Storage growth isn’t the problem. Unstructured data economics are 

Recently, we sat down with the editor of Blocks & Files, Chris Mellor, to discuss what we argue is the most urgent storage challenge facing enterprises today: namely, that the continual increase in data storage capacity is becoming economically and operationally unsustainable. 

The rationale is straightforward. Firstly, the rise of machine-generated data and AI workloads has changed the economics surrounding unstructured data. Storage spend is being swallowed by the need to retain unprecedented volumes of unstructured data, which is totally at odds with the assumption that growth can always be accommodated by simply adding more capacity. That assumption looks shakier than ever right now, with AI demand driving up storage prices — a squeeze expected to persist into 2027, and stretching procurement lead times. But even when prices normalize, the underlying problem remains: organizations cannot buy their way out of unmanaged data growth. 

This is not a temporary challenge. It is a structural problem, and it is growing more severe as three forces converge simultaneously. Unstructured data continues to grow faster than organizations can understand or govern it. Cyber risk escalates in parallel, as expanding data volumes widen the attack surface and make it harder to identify sensitive data or limit breach impact. And AI initiatives, the very initiatives organizations are counting on to drive competitive advantage, depend on trusted, accessible, and well-governed data that, in most enterprises, simply does not yet exist in that form. 

Secondly, in many environments, this challenge is compounded because organizations address it from the wrong perspective. Storage strategy is no longer a question of how much data exists, but of what value that data holds and how it should be managed. The default posture of storing everything indefinitely is no longer a sensible option, not operationally, not financially, and increasingly not even practically, as capacity itself becomes harder and costlier to secure. Treating retention as the path of least resistance is now a decision with real consequences, not a neutral default. 

This moves the conversation away from storage devices and towards the economics and lifecycle of the data itself. For Datadobi and our customers, this is an increasingly common and important discussion as the emphasis shifts from infrastructure to intelligence. In our experience, organizations are increasingly recognizing that understanding their data assets – is becoming just as important as the infrastructure used to store them. 

More capacity can no longer be the default answer  

Exponential growth in unstructured data has exposed the limits of relying solely on infrastructure decisions. 

As the Blocks & Files article explains, when storage capacity was the easiest problem to solve, few organizations had (or indeed needed) tools to understand what data they possessed or how actively it was being used. Today, however, this distinction is front and center because understanding the economics of data requires insight into the data itself, not simply the infrastructure supporting it. 

Our position, and that of our customers worldwide, is that this shift represents a significant opportunity. Organizations with a clear understanding of their data estates will be far better placed to control costs and derive greater value from the information they already possess. 

“The answer is not simply more capacity. The answer is visibility, intelligence, and orchestration. Organizations cannot buy their way out of unstructured data growth. Adding more infrastructure may relieve short-term pressure, but it does not solve the underlying problem: data remains fragmented, unmanaged, duplicated, and poorly aligned to business value. And even for organizations with the luxury of expanding capacity at will, the true cost extends far beyond the storage itself. Every terabyte retained must also be protected, backed up, snapshotted, replicated, multiplying the footprint and the spend. Worse, every terabyte retained is a terabyte exposed: in the event of a breach, unmanaged data is not just an overhead but a liability, expanding the attack surface and the potential blast radius of a cyber incident.” 

In practice, that means doing three things. First, organizations need to discover — transforming fragmented data into actionable intelligence by gaining visibility into what data exists, where it resides, who owns it, and the value and risk it represents. Second, they need to align — matching data with the right platforms, tiers, and access requirements based on business value rather than default retention. Not everything belongs on high-performance infrastructure, and the cost and risk of treating it as though it does is becoming untenable. Third, they need to operationalize — turning insight into action through policy-driven orchestration that automates data lifecycle decisions, strengthens governance, and ensures data remains aligned with compliance, sovereignty, and business requirements over time. 

Critically, this is not simply a cost reduction exercise. Organizations that take this approach extend the life of existing assets, defer unnecessary capital expenditure, reduce data exposure, improve governance, and make trusted data available for AI and future business initiatives. The business outcome is far broader than budget savings alone. 

The bottom line is that data growth will continue, and AI  isonly accelerating the trend. The objective is not to stop accumulating data or even limit how much is retained, but to understand the information that already exists and manage it much more intelligently. This is not a one-time cleanup exercise. It is an ongoing discipline — and it is why we believe the Intelligence and Orchestration Layer for Unstructured Data is becoming one of the most important strategic capabilities an enterprise can develop. The organizations that get this right will be better positioned to control cost, reduce risk, and unlock the full potential of their data for AI and innovation. Those that don’t will continue adding infrastructure to a problem that infrastructure alone cannot solve. 

Chris explores these themes in greater depth in his Blocks & Files article. You can read the full piece here.