluent in family-business language.
Structure Architect is the owner identity that survives the barbell. They design the architecture in which the business runs without operating it day to day. They do not see every shift, they read dashboards. They do not know every customer personally, they define the customer class. They decide less often, but with greater consequence. They sit closer to investor than to master.
Both identities are anchored in the DACH tradition. Both carry honor. But they lead to different architectures, and they rarely combine in one person.
An owner building the premium pole can stay a Workshop Master, with one extension. They must additionally become a brand architect, deliberately cultivating their own identity’s visibility. The Patek lesson. Workshop inside, brand outside. Extension, not switch.
An owner building the operator pole must become a Structure Architect. Here it is a switch. Workshop identity does not carry in an operator architecture, because the operational layer is no longer visible to the owner. Anyone unwilling to accept this builds the architecture half-heartedly, which is worse than not building it at all.
Anyone staying in the middle has chosen the Workshop Master identity without the brand extension, in a market that no longer pays for that position. That is not weakness. It is the consequence of a decision that, in most cases, was never made consciously, because the market position used to be stable.
The decision is no longer optional. Whoever does not make it makes it implicitly, in favor of the third position, which does not carry.
For forty years, owners did not have to ask the identity question, because the market did not ask it. Now the market does, and the answer is an owner decision, not a consultant’s recommendation. That is the reason this article does not contain a course of action. A course of action would assume the question is operational. It is not.
Whoever wants to know at the end of this article what to do has not read it carefully. Whoever wants to know who they want to be has understood the point.
It’s not about the middle. It’s about identity.
And there is no middle anymore.
