Digital (Non)Sovereignty: A conversation

In January 2025, Seda and Femke prepared their input for a workshop organised by Plixavra x by a way of this epistolary exchange. Between many other things, we tried to grapple with the re-emergence of Digital Sovereignty in a desperate attempt to safeguard Europe's competitiveness.

---

Dear Seda,

I wanted to write you yesterday but ended up hanging out with Heura and her friends in a park.

By the time it was getting dark, there was a conflict between them (I don't really know what happened as my Spanish is still basic) and they came to stand around us adults, reporting together that something was up.

Then (the kids) found some chairs and made a meeting in the dusky park, taking time to listen to each other and how they each felt, and then proposed and agreed upon a reparation. Ten minutes later they were playing again.

Jara explained that conflict resolution is something these kids learn at school, an anarchist pedagogical project in the neighborhood.

"to be exposed to one’s exposure is not equal to being destroyed by it. Intentionality minimizes loss." (Berlant)

Thinking of what "digital sovereignty through a feminist lense" might be, the amazing frankness by which these kids spoke about their emotions (they are 6 years old) while simultaneously taking care of the needs of the group, really struck me.

I think sovereignty is a tricky term, both in its negation and in its fiction but I am trying not to get hung up on it. If these kids have learned how to mix autonomy with interdependence, can we think of digital infrastructures that do that?

---

Dear Femke,

I have been sitting with the warm account of your Friday afternoon. In fact, I am reading Berlant to make your story of the afternoon with Heura and friends more legible to me. I can tell the story is an articulation of what it means to move in the universe, or would she call it lifeworld Berlant proposes we create.

And, I am reading you to make Berlant more legible to me, as her verse and her propositionality (rather than making a claim to truth) requires a switch in thinking and action that comes easier if you hold my hand with your stories.

I am also keenly aware that you and I have this ambitious plan to bring Berlant's concept of non-sovereign relationality into the workshop room, oh and at different scales. When we add to that that we want to also bring in a bit of computer science and a bit of creative practice and who knows what else to a precious room Plixavra already potentially stretched to its rims with her ambitions to use feminist legal theory to take a stab at digital sovereignty, we can without reservation say: we are aiming big.

To keep it grounded in the invitation, I agree that digital sovereignty is a tricky term. My go to would be to critique it. Maybe something like this:

[Put on a voice of authority, maybe one register deeper than usual] Digital sovereignty uncomfortably aligns with protectionist forces in ever more right-wing European states. Its sister notion "strategic autonomy" allows ex-neoliberals to piggy back on the protectionist conservatives to mobilize billions of Euros to sustain European competitiveness. Talk about tricky. In the process, these parties justify putting financial resources into speculative innovation by tech companies instead of any project that may actually respond to social needs to attain some sort of distributive, racial, ecological justice. In 2018, French nationalists talked about digital sovereignty as a path to decolonizing the otherwise enslaved Europe. This was picked up again and again at an event at the EU Parliament on "European Digital Independence: Building Alliances for the Eurostack" organized by Francesca Bria and Cristina Caffara, two women claiming a feminist space in and around the commission. While I appreciate the scale at which the two move, why not, I struggled with the idea that the more progressive claim to digital sovereignty is that "we" pursue a European version of the computational infrastructures made using an open source stack. This suggests a blindness to the way these computational infrastructures are expected to transform industrial production and with it the economic basis that we currently rely on, towards something that will even less benefit most people in the world...the frustrating and maybe exhausting oxymorons are too many to count.

But, my understanding is that Berlant proposes to start somewhere else that is more energizing. Not necessarilty to start with the inconveniences of all these people who are pushing digital sovereignty into our eyes, ears, mouths and political agendas, but to engage the inconveniences of the people and things we are driven to.

Berlant says: "We know why threatening things threaten us; it’s harder to know why it’s difficult to live with the ­things we want. It can’t just be that we fear losing them, we fear having them too."

I want to find out more about this with you and the others. This call for establishing a non-sovereign relationality. And, did I say already, Berlant is very difficult to read, but I understand how in her propositions, she invites us to create rather than consume our energy in resisting for the sake of resistance and critique. She asks us to look at the ordinary and the inconveniences it throws as places to start creating that transitional infrastructure in which we recognize our interdependence and can expose and be exposed to our inconveniences. Like Heura, the 6 year old and her friends.

I fear though this all sounds quite abstract. So, let me pose a question to you, how do you and I create a transitional infrastructure that invites the room with Plixavra and a bunch of lovely others to think with the world of Berlant, to think about non-sovereignty in place of digital sovereignty? Do you maybe have another story?

---

Dear Seda,

I think both holding the need for technological self-determination (to use another word than sovereignty) and a fierce critique the devastating nightmare of digital sovereignty is what is at stake in nauseating times like these.

I have read Berlant’s piece on Infrastructures for troubling times so many times by now, with students and friends and comrades and every time I find different images and story-lines to struggle against despair.

Here's one that made my morning, resonating with the one I sent you yesterday:

“To take something in is to be nonsovereign in relation to it, but that’s not equal to being destroyed by it. If we can distinguish mode from method, this mode allows presenting through movement, and not just movement in general but through digestion and extrusion of infrastructure at many material scales, like a worm.”

But you asked another story, one to respond to the questions that is closer to the ground.

As you know, I have been following the Nextcloud project for some years, trying to process my resistance to it while also being a daily user and often finding myself advising collectives to use it when they want to move their files out of Google or DropBox.

Two months ago, the board of Nubo received an email from Kenny, who is system administrator for several collectives working out of a basement at Maison de la Paix in Brussels. One of these collectives is Nubo, a small cooperative that offers locally hosted email and digital storage services for a reasonable price, and I am a member of the hands-on board of this cooperative. He ends his email to the Nubo board with a terse statement. “Making the contact synchronization work again, is something I don't want to do, nor maintain.” he writes.

What Kenny explained in his email is that due to continuous updates in Nextcloud, which makes up an important part of Nubo’s software stack, some of the code that he developed for synchronizing contact databases that are stored in Nextcloud and used by Roundcube, a web-based mail client, keep breaking. In order to keep the fragile ecosystem working, he cannot update to the Nextcloud component which proposes a security issue and results in alarming error messages that appear on the phones of subscribers.

He proposes to either disconnect the two applications or to abandon Roundcube altogether in favor of a plugin offered by Nextcloud which allows users to check their emails through the same interface as they use for their calendar, photo gallery and file storage. With Emmanuel, also part of Nubo’s board, we are seriously concerned about this development. It is a difficult conversation to have because due to the modest scale of Nubo, there is no financial margin to pay for extensive in-house development, and we also want to be careful with Kenny’s time and motivation. He maintains the Nubo infrastructure mainly in the margins of his employment for other collectives gathered in the same basement, and this means we are indirectly depending on decisions he makes for other structures.

As part of it’s mission to “honor people’s time and mental load”, Nextcloud now also offers an integrated webmail interface that could technically replace Roundcube’s functionalities and would reduce the risk of bugs, and security issues also because the technically complex work of bridging between applications is shared with the Nextcloud community and not carried by our tiny team alone.

But letting go of Roundcube would mean first of all deserting a project that by now is the last remaining possibility for hosting your own mailservice using Free Software.

Our main concern is though that by putting all our eggs in the basket of Nextcloud, Nubo becomes dependent on a single software even when it is an expansive one. Our project will be technically nothing more than an interface, a platframe, on top of a continuously expanding suite of functionalities decided by the Nextcloud community and a German GMbh. By giving up on the attempt to thread different software projects together, the Nubo infrastructure risks to become yet another technical monoculture that is counter to our desire to flourish a diverse and reconfigurable digital infrastructure in lieu of offering the kind of monolithic, seamless integration provided by Big Tech. “But for our subscribers it would also be more easy to use the same interface, like in Teams”, Kenny argues. And he might be right.

There are some more twists to this story, but I leave it here for now.

What is difficult to stomach is that Nextcloud has decided to compete directly with Microsoft teams, so it is always in a frustrating race to provide a similar experience, and in the process projects with a very different approach like Nubo can not but follow. Nextcloud’s most used argument, repeated in testimonials by commercial and government clients, is that unlike GAFAM, Nextcloud is “putting IT back in control”. “We address the limitations of other platforms by delivering true digital sovereignty and full transparency” they write on their website.

And yes, as with the European Stack proposal that you mentioned, for Free, Libre and Open Source software communities, “digital sovereignty”, sometimes “technological sovereignty” takes on a significance more close to food sovereignty movements, so it is about the struggle for people's right to sustainable and appropriate technologies, and also to have ways to define their own tools and systems.

Benevolent projects are playing dangerous games when they deliberately or not, confound the desire for a sense of self-determination when it comes to digital tools and computational infrastructures, with the deathly claim to digital sovereignty without even trying to address the power moves going on where all of the worlds resources are being oriented towards always more compute.

The issue I am trying to get at, is how to concretely get at transitional infrastructures at a scale that make sense to people here and now, that is of use, when not only the imaginaries but also the modes are set by industrial giants of computation?

---

Dear Femke,

Yes yes, these companies are good at hijacking our imaginations and speculations. But I see in this conversation, how we can find openings to set them free again, maybe also through roaming into and out of the shadows of these industrial giants.

Based on our work, collaborations that cross TITIPI, Martha Poon and Programmable Infrastructures, (will mention in intro) I want to make two proposals that, if helpful, can help identify what transitional infrastructures we need.

Proposal One: despite all the focus of digital sovereignty on clouds, I propose that we talk about computational infrastructures, which are not just clouds, but clouds plus end-devices. End-devices can be smartphones, or your laptop, or a sensor, or a machine in a factory with some programmability built into it.

The term the cloud is of course as lofty as the literal clouds in the skies. But the cloud is often understood simply as the provisioning of storage, such as AWS’ S3 and compute, as in EC2. In the popular imagination, the cloud lives in data centres. And indeed, the increasing construction of data centres, and their thirst for energy and water, but also their reliance on raw materials sourced all over the world, has become a point of contention in a world confronted with climate change on the one hand, and collapsing supply chains that have necessitated a geo-political re-ordering of the global logistical space. The latter is what also prompted the original motive behind the desire for European sovereignty.

Clouds, however, are much more than just storage and compute situated in datacentres. It is also more than the network links criss-crossing the globe. It increasingly includes devices for edge-computing and end devices, such as smartphones and IoT devices. They are woven together in an intricate and interdependent mesh, by now, you cannot have one without the other. I propose to keep this complexity in our minds, I propose thinking of computational infrastructures as cloud plus end-devices as its accessories.

This proposal hopes to bring our interdependencies with clouds even better into view. I will explain with a short anecdote. In December, I finally ordered a new computer through the university. My ambitions to continue using my previous computer into its 8th year failed once its screen went dark on me. When TU Delft finally handed over the device to me, they had installed on it software to control my machine remotely "through the clouds". With this, they can check what software I am running, they can control the licences on some of my software, and they log every adminsitrative action I take on the machine, requiring me to document it before doing so. I smiled at them when I picked up the computer and left thinking I would do a clean install when I got home. Only to find out that the machine was locked down to the point that I could not do a clean install. I was no longer smiling. I could not even log into my machine without being on the TU Delft network. Frustrated, I checked with kind colleagues in a chat who all lamented the various ways in which the remote control had bricked their computers or turned disfunctional and one said they were able to ask the ICT for a clean install without the remote control due to these problems. The chat turned "collective against remote control of our computers" helped me diligently list grievances I/we have with the remote control software, a range of complaints from all of us, written as coming from me, featuring concerns about not being able to work and surveillance. As they filled the chat with popcorn emojis, I switched windows to put in a request to ICT to please give me a fresh install. This was quickly sent to our security officer, who said that this could only be done if my department head asked for it. As I kicked off a process with my department head, someone from ICT contacted me and asked me why I struggled with the remote controls. I figured this person had read my list of grievances, but decided to patiently list th em again. In response, he said, but those don't seem like a big issue (!), and went onto a chatbot mode asking me what is the issue. I later got a copy of his messages to me from my department head who asked me now a third time, to list my concerns about the remote control software. Sidenote: maybe I also propose "administrative gaslighting" as a useful term. Long story short, I now have permission to get a fresh install, one still pending since it requires me to give up my computer for a day, a logistical act.

Clouds were not always like this. As the clouds evolved, they came to not only serve services, but also to become a location from which to manage the devices to which those services are serces. In the process, our personal devices switched from "connecting to the internet" for "internety things" to being completely dependent on the clouds to function, or even to be. This, we are told, is to manage a series of inconveniences, ICT personnel costs, high profile security breaches and bad reputations of the organizations we work for. Clouds hence replace inconvenient interdependencies with its own, but under the disguise of convenience, efficiency, and continuity.

Second proposal. The deeper we look into computational infrastructures, clouds plus end-devices, the more we see that they are a site of a transformative shift in how (economic) value is created. By this I do not refer to notions such as surveillance capitalism, as this puts an undue emphasis on the relationship between businesses and end-users. Rather, computational infrastructures, how they exist today, are an engine continously transforming how software is produced and deployed. What you and I experience as digital services is not an inevitability, and it has fundamentally altered how businesses digitalize and how, therefore, they operate, or exist in the world, also economically.

To put it succinctly: computational infrastructures are a production environment. Computational infrastructures define the parameters under which businesses can operate. Computational infrastructures tend towards long-term relationships that are insufficiently capurted by notions such as rent-seeking behaviour or vendor lock-in. To operate a software driven business today means to operate under the production conditions that the large providers of computational infrastructure dictate.

Computational infrastructures are about transforming how businesses operate in a way that zips these infrastructures to the heart of target organizations. This creates a perverse sitaution, where the revenue of tech companies come to depend on the continous revenue of the businesses they target, an interdependency the scale and inconvniences of which we do not yet know how to spell out. Yet, every time some news about DeepSeek or Cloudstrike hits their stock prices, with the potential domino effects on our retirement funds, on the countries some of us come from summarized as "emerging markets", or simply the local economies that pay some of us our salaries and others their social benefits (or nothing at all), we feel that it is not just the tech companies that are showing their volatility in this scheme, but all the businesses that have entered an interdependence with them through a dozen of interfaces, also called APIs.

If we think with these two proposals, we see that the problem with digital sovereignty is not that they are American companies, and they have our data, and with that data they are the ones that can develop these services and take decision making power away from poor Europeans, which we can take back if we only develop our own infrastructure. On the contrary, by going for simple replication of its production model, in which the tech provider and the target company are zipped together, efforts subsumed under digital sovereignty are likely to be replicating their rather extractive economic model without even understanding how they are doing so in the process.

It remains an exhausting and some days dreadful project to study how CI is transforming how we create value. But, with this conversation, I understand that their value is not only in providing a more robust critique, but in identifying the questions that we maybe want to be asking. And those are the following:
1) what is the model of production we wish to live with?
2) what is the relationship of immediacy, efficiency and intimacy that we may want to have with our devices and, maybe more importantly, with each other?

---

Dear Seda,

Today Helen said: 'Meso is the best'!

On another track, Helen and me tried to imagine how TITiPI could take on feminist economies. Basically we were wondering ... can we find ways to practice other value systems for computation. I think we need to keep doing those experiments, exactly because of the questions you ask at the end of your message: What modes of production? How to change our relation to immediacy and efficiency? Easier said than done!

I'm off to sleep now. I might write you more tomorrow morning, or maybe leave it here for now. We'll speak more when you are arrive.

x F