In the first article of this series, I explained how successive waves of technology create what I call the tech sediment. In the second, I looked at the origins of technological lock in and how companies deal with it at the micro level. The essential message is that smart IT managers make conscious lock in decisions. They know that every technology creates some form of lock in anyway, that avoiding lock in creates costs today that may never pay off, and that future technological change can render much of today's precaution obsolete.

We are in a Macro Trap

The result is that microeconomic actors, in aggregate, consciously create a massive macroeconomic lock in to mostly foreign technologies, while individually benefiting from not investing too much in precautions against it.

At the macroeconomic level, this creates enormous leverage over the entire economy. But where exactly does this leverage come from? Most of these technologies are, in principle, replicable. There is not much secret sauce in providing a workplace collaboration suite to consumers or a database to corporations. The leverage comes from the asymmetry between how quickly a foreign supplier can restrict access to a technology and how long users need to migrate to an alternative. For a cloud service, access can be restricted immediately by shutting it down. For other software, disruption can happen almost as quickly, for example through a faulty update.

While it is technically possible to maintain a hot standby solution that a company could fall back on after such a kill switch, doing so would be prohibitively expensive and economically irrational for almost every individual actor.

This is the trap: companies and consumers optimize rationally at the microeconomic level and, in aggregate, produce systematic macroeconomic unsovereignty. And in a decentralized, free market economy, we actually want microeconomic actors to optimize for their own interests. This is what drives slack out of systems and makes companies productive and competitive. And that, in turn, is a major source of our wealth and economic wellbeing.

Economically, this is an externality: individual actors capture the benefits of accepting technological lock in, while part of the resulting geopolitical dependency and its risks are borne by the economy and society as a whole.

The Obvious Ways Out Do Not Work

The first reflex is to moralize. We tell companies that they should accept additional costs for the overall wellbeing of society. But moral appeals have not solved problems such as the gender pay gap or climate change, and there is little reason to believe they will solve digital dependencies. Companies will continue to optimize for their own immediate interests.

The second reflex is to focus on open source. As the second article showed, the leverage created by commercial lock in can be significantly reduced with truly free and open source software (FOSS). But this has been known for decades, and microeconomic actors still choose proprietary software and services. Consumers accept proprietary services because they are convenient and often free, even if they pay with their data. Companies often prefer them because the total cost of ownership are often lower than operating open source alternatives themselves, especially if they lack the internal expertise or willingness to do so.

The next intuition is to build European alternatives. These already exist, ranging from very small providers to sizeable European technology companies. Yet their overall market share has declined for years, while growth data from hyperscalers (Oracle 93%, GCP 82%, Azure 42%, AWS 37%) suggest that this trend is not reversing. Scale matters enormously in this market. Much of the hyperscalers' portfolio is software, and software scales extremely well. Cloud is also a mature market in which offering a kind-of equivalent product (like the EU clouds) is not enough to change market shares substantially.

Then there is public procurement: governments could simply buy European alternatives. This may strengthen individual European suppliers, but it is unlikely to untrap the economy as a whole. The public sector is relatively small compared with the broader market and notoriously slow in adopting new technologies. Many public organizations are still caught in debates about cloud versus non cloud that resemble discussions from the 2010s.

The final intuition comes from the EuroStack (the industry version of it): use industrial policy to combine private and public demand and thereby create scale for European technology suppliers. But assuming that microeconomic actors remain rational, this would require substantial incentives to overcome their TCO calculations. Remember the tech sediment. Organizations do not lightly replace technologies that have become deeply embedded in their operations. They keep mainframes running for decades and maintain old client server spaghetti applications voluntarily for exactly this reason.

More importantly, these organizations have more immediate concerns than optimizing tech commodities. Banks need to compete as banks, insurers as insurers and car manufacturers as car manufacturers. They are investing heavily in AI because they expect it to affect their competitiveness and growth. European technological sovereignty competes with these priorities for money, management attention and engineering capacity. And, as the German saying goes, das Hemd ist näher als die Hose: immediate interests tend to win over more distant ones.

We Need to Address the Tech Sediment

All of this may sound like a call for inaction. It is not. Rather, it should serve as the starting point for brainstorming solutions by defining the constraints within which those solutions have to work.

There are two fundamental ones:

• Microeconomic actors act in their own interest. Their primary objectives are to succeed in their business and to remain compliant. Within these boundaries, technology decisions are largely driven by TCO and cost benefit considerations.
• The tech sediment is a given constraint in these calculations. Existing technologies, processes, skills and dependencies cannot simply be wished away. Any intervention that requires organizations to change them has to account for the resulting costs.

Put differently, the economic cost of sovereignty is the cost of making microeconomic actors deviate from the choices their own TCO and benefit calculations would otherwise produce. The greater that deviation, the more expensive sovereignty becomes.

The question, then, is not how to convince microeconomic actors to stop behaving rationally. It is how we can achieve macroeconomic sovereignty while working with their incentives rather than against them. What could such solutions look like?

1: Immediate, unavoidable Pressure.

When organizations face immediate and unavoidable pressure to change, they change. Huawei developed its own mobile operating system after US sanctions restricted access to Google services. Companies across Europe adapted when GDPR came into force. Organizations updated legacy applications when the Year 2000 problem made inaction impossible. When Chinese companies cannot access leading US LLMs or GPUs, they develop domestic alternatives and architectures that use fewer tokens or less computing power.

The most radical solution would therefore be simple: cut off the supply of US technology ourselves. Europe could, in effect, trigger its own kill switch. This would certainly force companies to migrate, build alternatives and break up parts of the existing tech sediment.

It would probably also be economically suicidal or at least masochistic. We would have to redirect enormous amounts of economic capital into replacing technologies that currently work. Money that today goes into things such as tax breaks for commuters, subsidized early retirement, or subsidies for company cars would instead have to finance hyperscaler replication, old generation semiconductor fabs and plants for storage and network hardware.

So immediate pressure works. The problem is not effectiveness. The problem is the price.

2: Break up the sediment through regulation

Economically speaking, the problem is that the tech sediment prevents markets from doing their efficiency job. If we could find a magic formula and put it into regulation, we could make the market work again.

I still remember when consumers in Germany were effectively tied to their local electricity supplier. Then the EU liberalized the market, forced suppliers and network operators to work through standardized interfaces and processes, and made switching possible. Today, consumers can compare electricity contracts online and switch providers within minutes. Something similar happened with mobile phone numbers: number portability substantially reduced one important barrier to switching providers.

Could we do the same with technology?

Having spent my entire professional life in the engine room of IT, I am skeptical. The tech sediment is not just a contractual relationship between supplier and customer. It is a messy combination of people, skills, functions, processes, data, software and infrastructure that has accumulated over decades. I find it hard to imagine a regulatory formula that could magically turn this organism into something where individual components can be exchanged quickly and cheaply.

But I am not saying that it is impossible. I am saying that I lack the imagination to see how it could work. Going deeper into this question would certainly be worth a study or two.

3: Leverage the Nature Of Technology

Some patterns are remarkably stable in tech. Three of these are:

• Every few years, something new comes along that changes “everything”. While it never changes “everything” it significantly shifts what you can do.
• If you have an engineer at your side who knows their stuff, even difficult things suddenly become possible.
• If management really wants something, things move surprisingly fast.

The wave changing everything right now is clearly AI. And I have never seen a new technology that is so well suited to taking on the tech sediment. Combined with knowledge graphs and smart software, AI gives enterprises a real chance to finally catalogue and understand their accumulated mess of people, functions, processes, software and infrastructure. Two ingredients are still missing: managers who really want to do it and engineers with the expertise to make it happen.

Managers really want something when it promises growth or profit. If we managed to tie IT modernization to a major business objective, managers would have a reason to hire the right engineers, give them budget and freedom, and put pressure on the organization to deliver. Modernization could then break up parts of the tech sediment along the way, creating less opacity, simpler dependencies and a greater ability to switch. And that anyway ongoing change would entirely change the TCO considerations of integrating open source and European technologies.

This is not a solution yet. But it is also worth a study or two. Could we use the subsidies of the next crisis, which will surely come, to incentivize microeconomic actors to build new businesses that require the modernization of their IT? Instead of paying companies to migrate for the sake of sovereignty, we could help them do something they want to do anyway: create new products, revenues and businesses. And while doing so, they have to turn around their tech sediment and give European technology suppliers a fresh chance to compete.

Make Sovereignty the Rational Choice

The Micro Macro Trap is real. We in the sovereignty bubble pursue goals that are different from those of the companies and consumers who create the actual GDP. If we want to have an effect that scales, we need to surrender to their thinking and their reality. And that reality is simple: the tech sediment exists, and microeconomic actors make decisions based on TCO versus expected benefits, not for the public good.

If there were immediate and unavoidable pressure to change, as Chinese companies experienced after US sanctions, the problem would become easier to solve, but enormously expensive. Since we do not want to create that pressure ourselves, we need to be more creative.

Regulation could be part of the answer. But if the tech sediment is the context, software will almost certainly outsmart regulation. Its complex web of technologies, processes, dependencies and people offers too many ways for even well made rules to miss their target.

So perhaps the more promising approach is to leverage forces that already operate within the world of microeconomic actors: new waves of technology, profit, growth and competitive pressure. Instead of asking companies to act against their economic interests in the name of sovereignty, we should look for opportunities to make our macroeconomic goals part of decisions they want to make anyway.

If we manage to align the two, we may have a chance of achieving more technological sovereignty at scale over the course of the current platform shift without having to pay the enormous price of forcing microeconomic actors out of their own incentives.