EuroStack has published a new paper on how to unlock demand for European technology. It gets lots of things right but runs into limitations because it tries to coordinate replacement of software workloads in a world that runs on layering. This article analyses the strengths of EuroStack and suggests actual business strategies to become successful despite existing challenges.
What is EuroStack and what does the new paper propose?
EuroStack is an initiative of more than 600 mostly small and medium-sized European technology companies, led by competition economist Cristina Caffarra and supported by selected industrial players such as Airbus. Its core logic is demand-driven: European tech providers don't lack talent or innovation — they lack demand to scale.
The new paper tackles exactly this problem, framing it as a coordination challenge. Individual companies hesitate to move because they fear a “first-mover disadvantage”: whoever switches first bears the cost and risk of migration, while others can wait. To overcome this, the paper advocates for coordinated procurement through buyer alliances, inspired by models such as the Open Compute Project and the O-RAN Alliance. The idea is that aggregated demand can reduce individual risk, create scale, and make European alternatives commercially viable.
EuroStack’s proposals rest on three assumptions that show a limited understanding how enterprise technology organizations actually work.
A: Hyperscalers are not the real lock-in
The EuroStack paper mentions the word “hyperscaler” exactly 28 times and focusses on them as the original source of lock-in. The analysis is: “Proprietary standards and closed interfaces between providers make legacy systems expensive... Enterprise data stored in formats that make migration technically complex and costly." This is not wrong but totally misses the point: Cloud transformation in global enterprises has been remaining at 20–30% for years (check here on slide 58). The real dependency is within in the demand-side actors themselves. It sits in how every piece of tech is entangled with organizations, processes, and business models: Windows, macOS, Android, iOS, VMware, NetApp, Cisco, Nvidia chips, CUDA, Intel, VMware vSphere, IBM MQ, ServiceNow, Atlassian, Miro, Oracle, SAP HANA, MS SQL, Neo4j, MS Dynamics, Workday, Active Directory, Auth0, Splunk, Dynatrace, Tableau, Salesforce, Shopify, MuleSoft, Qlik plus >100 external SaaS per enterprise. Every company combines all of it in a unique fashion.
B: Demand is not really unhappy with hyperscalers
The EuroStack paper states that “most European businesses want to reduce their reliance on US hyperscalers.” In reality, most tech buyers complain about price hikes from vendors like Microsoft or VMware and dislike their weak negotiation position. But that’s just one of many “discomforts” management deals with: employees forming workers’ councils, engineers preferring Macs and Slack over PCs and Teams, security incidents, talent churn, growing shadow IT, and increasing regulatory pressure, classic data center providers taking ages to deliver, human bottlenecks going on parental leave, contract negotiations with SAP. Of course, companies complain about hyperscalers, but at least they deliver super-reliably. And if you ask companies, if they want to go back the pre-AWS era, you will see fear in their eyes. Cloud is the devil you know.
C: EuroStack forgot the relevance of opportunity costs.
Managers in European enterprises are no victims. They employ strong architects and hordes of engineers who regularly present alternatives: moving from Oracle to PostgreSQL, from VMware to OpenStack, from proprietary systems to more sovereign architectures. The options have always been there. But management consistently decides against them. Not because they are unaware or incapable, but because they have better uses for their best engineers: AI adoption, product innovation, regulatory compliance, and core business execution. The problem is something that especially tech-minded people forget: Engineering resources are scarce, you don`t easily task them to change infrastructure that works.
The problem is not hyperscalers (A), nor latent dissatisfaction waiting to be unlocked (B). And it is certainly not a lack of available alternatives (C). The real constraint sits in the specific, messy setup of each company: its processes, integrations, accumulated code, and the layers of systems built over decades, including every technology that came before the cloud and sits on top of it. Even if Europe had the perfect hyperscaler —20% cheaper, fully sovereign—it would not trigger mass migration. Because the real cost is migration: risk, effort, and, most importantly, the opportunity cost of tying up scarce resources.
If we want to succeed, we need to play with innovation cycles and total costs
So, the situation is even messier than EuroStack suggests—and European demand is, in a way, comfortable where it is, because the alternative looks even messier. Shall we then capitulate? Of course not.
But before jumping to conclusions, we need to establish two basic truths that any viable solution must comply with or it will fail. These truths are consistently ignored: many engineers don’t see them because they are inside the wave, not looking at the tide, and economists don’t see them because they don’t fight the daily battles in the corporate tech jungle.
1: Technology evolves in cycles
Innovation comes in waves. New entrants challenge incumbents (Creative Destruction, Joseph Schumpeter, 1942), incumbents defend their models (Innovator’s Dilemma, Clayton Christensen, 1997), and adoption follows distinct patterns (Innovation Adoption Curve, Geoffrey Moore, 1991). But each wave brings its own infrastructure, operating model, and economics. You don’t replace old tech, you rebuild around the new one. Remember when trains replaced the Pony Express, they required rails and did not use the same horse change stations. Please, everyone listen to Benedict Evans.
2: Total cost reduction is what counts
Companies make these change decisions based on their own constraints: limited resources, acceptable risk exposure, and opportunity costs. They evaluate trade-offs pragmatically, typically within short time horizons of 3 months to 3 years. Demand moves when managers conclude that the business case justifies the changes in infrastructure, operating model, and economics including migration and opportunity costs. That is why companies still have paper and email while putting WhatsApp and Slack on top because every new technology itself made sense, but migrating from one to another didn't.
Cloud migration isn’t component procurement
The result: hardly any replacement, lots of accumulation. New technologies don’t wipe out the old; they layer on top. Mainframes, COBOL, client-server, Windows - they’re all still there. Even major shifts follow this pattern: mobile reduced the dominance of desktops without eliminating them; cloud challenged classic virtualization but did not remove VMware. Remember Benedict Evans? Cloud adoption is at ~20-30%.
In this light, also the buyer consortia EuroStack points to (e.g. Open Compute Project or O-RAN) make sense: They worked precisely because their upside was clear and individually attractive and outweighed change effort and risk. And it worked, because the buyer consortium could unify their demand around a clearly defined and interchangeable component. But such a thing doesn`t exist in a world of software-defined whatsoevers.
Hyperscaler compete on workloads. Every workload is a single fight that needs to be fought, and where a migration must make sense for the company that pays for it. Life of everyone in corporate IT would be so much easier if one could procure system migrations like Tioga Pass OCP Sled Servers.
EuroStack needs a business strategy, not macro arguments
If you accept that thinking cycles is key, and that customers move with TCO, then a viable strategy follows naturally.
There are four entry points for new technology vendors in companies:
1: New Capability (early cycle)
This is the innovation phase. A small group of experts within the organization adopts new technology before the business case is fully understood. This is what we see today with AI tools like Cursor, Claude, or n8n. The key is accessibility: adoption must be free to try because financial controllers don’t fund “crazy ideas” of experts. They just don’t get what all these new acronyms are about. That’s why Claude subsidized a $200 plan with compute ressources worth $27.000.
2: Capability Consolidation (mid cycle)
Once a capability becomes essential, organizations move to standardize it. What started as scattered tools—Claude, Cursor, ChatGPT, Copilot—gets consolidated. Now even finance accepts that these tools are necessary. Budgets get formalized, and vendor reduction becomes a priority. At this stage, bundles often win. Companies will accept an inferior tool like Microsoft Teams over Zoom & Slack if it reduces complexity and supplier management.
3: Cost Reduction (late cycle)
In mature phases, the focus shifts to efficiency. Companies systematically review cost structures and look for savings. This creates opportunities for cheaper alternatives—but only if they win on total cost, not just component price. This includes migration costs, operational risk, and opportunity costs. Because these constraints do not exactly incentivize major changes, organizations like Mastercard or Lufthansa still run legacy systems such as mainframes and COBOL is still a thing.
4: External Constraints (late cycle, irregular)
Sometimes change is forced by external shocks: Holy datenschutz, lack of gas supply, or rare earths. These can trigger movement but still change is only driven at minimal necessary effort. Companies do this only, when required. They did neither change for women’s equality nor for avoiding climate change. They will not do it for European sovereignty.
If you look at the 4 entry points from a business strategy perspective, you will see that EuroStack's macro arguments are an awesome opportunity to create Attention. The geopolitical fear, the sovereignty narrative, the scary headlines — they get you in the room. But AIDA doesn't stop at Attention. To move through Interest, Desire and Action, you need to switch language. From macro arguments to business cases. From collective good to individual incentive. EuroStack has mastered getting the meeting. What happens next determines whether anything actually moves.
I suggest two strategies aligned with how the B2B market works
The four entry points suggest exactly two strategies. Entry points 1 and 2 — where new capabilities are adopted and consolidated — call for creating new demand. Entry points 3 and 4 — where cost pressure and external shocks dominate — call for competing on total cost. Not sovereignty arguments. Business cases.
I. New Capability Bundle
EuroStack has ~600 signatories—mostly small, fast-moving companies—with access to policymakers and private capital. Use that leverage to create new demand, not coordinate old demand. Instead of pushing late adopters like banks into buyer consortia without clear upside, build something they actually want: a new capability. Right now, that’s agentic workflows. There is Celonis with enterprise reach, n8n as a core enabler, and a broad base of European consultancies. Bundle this into a concrete offer and let EuroStack act as the access layer—one entry point, one go-to-market. Get corporates into this capability and you control the next layer, where productivity gains and budgets sit.
II. Total Cost Reduction Bundles
Hyperscalers aren’t cheap at the component level: Hetzner outperforms many US clouds on price, Oracle databases are far more expensive than open source. Everyone knows it. Yet corporates seldom move, because total cost wins: switching is risky, expensive, and resource-intensive. So compete exactly there. Build full migration bundles, not components: “We untangle your Oracle stack including stored procedures with AI, move it to PostgreSQL, run it on STACKIT—one contract, one bill, minimal disruption, ROI after 18 months.” No abstract macro arguments, just a better business case. Once that works, optionality opens. European bundles will become sticky and the STACKIT ecosystem can pull more business into its ecosystem.
I’d bet an expensive bottle of good old European champagne: Our conversations with demand representatives would completely change if it started thinking in business offers instead of sovereignty arguments, simply because we would speak their language: Capability, cost, risk, and individual incentives.
Don´t get me wrong – I love EuroStack
EuroStack gets many things right. Europe needs to move into build mode. There is no lack of talent, and what US Big Tech has built is not magic. It is entirely reasonable to aim for a larger share of the value creation to remain in Europe. The idea of buyer cooperation is not wrong either.
But if you look at how enterprise IT actually works - how procurement decisions are made, how migration projects unfold, and how industrial organizations deal with risk and complexity—it is worth questioning whether the problem is really a “first-mover disadvantage,” and whether buyer alliances can realistically trigger a self-sustaining demand cycle.
EuroStack’s real assets are attention, network, and timing, something that the Greeks called Kairos. Combined, these provide rare access to top-level decision-makers across Europe. That access should not be wasted to persuade managers to abandon the patterns that made them successful. It should be used to meet them where they are.
Play the game they already understand: build compelling business offers, focus on clear capability gains and total cost advantages, and package them into concrete, integrated solutions. Not abstract sovereignty arguments, but deals that make sense on their own terms.
