SAP’s new EU AI Cloud approach is an example of how seriously large vendors are taking AI sovereignty in Europe. Working with partners such as Cohere, SAP is packaging agent-style and multimodal AI models so that customers with strict data residency requirements can still adopt modern AI. These capabilities are exposed through SAP Business Technology Platform, giving enterprises more control over performance and where workloads actually run.

A key theme is deployment choice. EU AI Cloud is built on top of SAP Sovereign Cloud, which lets customers pick different levels of control. Some will run on SAP’s own infrastructure in European data centers; others may use a sovereign on-site model where SAP manages the stack inside the customer’s facilities. There are also options for mixing this with selected hyperscalers or region-specific offerings like Delos Cloud in Germany. The common thread is keeping operational control and data locality aligned with European regulatory expectations.

For teams building AI-heavy systems, this highlights a broader pattern: architecture and vendor choices now have to account for data jurisdiction as a first-class constraint. Sovereign cloud offerings won’t replace mainstream public cloud, but they are becoming an important tool for organisations operating under EU privacy, retention and access rules — especially those in the public sector, healthcare, and finance.