The European Commission recently announced that it had joined W-Social, presenting it as an independent social media platform based in Europe. Ursula von der Leyen also announced her move to W-Social, along with other officials like Henna Virkkunen, joining the network. The platform has since launched a public beta, opening to a broader audience for the first time.
These moves come amid a broader debate about the risks of dependence on technologies and platforms that expose users to extensive surveillance or the possibility of geopolitical disruption. They also come against the backdrop of the Commission’s own Tech Sovereignty Package, which places open source and decentralized social media at the core of its digital sovereignty approach.
Social media platforms have been at the sharpest edge of sovereignty-related concerns. Platforms such as X have become associated with engagement-driven recommendation systems weaponized to influence political discourse, the disproportionate amplification of far-right content, and, more recently, scandals involving AI-generated non-consensual sexual imagery.
In this context, and amid calls for “European alternatives”, W-Social is being positioned as a potential option. But it may be too early to treat it as one. Based on current signals, it is worth scrutinizing whether W-Social is building toward an open ecosystem or toward platform capture while using the architecture of openness and the language of sovereignty.
W-Social provides a useful case study for examining what digital sovereignty should mean in the context of social media. This piece argues that two issues deserve particular scrutiny: the closure of infrastructure that is presented as open, and the continuing dependence on infrastructure that remains concentrated in a single provider. Together, these illustrate that “sovereign” social media depends on open infrastructure and a credible path towards reducing structural dependencies, rather than European ownership alone.
What’s the deal with W-Social
W-Social is built on the AT Protocol, also know as ATproto or ATP, the open protocol that also underpins Bluesky. From a sovereignty perspective, this architectural choice is arguably more important than the fact that W-Social is European. AT Protocol disaggregates the different elements of a social media platform (such as identity, hosting, moderation, and user-facing applications) with the aim of limiting the power of a single entity over the network.
In our recent policy brief, we argued that the path toward a better social media ecosystem is not simply replacing American platforms with European ones. The challenge is not ownership alone, but closed platforms that lock users in, built on engagement and surveillance-driven business models. Alternatives require building on open infrastructures that enable fundamentally different forms of governance and curation.
That requirement for openness and plurality in infrastructure and curation is also what makes some of W-Social’s recent decisions worth scrutinizing.
When first announced at the World Economic Forum in Davos, W-Social did not disclose, and later appeared to obscure, the fact that the platform would build on ATProtocol. The platform has also recently moved to a closed-source development model.
This approach seems misaligned with the broader push towards openness and digital sovereignty. Transparency, scrutiny, and the ability to build on shared infrastructure are central to the value proposition of open ecosystems.
When the Commission announced it had joined W-Social, it did not create a new presence on a new platform. It moved an already existing ATproto account from one data storage provider (Bluesky) to another (W-Social). In ATproto, this storage infrastructure is called a Personal Data Server (PDS). It is where a user’s content, identity, and social graph live. One of the protocol’s core design principles is that any PDS should work with any application built on it, decoupling where data is stored from which apps can access it.
There are reports, however, that W-Social requires users to remain on its own PDS. This means that the Commission cannot use W-Social without also hosting its data on W-Social’s PDS. This architecture undermines the portability that the protocol was designed to guarantee, tying identity and content to a single provider in the way that conventional platforms do, and that open protocols were designed to prevent.
Open protocols derive their value from the ability of developers, projects, and users to build on shared infrastructure. A platform that restricts portability, closes its source code, and offers no transparency about its infrastructure roadmap while benefiting from the collectively produced network risks extracting from the ecosystem, instead of strengthening it. In that context, the language of sovereignty becomes rhetorical, not architectural.
Sovereignty is about infrastructure, not ownership
Whether a platform constitutes a meaningfully “sovereign” alternative is a separate question from where its headquarters are. ATproto separates into distinct layers what conventional platforms bundle together. Sovereignty, in this context, is a question of control over these interconnected layers of infrastructure.
Personal Data Servers hold each user’s data. Relays aggregate updates (posts, likes, shares) from Personal Data Servers and distribute them across the network. AppViews organizes and indexes this information so apps can display it to client apps used by users. Moderation services determine what content appears and under what conditions.
As things currently stand, W-Social appears only to offer EU-based Personal Data Servers. While hosting user data in Europe may provide a limited form of jurisdictional control, it is only one component of the broader infrastructure stack.
The limitations of that partial control have already become visible. Last week, new W-Social users were unable to post, like, comment, or edit their profiles for several days. W-Social had been rate-limited by Bluesky’s Relay, the infrastructure it still depends on to distribute content across the ATproto network. Rate-limiting of this kind is usually a spam-prevention measure, but this incident illustrates what incomplete infrastructural sovereignty means in practice. A platform can be headquartered in Europe while remaining subject to operational decisions made elsewhere.
Further up the stack, the dependencies become more consequential, with AppViews and moderation systems as the most resource-intensive components, and the ones where control matters most. Without its own AppView, W-Social remains dependent on Bluesky’s AppView and moderation infrastructure, which gives the latter ultimate authority over the actors, behaviour, and content permissible on the network. Even with an independent AppView, moderation infrastructure still requires substantial resources — a commitment W-Social has not publicly signalled.
If key moderation decisions remain dependent on infrastructure operated by Bluesky PBC, claims of sovereignty and independence are difficult to sustain. More fundamentally, dependence on external infrastructure constrains the possibility of developing meaningful alternative recommendation systems and curation models.
Issues related to infrastructure dependency and business models are not unique to W-Social, but they are issues that any initiative claiming to offer a sovereign alternative must address. Building on an open protocol while depending on external infrastructure, and funding operations through advertising (as W-Social has announced they will), risks importing the same incentive structures that define the platforms such initiatives seek to replace.
What distinguishes a credible alternative is not the absence of dependencies, but the willingness to acknowledge them and articulate a path beyond them. For instance, the Eurosky initiative, which already runs an EU-based Relay and maintains a development plan that more clearly signals their objectives and vision, including building an independent AppView and moderation infrastructure. Mastodon, building on the ActivityPub protocol, has long offered the ability to run independent infrastructure. The recently announced European Social Stack brings together multiple initiatives that build on different open protocols and explicitly signals similar commitments.
A gateway, if not an alternative?
One of the strengths of open protocols is that gains for one provider can benefit the ecosystem as a whole. Every user who joins W-Social is also joining the broader ATproto network. They can interact with users on Bluesky, Blacksky, Eurosky’s Mu, and future services that have not yet been built. With bridging infrastructure, cross-protocol exchange — between ActivityPub and ATproto applications, for instance — is also possible.
In that sense, W-Social could become another gateway to a more decentralized social web. With its heightened visibility and greater resources, it is possible that W-Social will eventually invest in independent AppView infrastructure, moderation services, or other parts of the stack that would further decentralize control over the network. If these investments materialize, they could represent a meaningful contribution to social media sovereignty.
A clearer signal of that intent, with a public infrastructure roadmap and a renewed commitment to open-source development and interoperability, would strengthen sovereignty claims around W-Social.
That potential sits alongside a present reality: users moving to W-Social from Bluesky are, on current evidence, moving toward less openness rather than more.
There still remains a risk that attention and political support become concentrated around projects whose visibility rests more on narrative than on demonstrated commitment to open infrastructure, while other initiatives that have been more committed to the ethos of open infrastructures struggle to gain visibility and resources.
The Commission and its officials’ decisions to join W-Social, and the visibility that follows from it, is a case in point. This visibility shapes which projects receive funding and support, and which understanding of “sovereign” becomes the default. If it is not accompanied by clearer standards — around openness, portability, infrastructure independence, and business model — it risks rewarding the narrative rather than the practice.
Different models, not just different flags
How we design our social media ecosystems determines the conditions under which people encounter information, form opinions, and participate in public life.
Platforms optimized for engagement and funded by surveillance advertising shape those conditions in ways that are largely antithetical to meaningful exchange. Closed infrastructures compound this by design: they are opaque to users, unaccountable to democratic oversight, and premised on making exit costly.
Architectural openness creates the possibility of alternatives, enabling algorithmic plurality and giving users a choice between different curation models, but it does not guarantee them. That still requires deliberate choices about product design, governance, and business model that go beyond the protocol layer. Alternatives require open protocols and interoperable infrastructure, but also recommendation systems and business models not structurally dependent on capturing and monetising attention.
If sovereignty is reduced to the nationality of a platform operator, the EU risks reproducing the same dependencies and structural dynamics it seeks to escape. A European platform built around closed infrastructure and engagement-driven recommendations does little to address the threats to democracy and the safety of people online, regardless of where it is incorporated.
The building blocks are already emerging — in open protocols, in the initiatives already investing in independent infrastructure, and in the Commission’s Tech Sovereignty Package that identifies the right priorities. In a new call for proposals this week, the Commission commits €1.48 million towards a protocol-based social media service “where users have greater choice about where their data is stored, how their content is moderated, and which apps they want to use to connect with their peers.”
The challenge now is to ensure that such investments and endorsements strengthen the openness and interoperability of the social media stack, rather than consolidating new dependencies within it.
