Enterprise cloud strategy is undergoing its most significant recalibration since the shift to public cloud began over a decade ago. A recent survey found that 66% of respondents moved AI workloads from public cloud back to on-premises or private cloud infrastructure within the past 12 months. This is not a retreat from cloud computing or artificial intelligence — it is a maturation of digital strategy, driven by cost discipline, data governance, and regulatory exposure. For CFOs, General Counsel, and CTOs steering digital transformation programmes, this shift carries direct implications for capital allocation, vendor contracts, and compliance posture.

The Reversal: From Public-Cloud-First to Hybrid by Design

For much of the past decade, cloud migration strategy defaulted to public cloud. AI adoption in enterprise environments has disrupted that default. Training and inference workloads involving proprietary data, customer records, or regulated information are increasingly returning to private or on-premises environments, where organizations retain full control over data residency, latency, and audit trails. ISG’s research into the German market confirms this trend is not confined to global tech giants: mid-market industrial and financial services firms are pairing cloud modernization initiatives with explicit AI deployment and data-control requirements, effectively designing hybrid architectures from the outset rather than retrofitting them after a public-cloud rollout.

This is a rational response to three converging pressures: unpredictable public cloud AI compute costs, heightened scrutiny of cross-border data flows under GDPR, and the compliance obligations introduced by the EU AI Act, which enters into force in stages through 2026 and imposes documentation, risk-classification, and human-oversight requirements on high-risk AI systems. Sovereign cloud — infrastructure guaranteeing that data and processing remain within EU jurisdiction — has moved from a niche procurement preference to a board-level requirement.

Vendor Response: Multi-Cloud as Risk Mitigation, Not Just Scale

Major technology vendors are adapting their commercial strategies accordingly. Verizon’s adoption of Google Cloud’s full-stack AI, including Gemini Enterprise, illustrates how large enterprises are scaling AI while retaining architectural flexibility across cloud environments. Oracle and AWS have expanded Oracle AI Database@AWS to 22 regions across Asia Pacific, Europe, and the Americas — a clear signal that multi-cloud interoperability, not single-vendor lock-in, is now the expected baseline for enterprise digital strategy. Meanwhile, SAP reported a 26% rise in its European cloud backlog, as core finance, procurement, supply-chain, and HR systems continue migrating to platforms explicitly designed to support both AI deployment and regulatory compliance.

Taken together, these developments show that innovation management in large organizations increasingly means orchestrating a portfolio of cloud and AI providers, rather than betting on a single hyperscaler relationship. This has direct consequences for contract negotiation, exit clauses, and data portability provisions — areas where legal and procurement teams must now engage earlier in the technology selection process.

Implications for Business Leaders

Boards and executive teams evaluating digital transformation roadmaps should consider the following:

  • Reassess total cost of ownership for AI workloads, factoring in repatriation costs if workloads later move from public to private infrastructure.
  • Map data flows against the EU AI Act and GDPR before finalizing AI deployment architecture, particularly for high-risk use cases in finance, healthcare, or HR.
  • Negotiate multi-cloud and exit provisions into vendor contracts now, rather than after workloads are embedded in a single environment.
  • Align M&A due diligence checklists with target companies’ cloud and AI architecture, since sovereignty and compliance gaps can materially affect post-merger integration costs.
  • Establish cross-functional governance linking CTO, General Counsel, and CFO functions to oversee emerging technology decisions with regulatory and financial consequences.

Key Takeaway

The shift toward hybrid and sovereign cloud is not a rejection of cloud migration or AI adoption in enterprise operations — it is a more disciplined, compliance-aware phase of digital strategy. Organizations that treat cloud architecture and AI governance as a single strategic decision, rather than sequential IT projects, will be better positioned to scale AI responsibly while managing regulatory and financial risk across European and global operations.