Nextcloud is not a suite, it is a platform

Nextcloud is standing in front of an open window. Most people, including many inside the ecosystem itself, have not yet noticed it is open. What started as a safe alternative for file storage has quietly grown into something fundamentally different. The technical reality is already that of a platform. The name is the only thing still trailing behind. And in technology, as history teaches us, naming is not cosmetic. Naming is load-bearing. Windows close.

The thesis

You buy a suite. You use a workspace. You inhabit a platform.

Ruben van der Linde, CTO Conduction

Suite, workspace, platform

Three words that get used interchangeably. They mean very different things.

A suite is a box with predetermined tools. Mail, files and office apps live in a single box. Integration between those tools happens in the user's head, or in the open browser tabs on their screen. The rest you solve yourself.

A workspace is one coherent surface where data, identity and workflows are stitched together. Chat, video, project management, knowledge management and code hosting live alongside mail and files. One login, one permission model, one user experience. The user no longer switches between ten systems.

A platform is the infrastructure on which an organisation builds its own long tail. Not the vendor. The organisation itself. A platform has a marketplace, a composition layer, an identity layer, an AI substrate and an economic model that feeds an ecosystem. On a platform, software runs that nobody at the original vendor ever envisioned. That is the definition.

Let's be honest. The difference is not a marketing trick. It determines which question an organisation asks at the moment it buys. It determines where money flows over the next decade. And it determines something many people underestimate: whether an open-source ecosystem will beat Microsoft Power Apps on the terrain that matters most tomorrow.

Tomorrow's functionality will not be built by vendors anymore. It will be built by users with AI. And in that world, scale wins. Nextcloud already has that scale: more than 16 million users worldwide.¹ Power Apps has around 56 million.² The gap is smaller than it looks, and the open side plays with a fundamentally better business model.

At Conduction we have made our choice. We are putting our time, our money and our engineering talent into two things: helping Nextcloud evolve further into a real workspace, and building the underlying platform infrastructure that lets developers and civil servants ship their own applications. No fork, no separate product, no lock-in. Just acceleration of what is already there.

The anatomy of a platform moment

Software categories do not tip on features. They tip on framing. Three examples from the past thirty years show the pattern.

WordPress broke through by keeping the core small and the ecosystem large. The features did not win, the ecosystem did. Today more than 43 percent of the entire web runs on WordPress.³

Salesforce launched the AppExchange in 2006. Four years later the central message was no longer "Salesforce is a CRM" but "Salesforce is a platform on which an ecosystem runs". Today 91 percent of customers use at least one app from the marketplace.⁴ The moat is no longer CRM functionality. The moat is the marketplace.

The iPhone is the best-known example. In 2007 Steve Jobs positioned the device as "the platform that runs apps", at a moment when the App Store did not yet exist. The naming ran ahead of the technology. The technology followed.⁵

And then the cautionary example: SharePoint. Microsoft positioned SharePoint for years as a platform, but never built the economic architecture that feeds an ecosystem. No marketplace that works. No revenue share that rewards developers. No citizen-developer on-ramp. It became a suite that called itself a platform. That is not a platform, that is theatre.

The pattern is clear. Platform moments emerge when three things are true at the same time: the technical substrate is in place, the story shifts, and the ecosystem gets economic oxygen. For Nextcloud, two of those three are now true. The third is within reach.

Year one versus year three

When a Dutch municipality or a German ministry thinks about digital sovereignty, it asks a very different question in year one than in year three.

For year one, a suite works. For year three, you need a platform. A workspace sits in between. The bundle wins year one and loses year three. That is the trap European governments are walking into right now, with the best intentions, while the procurement clock keeps running.

The substrate is already there

What is special about Nextcloud is that the platform substrate is already technically in place. Anyone who looks critically sees it immediately.

Files with version control, sharing mechanisms and external storage connectors. Calendar and Contacts via open standards (CalDAV, CardDAV, IMAP). Talk for chat and video, federated via Matrix. Mail as an integrated inbox. LDAP, SAML and OIDC as an identity layer that hundreds of organisations already run on. ExApps as the architecture to plug external services into the workspace as first-class citizens.

On top of that substrate, components have lived for years that you would call "apps" if Apple were selling them. A wiki via XWiki. Project management via OpenProject. Code hosting via GitLab. A learning record via Scholiq. Federated chat via Matrix or a team environment via Mattermost. Automated workflows via n8n, or Windmill for the developers who prefer working in code. PII anonymisation via Presidio. Single sign-on via Keycloak. AI inference via Claude, Mistral, or locally via Ollama.

This is not a suite. A suite is a box with predetermined tools. This is an infrastructure on which an organisation builds its work.

And yet, in nearly every sovereign-workplace bundle in Europe (openDesk in Germany, Mijn Bureau in the Netherlands, La Suite Numérique in France) Nextcloud is positioned as one of the components. Not as the underlying layer.⁶ That is a framing mistake with large consequences, because the buyer behaves accordingly. Anyone who buys Nextcloud as "the files component" locks themselves into year one while the problem plays out in year three.

Anyone who buys Nextcloud as "the files component" locks themselves into year one while the problem plays out in year three.

Ruben van der Linde, CTO Conduction

The sovereignty calculus around AI

The next question every government asks in 2026 is: "how do we do AI?" And right here the architecture tips in a way most decision-makers have not yet internalised.

Microsoft Copilot routes data to cloud infrastructure for inference. Google Workspace does the same. The sovereign variants (Bleu in France, Delos in Germany, Microsoft Cloud for Sovereignty) reduce that exposure. Even so, they hit the CLOUD Act legally, because the parent company is American.⁷ For some organisations that is acceptable, for others it is not. For the ICC it demonstrably was not.

Nextcloud flips the entire architecture. The starting point is that the data already lives on infrastructure the customer controls. The AI substrate comes to the data, not the other way around. That sounds subtle, only it is a world of difference. It means a municipality can choose a local model on Ollama when the data demands it, Mistral in a European cloud when the sensitivity level allows it, or Claude via a controlled gateway when the use case needs a frontier model. With Presidio in front to anonymise PII where required. With audit logging on every tool call. With the operator choosing the model, not the vendor.

This is the sovereignty calculus that Microsoft and Google cannot run. Not for technical reasons, only because of their business model. Their margins depend on inference in their cloud. Anyone who moves inference to the customer also moves margin away from headquarters. You do not do that voluntarily.

The citizen-developer wave

And then the third piece, the piece that completes the platform moment. A platform only comes alive when its users build on it themselves. Not just developers, all users. Microsoft has bet massively on this for the past ten years, and the numbers are genuinely impressive.

Power Platform grew from 33 million monthly active users in 2023 to 56 million in 2025. Microsoft has a public target of 500 million users in 2030.² The 2026 release bundles Power Apps, Power Automate, Power Pages, Copilot Studio and Dataverse into a single AI-native citizen-developer environment. The internal framing is candid: "low code is dead". It is being replaced by natural language and AI agents that build for the user.

Microsoft is running from 56 million to half a billion in five years. That is the trajectory of the other side of the race. The race is on. Either Europe ships an answer that scales at the same rate, or in 2030 every leave request, every supplier onboarding flow, every visitor register on the continent runs on someone else's runtime.

Here comes the interesting math. Until recently the reasoning in every comparison with Microsoft was: they have a bigger R&D budget, more engineers, a larger product organisation, so they win on functionality. That math no longer holds. Functionality will not be built by product organisations going forward. Functionality will be built by end users who use AI assistants to make their own apps. And in that new math, the side with the largest engineering team does not win. The side with the largest citizen-developer pool wins.

Tomorrow's functionality will not be built by vendors. It will be built by users with AI. In that world, scale wins.

Ruben van der Linde, CTO Conduction

16 million is a headline. The strategy lives in two segments.

The bare 16M-vs-56M gap is interesting, only it hides the actual question. We do not need to convert every Nextcloud user into a citizen-developer to win. We need to convert the segments where the gap to Microsoft is closeable, where the procurement door is open, and where Microsoft's own business model blocks them from following us. Two motions matter. Both are uniquely accessible to an open platform.

One French civil servant who uses Claude to build a supplier-onboarding app delivers more functionality than a team of five product managers in Redmond. One MKB owner in Almere who uses Mistral to wire a quote-to-invoice workflow across mail, files and a register saves their accountant three hours a month. Multiply that by 2.5 million French civil servants, plus the 160,000-and-growing German base, plus an MKB tail measured in tens of millions across Europe, and the race takes a very different shape. The 56 million Power Apps users is the headline number. What we are building is a wedge into the segments Microsoft cannot follow.

On top of that, the open side plays with a fundamentally better business model. Microsoft has to monetise every citizen-developer through license fees, or the business case burns down. An open-source platform monetises through support, integration, hosting and data-sovereignty guarantees. The citizen-developer costs the platform almost nothing. At Power Apps that same user costs Microsoft license revenue they cannot wish away.

That is not a small asymmetry. That is a structural advantage the moment the scale is in place.

The open-source world already has pieces of the answer in place. n8n for no-code automation. Windmill for developer flows in Python, TypeScript, Go or Bash. OpenProject for workflow-oriented project work. Matrix and Mattermost for communication. They are loose puzzle pieces. What is missing is the composition layer that lets a civil servant ship a supplier onboarding app in an afternoon on the Nextcloud substrate, without code, without forks, without the next Nextcloud upgrade breaking everything.

A European civil servant who builds a leave-request app in Copilot Studio in 2027 will be impossible to pry loose in 2030. Not because the technology forbids it, only because the mindshare sits elsewhere. Mindshare is the most precious window a platform has. It is also the window that closes the fastest. We have that window now, not later.

Where Conduction is putting its weight

At Conduction we have made the strategic choice. We do not believe in a Dutch alternative next to openDesk, La Suite or Nextcloud. We believe the winning move is to help Nextcloud itself grow into the European platform it already has the potential to be. Concretely, we are putting our time and our engineering talent into two parallel tracks.

These are not loose projects. They are two sides of the same coin. A workspace without a platform stays a suite with extra apps. A platform without a workspace stays an empty room without an audience. Together they form the European counterpart to Power Platform, built on foundations we already have.

And the best part: we do not have to do this alone. ZenDiS in Germany works on comparable pieces. La Suite Numérique in France ships components we plug into. Nextcloud GmbH itself keeps building the core. Conduction delivers the Dutch contribution to a European movement, not a Dutch version of something that already exists.

Component overview

For anyone who wants to see it concretely, the stack looks like this.

This is not a suite. A suite ships in a single box. This is a platform on which an organisation builds itself, with components that are themselves vendor-independent, and an AI layer where the operator chooses the model.

Naming is load-bearing

Back to Jobs and the iPhone. The name ran ahead of the technology. The name was load-bearing. It still is.

For Nextcloud the platform reality is on the table technically. Schleswig-Holstein is running. The Bundeswehr is running. The ICC is running. RKI is running. ExApps works. The AI architecture is fundamentally different from the American alternatives. The component ecosystem is in place. The composition layer is under construction.

What is missing is that we start calling it that. Not one of the components in a sovereign bundle. Not a files replacement. Not a suite. A platform. Because as long as we keep calling it a suite, governments keep buying it as one. And in year three they stand empty-handed when the real question comes up.

As long as we keep calling it a suite, governments keep buying it as one. In year three they stand empty-handed when the real question comes up.

Ruben van der Linde, CTO Conduction

What this means for Europe

The question is not whether Europe can build a Power Platform equivalent. The question is how fast. The window is eighteen to thirty-six months before a generation of European civil servants becomes citizen-developers inside an American ecosystem. After that, the window closes. Breaking out gets exponentially more expensive. Mindshare is the one form of infrastructure you cannot retrofit.

What governments can do now

In closing

A suite bundles mail, files and office apps. A workspace adds an integrated ecosystem on top of that substrate. A platform is something else. A platform is the infrastructure on which an organisation builds its long tail, deploys its AI on its own terms, and lets its citizen-developers move without the vendor holding the strings.

That is what Nextcloud already is, only under the wrong name. 16 million users worldwide form the largest potential citizen-developer pool the European open-source world can bring to the table. When that pool gets the right composition layer, and the right AI substrate, functional parity with Microsoft Power Apps is not a dream. It is arithmetic.

At Conduction we are choosing our role deliberately. Workspace on top of the Nextcloud substrate. Platform infrastructure underneath, where developers and civil servants can build. Both in the open. Both plugging into what is already there. No separate Dutch version of something that already exists.

The choice is easy. The time is now.

Sources