Takeaways from the EU Open Source Policy Summit 2026

Working at the intersection of digital transformation and public policy, I found EU Open Source Policy Summit 2026 deeply enriching. It is rare to see policymakers, open-source advocates, and tech practitioners share the stage and genuinely engage with one another. Below are the insights that resonated most with me, along with my personal reflections. The summit revealed both Europe’s extraordinary potential in open source and the urgent challenges we face in bridging the gap between policy ambition and technical execution.

The foundational principle: Public money, public code

The opening keynote captured exactly why this event was necessary. As Frank Feldmann put it: “Some people speak coding while some speak policy, but they do not speak with each other.”

This observation proved prescient throughout the event. Policymakers articulated policy frameworks with clarity, but when discussing technology, their language became noticeably vaguer, gesturing toward a distant galaxy that needs to be more this or that way. Conversely, technologists engaged deeply with technical topics but often treated policy as a product to purchase: we need more policy for this, less policy for that. While this is admittedly a generalization, a common thread ran through participants’ approaches, shaped by whether they came from policy or tech backgrounds.

I particularly appreciated Daniel Glazman’s confession that he still “codes for fun” when time allows. It served as a valuable reminder that passion fuels innovation and that enjoyment matters in everything we do (at least, that’s how I interpreted it). “Coding for fun” is also the backbone of open source technology, which is often dismissed as a playground for solo developers. Yet Gabriele Columbro pointed to research showing that open source creates USD 8.8 trillion in economic value.¹ Open source is hardly marginal. I see it as similar to water infrastructure: we depend on it because we need it, yet economically it paradoxically appears to have no intrinsic value because it’s free.

*Thibaut Kleiner shared examples of cities in Denmark and Germany adopting open-source solutions: “Open source is happening all across the EU… but it is scattered. We need to share best practices and improve how tenders are written to target open source software rather than closed-source alternatives”. Technology isn’t the bottleneck, as it has often been implied but rather,it is its adoption within public bodies that hampers deployment at scale.

AI and the decentralized future

The conversation later turned to AI and the decentralized future. While Europe is most likely late to the foundational model race, Mike Milinkovich suggested we could be the first to develop an AI frontier model that runs on a laptop. Technically, this is already possible. For instance, I run the model gemma3:4b on my MacBook M1 through ollama: the performance hardly compares to cloud-based models, but it’s sufficient to digest policy documents under 10 pages published by the European Commission on Have Your Say.²

Instead of focusing solely on building the biggest and mightiest model, we should incorporate sustainability and decentralization. Massive data centers are being built (or planned), but what would happen to this technological advancement and its dependent services if they’re damaged—say, by an earthquake? I can already hear objections that innovation cannot be constrained by budgets, policies, or any limitations. But an old Latin proverb reminds us: Mater artium necessitas (Necessity is the mother of invention). Without digressing too far, many tech giants dominating the digital landscape were born during recessions: WhatsApp, Uber, and Airbnb, for instance, emerged in the aftermath of the 2008 global financial crisis. A 2023 study found that ”[…] recession startups are better able to retain their founding inventors and build productive R&D teams around them.”³

I believe that a future where everyone can run locally AI models to simplify daily tasks it is possible.

The monopoly in the forge arena

During the online chat, an interesting discussion emerged about discoverability of open-source software. Since I started working as a web developer, GitHub has been the go-to platform for open source software (OSS). It has supported the open source community through free code hosting, CI/CD pipelines, and financial backing (including for events like this). In 2018, Microsoft acquired GitHub with a stated commitment to embrace open source. However, Microsoft’s reputation in the open source community remains complicated due to historical patterns of aggressive competition and harsh vendor lock-in policies with proprietary software like MS Office.

I mentioned GitLab, SourceForge, and Codeberg as open source alternatives to GitHub, but OSS project availability, discoverability, and user experience still require significant work to become serious competitors.⁴ The European Commission has also entered the “open source forge” space with the EU OSS Catalogue⁵, but to be honest, it is hard to use: filtering needs heavy lifting, search functionality is very limited, and the user experience cumbersome. While it’s build by and for the European (including national) public administrations, such distinctions in open source don’t make sense—especially in the context of interoperability. Open source cannot be contained within regional or national borders: by nature, it is open, inclusive, and collaborative. Therefore, the idea of a “European open source” sphere contradicts the very essence of open source. However, the term “European open source community” might legitimately refer to people working on open source who live in Europe.

I left the session thinking that good tools exist, there’s dedicated budget, yet we keep reinventing the wheel, rather than building the rest of the car. For instance, during a webinar on public software interoperability, participants asked for real-life examples — essentially, digestible and human-readable documentation on how to use the “building blocks”.⁶ Once again, a European open-source network exists, but it’s fragmented and hasn’t “discovered itself.” More accurately, multiple European open-source networks exist, often divided by domain (e.g., open source for public institutions) or policy area, such as think tanks or NGOs running their own OSS directory.

I’m not suggesting a solution is simple, as many actors are involved with equal rights. To quote Gabriele Columbro: “We need to invest more in business open source.”: invest steadily and consistently in open source technology that’s actually used, following a long-term sustainable strategy. Open source cannot survive on “coding for fun” alone — at least not without constant financial support. Developers are people too. If we want reliable, secure systems that serve citizens for years, we need structured funding. For instance, the Android Open Source Project operates as an independent entity within Google; Ubuntu is developed and distributed by Canonical. I see nothing wrong with corporations developing and distributing open source software, provided they relinquish code ownership (with some exceptions, like prohibiting use of their software to compete directly against them). Open source software follows democratic values and as democracies, it needs to start from the grassroots to thrive, but it requires a political system of check and balances to sustain itself long term. Those check and balances are often found within the communities themselves, like the case of Drupal ⁷ as well as the guidelines on how one can contribute to the projects⁸.

What “public money, public code” really means

Reflecting again on the opening motto—“Public money, public code”—we need to be more serious about where and how public code is distributed. States should invest in developing their (our) own solutions rather than relying on closed-source software that often comes with vendor lock-in. A great example is the joint effort led by the French (DINUM) and German (ZenDiS) states, which created a bureaucratic workspace (not merely an office suite) distributed freely without strings attached (MIT license)⁹.

Given how much tech infrastructure is shared through interoperability, I thought that a “package-manager-style OSS forge for public code” could solve this discoverability issue: something like Homebrew on Mac, APT on Ubuntu, or Chocolatey on Windows.

Connecting the dots:

• GitLab Community Edition (CE) or Forgejo would provide the open source infrastructure;
• A public-private joint venture would offer a public service to maintain the project, with the main public benefit of lowering single-market barriers for new and existing companies through easy and accessible distribution of OSS;
• Ad hoc, call for contributors to make existing projects “ready-to-adopt” by creating multilingual documentation and training: rarely mantainers speak all 27 official EU languages.

However, that would mean reinventing the wheel once more. What is actually needed is to start a grassroots movement among practitioners within public bodies, who distribute their products on platforms used by the general public.

The courage to build

This event inspired me to act with courage and clarity. As Alexandra Geese stated clearly: “We need courage to build, not only to say what we have to do.” I underlined this twice. Talking about ideas or objectives is not the same as building them.

Conclusion

The EU Open Source Policy Summit 2026 Online served as both mirror and compass: a mirror showing that our challenges are shared across Europe, and a compass pointing toward collaboration, courage, and concrete action. If we make the effort to understand each other despite language differences and professional silos, we can help write the next chapter of European digital sovereignty, one open-source project at a time.