Why Melious AI moved its CDN and DNS to Gcore: a founder conversation about sovereign AI in Europe

For many startups, infrastructure decisions are mostly about performance, pricing, and developer experience. For Melious AI, they are also about trust.

Melious AI is a German startup building a European AI platform around privacy, transparency, sustainability, and digital sovereignty. Its core product gives companies an OpenAI- and Anthropic-compatible endpoint, so teams can switch by changing the base URL rather than rebuilding their applications from scratch. Behind that endpoint, Melious connects open-weight models, European infrastructure providers, and additional tools such as web search and code execution in a sovereign environment.

We spoke with Dean Fünffrock, Co-Founder & CPO, and Marc Hornbacher, Co-Founder & CMO, about why they started Melious, what sovereignty means beyond a marketing slogan, and why the team decided to move its CDN and DNS layer to Gcore.

"We make European AI infrastructure actually usable"

Dima Maslennikov, Gcore: Dean, Marc, thanks for joining. Let's start with a short introduction. Who are you, and what is Melious AI building?

Dean Fünffrock, Melious AI: I'm Dean, Co-Founder and CPO at Melious AI. We are a startup based in Germany, and what we build is a way to connect the software and infrastructure that already exists in Europe and make it usable and accessible for companies.

At the core, we provide an inference service. It works through an endpoint that is compatible with what many developers already use with OpenAI or Anthropic. In practice, that means companies can switch quite easily by changing the base URL.

But we are not only building around technical compatibility. Our values are privacy first, sustainability, and transparency. We try to work with data centers that mostly use green energy, and where that is not fully possible, we compensate through projects. We also want users to know what they are getting: which model was used, which provider was used, what energy was consumed, and what CO₂ was estimated for a request.

On top of the API, we also have a chat interface so that people can use the product without coding. And we are expanding with tools that AI applications often need, such as web search and code execution. The important part is that these tools are also hosted on sovereign infrastructure. We do not want to rely on Google APIs or Brave APIs for that.

Dima: And your role specifically?

Dean: I'm CPO, but at our stage, that means a lot of things. Right now, I mainly work on front-end development and talk to customers about what they need, what features matter, what does not matter, and where the product has the biggest potential.

Marc Hornbacher, Melious AI: I'm Marc, Co-Founder and Chief Marketing Officer. I take care of the marketing side at Melious. I'm also still at university and finishing my bachelor's degree this semester. After the summer, Melious will be my full focus.

From a privacy problem to a company

Dima: How did the company start?

Dean: The project started before the company was officially founded. I worked with our CEO, Simon, at another company in our city. He was building a solution that needed AI integration, but it had to be really private. There was no solution where we could be fully sure that no data was exposed to US authorities, logged, or used for training.

That was the beginning. We started with brainstorming sessions, met in the city, talked about the idea, and later brought in Tom, our COO. Simon, Tom, and I went to the same high school, which was also quite technical, so the idea resonated with us. Marc and I knew each other from making music a few years ago. That is how the team came together.

We founded the company in September 2025. Since then, we have been building the product and raising money.

Dima: Where are you now in terms of traction?

Dean: We have a live product with over 300 registered users, and we've seen especially strong growth recently — driven in part by events like the Fable shutdown pushing teams toward sovereign infrastructure.

Dima: What has been the main acquisition channel so far?

Marc: LinkedIn and events. A lot of it comes from talking to people in person. We are also about to start Google Ads, but until now it has mostly been LinkedIn and direct conversations at events.

"If you work with sensitive data, you should care where it goes"

Dima: Let's step back to the problem. What is the core issue you are solving?

Dean: AI is one of the biggest topics right now, and it will be integrated into almost everything. But if you deal with sensitive data, you really need to care about where that data goes, who is involved in processing it, and what happens to what you put in.

With many established solutions, companies face uncertainty. Data may be processed in environments where they do not fully control legal exposure. For example, there are questions around whether data can be accessed under the US Cloud Act if US-owned cloud companies are involved.

For us, that is the problem: companies want to use AI, but they need to know what happens to their data and who has control over the infrastructure.

Dima: Marc, who are the customers with this problem?

Marc: Basically, every company has to think about privacy and data. But some industries feel it much more strongly. Medtech, fintech, defense tech, govtech — these are obvious examples. Any company that works with sensitive data and cannot compromise on privacy is relevant.

At the same time, we currently focus a lot on startups because they are easier to reach at our stage. We can talk to them faster, learn faster, and understand where the product fits best.

Dean: The challenge with industries like medtech is that there are many compliance requirements. As a startup, you cannot just rush into that. It will take time. But in the long run, those areas are exactly where a solution like ours can have the biggest impact.

Sovereignty is not one thing. It has layers.

Dima: Sovereignty is a hot topic, but it is also becoming a marketing keyword. What does sovereignty actually mean for you?

Dean: For us, in AI, sovereignty has three layers.

The first layer is the model provider or the lab that trains the model. The second layer is the hardware the model runs on. The third layer is the entity that operates the hardware, provides the service, and controls the data center.

The part we are solving is control. We want to know what is happening in the infrastructure we use, and we want to be able to tell our customers what is happening to their data.

That is the core of sovereignty for us. It is not just about saying "European." It is about being able to understand and control the chain.

Why Melious AI moved CDN and DNS to Gcore

Dima: And that is also why you decided to move to Gcore?

Dean: Yes. We needed to make sure there was no US entity involved in the chain.

So we did some research and checked whether there was a provider that could give us what we needed. We found Gcore, and it provided everything we needed.

Dima: So for you, it was not only a technical infrastructure decision. It was part of the sovereignty requirement.

Dean: Exactly.

What comes next for Melious AI

Dima: What are your plans now? Are you looking for customers, investors, partners?

Dean: Our primary goal is to reach as many potential customers as possible who struggle with privacy and work with data where privacy really matters. We are looking for companies that do not want to compromise, that want to know what happens to their data, and that want to be sure it is safe.

Fundraising is a topic for us. We are well set for the moment, but we are preparing for a larger seed round toward the end of the year, and we are always open to discussions with the right partners.

What we really want is customers, feedback, and connections — especially into government organizations. We already got some insights into healthtech, but govtech would be very interesting for us as well.

The main call to action is simple: try our service. See how easy it is to switch. Give us feedback if something is not as expected, if you need a feature, or if there is a use case we should understand better.

The infrastructure question behind European AI

Melious AI is building in a market where "sovereign AI" can easily become a slogan. But for the team, the concept is concrete: who trains the model, where it runs, who operates the infrastructure, and under which legal framework the data is processed.

That is why the choice of Gcore for CDN and DNS was not just a backend decision.

It was part of the same product logic that shapes Melious itself: if companies are expected to trust AI with sensitive data, the infrastructure behind that AI has to be explainable, controllable, and aligned with European requirements.

For startups, this is a reminder that infrastructure is no longer invisible. In AI, it is part of the product, part of the compliance story, and part of the trust users are being asked to place in the technology.

Gcore offers global infrastructure built for exactly this kind of challenge. If you're a startup that can't afford ambiguity in your infrastructure stack, explore what's possible at gcore.com/infrastructure-for-startups.