AI is reshaping nearly every part of the enterprise, from how employees work to how organizations compete. Increasingly, it’s also reshaping the data architectures that underpin those efforts.
New global research from Cloudera suggests organizations are reassessing the data architectures that support AI. As AI expands beyond pilot projects and into core operations, enterprise leaders are reevaluating the architectural decisions that will shape how AI scales across the business.
The Great AI Re-Architecture captures a broader shift in enterprise AI: AI has reached a point where enterprise data architecture has become the determining factor in whether organizations can scale it. Organizations are redesigning their data architectures to support the operational demands of AI, from governance to increasingly distributed workloads.
AI Is Reshaping Enterprise Infrastructure
Many organizations are reaching the same conclusion: their existing data architecture has become a limiting factor.
What’s driving that reassessment extends well beyond compute. Security, governance and compliance rank as the leading driver of architectural change (42%), ahead of improving performance (35%) and scaling AI initiatives across the business (33%). Meanwhile, 84% of respondents report that AI workloads have increased infrastructure costs. The survey suggests enterprise AI is changing the priorities behind infrastructure decisions. Security, governance and compliance now outweigh performance as the leading driver of architectural change, emphasizing how closely AI strategy has become tied to data protection and regulatory requirements.
“Security and compliance increasingly shape the way organizations design their AI architectures,” said Sergio Gago, CTO at Cloudera. “Enterprise leaders want to expand AI without creating new risks for their most sensitive data.”
Supporting enterprise AI increasingly requires architectures that can evolve as business requirements change while protecting enterprise data wherever it resides. The architectural decisions organizations make today will influence how easily they can expand AI as technologies and business priorities continue to evolve.
Governance Is A Driving Force For Scaling AI
Governance is becoming a defining factor in how quickly new initiatives move forward. Nearly all respondents (95%) say they delayed or canceled AI projects in the past year because of data governance, compliance or regulatory complexity, while more than half (55%) report delaying six or more projects.
The survey also suggests that governance is becoming more difficult as AI becomes more deeply embedded across the enterprise. Nearly three-quarters (73%) of respondents say AI integration has made data governance more complex and harder to maintain as organizations manage data across increasingly distributed environments.
“AI changes the scope of governance,” said Gago. “As organizations expand AI across the business, they need a consistent approach to managing data regardless of where it resides. That consistency gives organizations the confidence to scale AI without losing control.”
Governance challenges increasingly intersect with security. As enterprise data becomes more distributed, organizations need a consistent approach to protecting sensitive information across every environment where AI operates. As enterprise data becomes more distributed, organizations must maintain consistent policies across environments without slowing AI development. That balancing act helps explain why governance is emerging as one of the defining operational considerations for enterprise AI.
For many organizations, governance has become a prerequisite for enterprise AI. The ability to apply consistent policies across increasingly distributed data environments can determine whether AI initiatives continue to expand or stall before reaching production.
AI Workloads Are Becoming More Distributed
Enterprise AI is prompting organizations to take a more deliberate approach to workload placement. Rather than relying on a single infrastructure model, they’re evaluating where AI workloads will deliver the greatest value.
One of the clearest signs of this shift is workload mobility. Two-thirds (66%) of respondents say their organizations have moved AI workloads from public cloud back to on-premises or private cloud environments within the past year. At the same time, hybrid environments have emerged as the most common location for AI inference workloads, narrowly ahead of public cloud. Looking ahead, planned investments span public cloud, on-premises, edge and hybrid architectures, showing how organizations are expanding their options rather than converging on a single infrastructure model.
“Organizations want the flexibility to run AI wherever it delivers the greatest value,” Gago said. “That requires an architecture that supports changing workload requirements without forcing a fundamental redesign every time priorities shift.”
That flexibility allows organizations to align AI workloads with business requirements instead of infrastructure limitations. Some AI workloads prioritize performance and low latency, while others are constrained by where sensitive or regulated data can be stored and processed. Those competing requirements are encouraging organizations to design architectures that support multiple deployment models without sacrificing security or governance.
Enterprise AI will continue to evolve, and the infrastructure supporting it has to evolve with it. Organizations are designing architectures that let AI workloads move as business needs change without requiring a fundamental redesign.
Building The Foundation For Enterprise AI
The Great AI Re-Architecture is already underway. Organizations are redesigning the data architectures that support AI to meet the demands of enterprise-scale deployment.
The gap between AI leaders and everyone else will increasingly be shaped by architecture. Organizations that build data architectures capable of bringing AI to their data, wherever it resides, will be better positioned to scale AI as the technology continues to evolve.
Learn how Cloudera is helping the world’s leading enterprises build the data architectures needed to scale AI.