The enterprise world is entering a new stage of artificial intelligence maturity. It’s a stage defined not by flashy pilot projects or limited experimentation, but by a rapid integration into business’ core operations at enterprise scale. But as that push for scale continues, enterprise leaders are increasingly confronted with a difficult reality—traditional data architectures are overdue for an AI-oriented refresh. And that realization has begun ushering in a Great AI Re-Architecture, where enterprise IT leaders are reshaping their data architectures for an AI-driven future.
To better understand how organizations are navigating this new frontier, Cloudera released its latest report, The Great AI Re-Architecture, surveying 1,500 enterprise architects, cloud infrastructure leaders, and data architects worldwide. The results are clear–AI is completely redefining how organizations structure their data architecture. In fact, 72% of those surveyed said their current data architecture requires a significant overhaul to meet their AI goals.
So, as enterprise leaders push for deeper AI integration and greater business impact, how are data architectures evolving to meet the moment? Let’s explore some of the key findings.
The Beginnings of a Great Re-Architecture
Considering nearly three-quarters (72%) of respondents said their current architecture requires a significant overhaul to support future AI goals, it’s clear the enterprise world is well aware of what’s needed. And another 75% said they have already moderately or significantly changed their data storage and architecture practices.
AI workloads demand seamless access to data that often resides across multiple environments, whether that’s on-premises, in public clouds, private and sovereign clouds, or spread across edge systems. Ensuring all of that data is managed properly and ready to bring to AI is costly. In fact, 84% of respondents said that AI workloads have caused their infrastructure costs to increase.
While enterprises generally are all feeling the pressure to overhaul their data architectures and capitalize on AI, the driving motivators behind those efforts reveal some variance. Forty-two percent of respondents pointed to data security, governance, and compliance requirements as the force behind changes to data storage and architecture practices. This was followed by 35% who were focused on improving performance and reducing latency, 35% who said real-time or edge-based AI capabilities, and 33% who said scaling AI initiatives across their business.
Whatever the reason, many organizations have come face to face with the fact that even if they are ready to leverage AI, their infrastructure simply is not.
An Emphasis on Governance and Trusted Data
The race to scale AI has placed renewed importance on governance and security—its positioning elevated from a matter of checkboxes on a compliance to-do list and instead becoming an AI enabler. As respondents know all too well, governance issues can be a serious hindrance to successful AI initiatives. In fact, nearly every organization surveyed (95%) reported delaying or canceling AI projects over the past year because of governance or compliance issues.
As AI finds its way into core business processes, the data it relies on must be trusted. That means enterprise leaders need to know where their data is stored, who can access it, and whether it can be used responsibly in complex environments. Without that visibility and control, organizations risk slowing AI adoption or introducing unnecessary risk into their AI efforts.
The findings expose a clear shift in the role of governance as an enabler. But keeping pace with the inherent change that comes with AI requires a unified data foundation that ensures control over 100% of organizational data, wherever it resides.
An AI-led Future Thrives on Flexibility
A notable trend in Cloudera’s report is that organizations are taking a more intentional approach to where AI workloads run. And many are even repatriating workloads away from public cloud environments. Among respondents, 66% said they had moved AI workloads from public cloud back to on-premises or private cloud environments in the last 12 months.
That number is not an indicator of any failure on the part of public clouds. Rather, it shows a preference for a hybrid approach rather than a cloud-first one. In fact, nearly one-third of respondents (31%) said they are running AI inference workloads in hybrid settings. Though this was followed closely by public cloud (30%) and private cloud (22%). Ultimately, what this shows is a preference to bring AI workloads to the environments that enable them to deliver the best results.
But what does a hybrid-oriented future mean for the plans of IT leaders? Over the next two years, roughly one-third (29%) of respondents said they still planned to increase cloud spend, but at the same time, one-quarter (25%) said they planned to place more emphasis on a hybrid approach.
Wherever AI heads, the future is sure to be much more mixed, as organizations seek data architectures that enable the most impactful business results. But that shift requires the flexibility to run workloads in the right place without sacrificing visibility, governance, performance, and cost control across the entire data landscape.
A New Foundation for Enterprise AI
AI has brought about the next phase in data architecture—a great AI re-architecture that will redefine the way every business approaches data and AI workloads. As AI makes its way deeper into core enterprise systems and continues to spread at scale, critical gaps must be addressed to deliver real, lasting value.
The enterprises that act with urgency and embrace a data architecture built with the flexibility, visibility, and control to support AI at scale will build a distinct competitive advantage for the future.
Explore the full survey report and discover how organizations are rethinking data architecture for the next era of AI.
