Hybrid is not a transition state

Every architecture diagram we see this year has an arrow on it. The on-premise systems sit on the left, the cloud sits on the right, and the arrow promises that the left side is temporary. The estate is described as "in transition," which grants everyone permission to treat the current state as not worth investing in.

It is time to say the quiet part: for most European enterprises, the arrow is decorative. The mainframe that clears your payments, the ERP that has been customized for fifteen years, the systems pinned in place by data residency or by the simple fact that they work, these are not late for the cloud. They are staying. Hybrid is not the journey; for the planning horizon that matters, it is the destination.

The cost of pretending it's temporary

Calling the current state a transition sounds harmless, but it quietly distorts every decision that follows. Anything labeled temporary does not get invested in, because why would you improve something you are about to replace? So the integration between on-premise and cloud, the part that will actually be load-bearing for the next decade, is built as scaffolding: quick, tactical, undocumented, and owned by no one in particular.

The result is an estate that runs its most critical connections on its least considered infrastructure. The systems on both ends are well-governed; the bridge between them, which everything crosses, is the neglected part. This is precisely backwards, and it stays backwards for as long as the organization tells itself the bridge is about to become unnecessary. It is not, and the pretense is what keeps it fragile.

What changes when you admit it

• The seams get investment: if on-premise and cloud will coexist for a decade, the integration between them is permanent infrastructure, not scaffolding, and deserves to be governed like it.
• Operations converge: one way to publish, secure, and observe an API, wherever it runs, instead of a modern discipline for the cloud and archaeology for everything else.
• Modernization gets honest: systems move when there is a business case, not because the diagram is embarrassing. Some cases will never close, and that is architecture, not failure.

An estate planned as "temporarily hybrid" gets two half-built operating models. An estate planned as hybrid gets one that works.

One control plane for the whole estate

The practical consequence is a control plane that spans the whole estate, gateways where the workloads are, governance that does not care about the hosting. The point is not to make the mainframe feel modern; it is to stop maintaining two entirely separate ways of running things, one for the systems you are proud of and one for the systems you are quietly ashamed of but depend on completely.

Our customers who made that shift describe the same relief: the estate stopped being a migration behind schedule and became, simply, the estate. Once governance, security, and observability apply uniformly wherever a workload runs, the question of where it runs becomes a business decision rather than an operational burden. The arrow can stay on the diagram. Just stop waiting for it.