Enterprises worldwide are hitting a wall in their AI journey not because they lack appetite, but because the data architecture beneath their AI ambitions wasn’t built for this moment. That’s the core finding of Cloudera’s new global study, “The Great AI Re-Architecture,” based on responses from 1,500 enterprise architects, cloud infrastructure leads, and data architects.
Adoption is high, but so is friction
While 77% of organisations are actively using AI, a striking 95% have delayed or cancelled AI initiatives in the past year, largely due to data governance, compliance, or regulatory hurdles. Nearly three-quarters (72%) say their existing data architecture needs a significant overhaul to keep pace with AI’s demands a clear signal that today’s infrastructure is buckling under new pressures.
Governance emerges as the real bottleneck
Governance is no longer a back-office concern; it’s foundational to AI success. The report finds 73% of respondents say AI has made governance more complex, and 55% have delayed or scrapped more than six AI projects in the past year over compliance concerns. Adding to the complexity, 97% of organisations move data across environments at least monthly, making consistent governance across cloud, on-premises, and edge setups increasingly difficult.
Hybrid is becoming the default, not the exception
In response, enterprises are rethinking where and how AI workloads run. Two-thirds (66%) have shifted AI workloads from public cloud back to private cloud or on-premises infrastructure over the past year, while a quarter plan to prioritise hybrid-first architectures over the next two years. Meanwhile, 84% report rising infrastructure costs tied directly to AI workloads.
The takeaway is clear: simply adopting AI is no longer enough. The organisations that will pull ahead are those rebuilding their data foundations with governance, flexibility, and hybrid infrastructure at the core, so they can run trusted AI on trusted data wherever it resides.