Rotterdam Puts Drupal’s Biggest Bets to the Test

Perhaps the most revealing thing about DrupalCon Rotterdam is that its programme does not tell one tidy story. Drupal is preparing to discuss its transformation into a production-ready Agentic CMS while another session asks, quite directly, why developers still do not choose Drupal. Canvas has become central to Drupal CMS, yet Rotterdam will also put it against Display Builder in a real-world project. Those collisions matter because Drupal’s next phase depends less on any one initiative succeeding than on whether these different bets can reinforce one another without carrying old friction forward.

Connecting Drupal to models and agents is no longer the difficult part of the AI story. Rotterdam moves quickly into hallucination, reviewer trust, orchestration, governed workflows and human control because production use raises harder questions about what an agent may do once it can act rather than merely answer. Permissions, auditability, validation and human approval are becoming part of the product itself. If the Agentic CMS idea is going to mean more than a collection of AI capabilities, those boundaries have to become as deliberate as the capabilities they constrain.

Adoption tests another part of the same argument. Drupal CMS 2.0, Canvas, Recipes and AI-assisted tooling have changed the product presented to builders, but improved technology does not automatically change how Drupal is perceived beyond its existing community. “Why Developers Don’t Choose Drupal” puts that gap unusually plainly. The useful evidence from Rotterdam will not be another inventory of features, but signs that newcomers can enter the system more easily, that organisations are choosing Drupal CMS rather than only inheriting or upgrading Drupal, and that the new experience removes barriers people outside the community can actually feel.

Digital sovereignty makes the question more demanding because Drupal is also being asked to apply its values inward. A Rotterdam discussion on FAIR asks whether sovereignty, governance and cost problems in Drupal’s own software distribution could be addressed differently, starting from Drupal.org’s role as the ecosystem’s central distribution point. That turns openness from something Drupal offers its users into something its own infrastructure may need to demonstrate. Arguments about control become more credible when the project is willing to examine where dependence, responsibility and operational burden sit inside its own systems.

Discovery may be where all of these threads become visible from outside. A Rotterdam discussion on Drupal’s representation in ChatGPT, Gemini and Claude starts from the recognition that developers and decision-makers increasingly encounter technologies through AI-generated answers as well as conventional search. Drupal therefore has to become easier to understand, evaluate and trust not only for people already inside the ecosystem, but also for the systems mediating their choices. Rotterdam will not settle all of that in four days. What it can reveal is whether Drupal’s current direction is beginning to cohere: a product easier to adopt, agents with clearer limits, infrastructure consistent with its sovereignty claims, and a project that can explain its relevance wherever technology choices now begin.

Follow The DropTimes on LinkedIn, X, Bluesky, and Facebook, or join #thedroptimes on Drupal Slack.

Kazima Abbas wrote and curated this issue of Editor’s Pick.