Weekly Backlog Week 26/2026
🧠Editorial While outside the asphalt is slowly turning into lava, the tech industry is once again …

Bavaria suddenly wants to become more independent from Microsoft.
Microsoft wants to know when you’re in the office.
Trump wants to get involved with OpenAI.
The wallet is coming - maybe?!
And all of Europe is discussing digital sovereignty, while half of the public administration probably couldn’t even make coffee without American cloud services.
One could say: The timing is remarkable.
Welcome to the Weekly Backlog. This week is about dependencies that you only notice when they become inconvenient.
According to several media reports, Donald Trump is considering a state AI fund that would hold stakes in companies like OpenAI or Anthropic. The official reasoning sounds enticing: The American population should directly benefit from the economic success of the AI revolution.
But the closer you look, the more questions arise.
Because the leading AI companies are now valued at amounts that are considered ambitious even in Silicon Valley. OpenAI and Anthropic are facing potential IPOs with valuations well over a trillion US dollars. At the same time, many of these companies continue to burn billions. The costs for data centers, chips, energy, and model training are rising unchecked. The hope for future profits is enormous. The actual profits so far are not.
This raises a simple question: Why should the state act as a buyer now of all times?
In any normal market, a state entry at peak valuations would be a significant risk for taxpayers. With AI, it is still completely open which providers will even be among the winners in five or ten years. The market will consolidate. Many of today’s celebrated market leaders will fail, be acquired, or miss their expectations.
That’s exactly why the idea seems less like an investment strategy and more like a political lifeline for an industry whose valuations are increasingly difficult to justify.
Even more problematic is the conflict of interest. How independent can a state regulate an industry if it is simultaneously its major shareholder? Who will control the development of AI systems in the future: democratic institutions or companies in whose success these same institutions are financially involved?
Trump sells the project as citizen participation in technological progress. However, his political past provides enough reasons for skepticism. Throughout both terms, one accusation runs like a red thread: Political decisions regularly benefit himself, his circle, or selected economic allies.
Therefore, another question should be allowed:
Is it really about involving the population in the prosperity of AI? Or is it about providing highly valued tech companies with a state-funded major investor before the market begins to reassess their actual values?
When the state and Big Tech move too closely together, in the end, it is not the citizens who benefit – but those who already own the shares.
Even when Microsoft announced this feature months ago, the criticism was great. Data protectionists, works councils, and many employees rightly wondered why a collaboration platform suddenly needs to know who is in which office and when.
Now the feature is actually supposed to be rolled out. Teams will soon be able to detect via the company WLAN whether employees are in the office and make this information available to others.
Does this surprise me? Not at all.
I think it’s naive to believe that a corporation like Microsoft would forgo the opportunity to satisfy the voyeuristic needs of some control fetishists. After all, there is a market for such features. Otherwise, they wouldn’t be developed.
Of course, the whole thing is sold as better collaboration. That is always the official justification for such features. In reality, however, it primarily expands the amount of data that can be collected about employees.
The infrastructure for surveillance is built long before it is used.
And that’s exactly why this isn’t about a single Teams update. It’s about the creeping normalization of control. What seems like a helpful feature today can already be considered a self-evident expectation tomorrow.
Germany will get a digital wallet in 2027. This sounds like a long-overdue step towards modern administration. In fact, it is already apparent today that Germany will introduce a digital identity solution known to lack essential functions before it even starts.
Qualified electronic signatures will initially not be available. Payment functions neither. Even privacy-friendly procedures like zero-knowledge proofs, which could be used to prove age without revealing the date of birth, are unlikely to be ready in time. At the same time, the security assessment is not yet complete.
Nevertheless, politics sticks to the schedule.
This reveals a pattern that has accompanied many state digital projects for years: It is not the completion of a system that determines the time of introduction, but the introduction determines what is considered “finished.”
This is particularly remarkable because this is not just any administrative portal. The digital wallet is supposed to manage IDs, driver’s licenses, and other identity proofs in the future. It should lay the foundation for how citizens digitally identify themselves to the state and companies. Anyone building such infrastructure should first prove security, privacy, and functionality and only then announce a start date.
In addition, there are reports of possible conflicts of interest in the project environment. The allegations may be justified or not. What matters is something else: A project of this importance should not even get into a situation where its independence is publicly questioned.
Even more problematic, however, is the political message that emerges from this approach. If it is admitted months before the start that central functions will be missing, then it was not the product that was adapted to the requirements. Then the requirements were adapted to the schedule.
This is where the real failure lies.
Because the digital wallet is not a prestige project and not a PR measure. It is critical infrastructure. Anyone introducing it while essential functions are missing and open questions about security remain risks exactly what a digital identity vitally depends on: trust.
Germany needs a digital wallet.
But above all, Germany needs the ability to sell digital infrastructure as a success only when it actually works.
For months, a deal was discussed that would have further deepened Bavaria’s dependency on Microsoft. The goal: to centrally bundle Microsoft licenses for authorities and municipalities and purchase them more cheaply.
Critics warned early on about the consequences. Those who increasingly tie their digital infrastructure to a single US corporation may save money in the short term. In the long term, they pay with dependency.
Now the state government is pulling the plug. The deal is on hold for now. Instead, there is suddenly talk about digital sovereignty, geopolitical risks, and the need for independent solutions.
Does this surprise me? Not at all.
I would be more interested in what Markus Söder has to say about it today. Just a few months ago, negotiations with Microsoft were defended against all criticism. Those who pointed out digital dependencies, lock-in effects, and geopolitical risks were considered naysayers. Today, the state government justifies its change of course with exactly these arguments.
Today, the very risks are named that experts have been warning about for years.
The idea that one could increasingly tie critical state IT infrastructures to individual American corporations without becoming politically and economically blackmailable was always a dangerous delusion.
Digital dependency does not arise on the day a provider pulls the plug.
The infrastructure for dependency is built long before it is used.
That’s exactly why this isn’t about a single Microsoft contract. It’s about the realization that digital sovereignty and maximum platform dependency are not compatible goals.
Those who want European agency must also be willing to seriously build European and open alternatives.
đź”—https://www.br.de/nachrichten/bayern/korrektur-der-staatsregierung-vorerst-kein-deal-mit-microsoft,VLTSkXp & https://www.security-insider.de/kritik-bayern-microsoft-vertrag-digitale-souveraenitaet-a-c65a5c1d275dd2cc93318a32d0690d17/

EU sovereignty package strengthens European infrastructure policy, regulates dependencies, open-source strategies, and regulation; central question: balance of power between states, hyperscalers, and open standards.
Shows how attackers misuse cloud infrastructure from hyperscalers to bypass controls; intensifies debate about dependencies and sovereignty.
đź”—https://thehackernews.com/2026/06/pcpjack-hijacks-230-aws-google-cloud.html
Local open-source model enables edge computing on laptops, potentially reducing dependencies on hyperscalers.
Report on Office alternatives LibreOffice/OnlyOffice; emphasizes digital independence, openness, and European sovereignty over hyperscaler dependencies.
đź”—https://www.metacheles.de/office-365-alternativen-digital-independence-day-6/
A remarkable signal from Switzerland: The Bernese professor and open-source pioneer Matthias StĂĽrmer reports on a political decision that deserves attention far beyond national borders.
The Swiss Council of States has voted with a clear majority for an impulse program to strengthen digital sovereignty – and thus against the recommendation of
🧠Editorial While outside the asphalt is slowly turning into lava, the tech industry is once again …
🧠Editorial This week, I repeatedly asked myself whether we in IT are actually solving problems or …
🧠Editorial The best time to think about digital sovereignty was ten years ago. The second best is …