Richard van Anholt, co-founder of GLBNXT, recently sat down with Dibran Mulder, CTO of Caesar Groep, for an informal chat over coffee. It started with a practical question: how do you get AI to do something useful inside the software a company already runs? From there it drifted to proofs of concept, procurement and where European cloud fits in.

Dibran talks about this as someone who builds it. He began as a developer and solution architect at Caesar Experts, then spent several years at Cloud Republic working on Azure and cloud development, where he also contributed to the company’s Dutch tech podcast. Today he is CTO of Caesar Groep, a Dutch IT group founded in 1993 with a long track record in software development and data analytics. He is also CTO of Yivi, the open source digital identity wallet whose development and operations Caesar took over in late 2024. He is level-headed about AI, and just as excited about what it is bringing now and what is still to come.

His starting point was simple. Most corporate data already lives in applications people use every day, such as CRMs and internal tools. AI becomes interesting when it works with that data where it sits instead of asking anyone to move it. So his advice for platforms like GLBNXT is to expect developers to keep asking for easy connections to the datastores and enterprise software they already run. In his view, that matters more than any single model or feature.

GLBNXT fits that picture for him. He described its building blocks as a good starting point for getting proofs of concept into the market quickly. What stayed with him was less the speed and more how the platform responded. When something was missing, like Neo4j or Jupyter Notebook, GLBNXT added it on request. That saved his team time and budget they would otherwise have spent building or integrating those parts themselves, and he was pleased with how neatly they landed in the platform.

Neo4j deserves a closer look. In Dibran’s view, you don’t make an organization’s information available to AI by breaking it into chunks. You do it by establishing an “information space”, and a graph database is a useful tool for that. It keeps the relationships between things, like which contract belongs to which client and which project that client runs, so AI can work with the context around the data and not only with loose pieces of text. For the use cases Dibran was testing, that made a real difference.

Sovereignty came up too, but as background, not as the main topic. Richard has noticed that the conversation about European cloud versus the big US hyperscalers has shifted. Dibran agreed, yet pointed out that the Netherlands still relies on US cloud more than Germany or France does. People say they want to move. Few actually do. Both agreed that the risk of non-European governments reaching European data is well known, and that in most meetings it gets acknowledged and then waved away.

Dibran thinks the obstacle is rarely technical. It is procurement. Approving a new vendor takes so long that purchasing teams often stay with whoever is already on file. Nobody decides data privacy is unimportant. The easier route simply wins. He has seen European alternatives hold up in practice and named Scaleway as a good sovereign cloud next to GLBNXT.

On GLBNXT itself, Dibran was upbeat without overselling it. “A platform like GLBNXT, one you can trust and that’s ready for enterprise use, is where I see this going,” he said. “I’m comfortable betting on that future.” The crowded market doesn’t worry him: “Competition keeps everyone sharp. What matters is who delivers, and so far they have.”

GLBNXT builds EU-hosted, sovereign AI infrastructure for regulated organizations across Europe. www.glbnxt.com