Jay Subramanian
GM of Core Storage Platforms, Hitachi Vantara
March 20, 2026
Most enterprises have made the commitment to AI. They’ve approved the budgets, stood up the pilots, and named it a strategic priority. So why are 95% of them getting zero return on $30–40 billion in GenAI investment? According to MIT research cited in Hitachi Vantara’s 2025 State of Data Infrastructure Global Report — which surveyed more than 1,200 IT leaders across 15 markets — the failure isn’t the model. It’s the infrastructure underneath it.
The report puts it plainly: the divide between organizations winning with AI and those burning through budget comes down to data foundation maturity. Only 41% of organizations have reached what the research defines as Optimized status: resilient infrastructure, clean data, AI-driven operations, measurable returns. The other 59% are either building toward it or stuck in fragmented, manual environments that actively undermine their AI investments.
The question for enterprise technology leaders isn’t whether to modernize infrastructure. It’s how to understand what modernization actually requires — and why getting it right is the most strategic investment they can make right now.
For decades, enterprise storage was built around reliability and capacity, with performance optimized after the fact. Generative AI, real-time analytics, and high-frequency transaction processing have all changed that, demanding consistent sub-millisecond latency at scale.
If your storage can’t deliver data fast enough, your AI infrastructure can’t deliver insights fast enough. The State of Data report reinforces this: organizations that cite the right infrastructure to support AI workloads as a key to business success are far more likely to reach Optimized maturity — and far more likely to cite data quality as the reason their AI projects succeed (48% of Optimized organizations versus roughly 25% of Emerging ones).
The problem is structural. Annual data growth rates of 29% — a figure confirmed by 451 Research S&P Global — aren’t evenly distributed. They’re concentrated in unstructured data types such as video, sensor streams, and model training sets that are particularly demanding on storage infrastructure. Organizations need systems built to absorb that growth without performance degradation.
Hitachi Vantara Virtual Storage Platform One (VSP One) Block High End is engineered for this inflection point — delivering up to 50 million IOPS with consistent sub-millisecond latency across an all-NVMe architecture designed for the exponential demands ahead.
The State of Data report surfaces one of the starkest findings in enterprise infrastructure research: 89% of Optimized organizations operate with high availability design, regular resilience testing, and AI-driven operations.
As Chris Millington, Hitachi Vantara’s Global Solutions Lead for Data and Cyber Resilience put it: “There is no such thing as cyber resilience without infrastructure resilience.”
The stakes compound when you factor in the security landscape. Internal AI breach concerns jumped 10 points year over year among surveyed IT leaders (from 31% to 41%) — now nearly matching external AI-enabled attack concerns at 43%. Meanwhile, confidence that employees use AI safely for business purposes dropped 12 points since 2024. The threat surface has expanded dramatically, and the organizations with fragmented, manually managed infrastructure are the most exposed.
That's not a theoretical risk — it’s the operating reality for the majority of enterprises today.
VSP One addresses this directly with eight nines (99.999999%) availability, FIPS 140-3 Level 2 certified protection, and integrated cyber resilience through CyberSense-powered anomaly detection, immutable snapshots, and automated clean-data recovery.
The State of Data research identifies a pattern that shows up consistently across maturity levels: organizations stuck at Emerging or Defined status tend to rely on simplicity to operate, which limits their ability to adopt sophisticated infrastructure, expand globally, and support advanced analytics. They also pay higher long-term costs because their legacy systems can’t optimize efficiently.
Ninety-four percent of companies in the study say they need third-party help on data infrastructure. The reasons vary, but the underlying problem is consistent: complexity has outpaced internal capacity. Platform proliferation, hybrid cloud sprawl, and the skills gap are converging to create organizations that are managing infrastructure rather than deriving value from it.
Workload consolidation is part of the answer. Organizations running separate systems for open systems and mainframe workloads are paying a complexity tax on everything from licensing and management overhead to training and integration. A unified platform that handles both eliminates that overhead and frees skilled teams to focus on innovation rather than infrastructure maintenance.
Intelligent management is the other part of the solution. VSP 360, Hitachi Vantara’s unified data management platform, delivers AIOps-powered observability across the entire data estate — monitoring capacity, performance, system health, and security from a single interface accessible via SaaS, on-prem, or mobile.
The State of Data report is clear that this kind of centralized governance and automation is what separates Optimized organizations from everyone else: strategic, automated management of data location for performance, compliance, and cost.
One of the most instructive findings in the report is about decision-making structure. Less mature organizations give line-of-business managers final authority over AI use cases and priorities. Optimized organizations empower the Head of IT to own ROI and investment decisions for AI, choose which models to use, and define use case priorities.
This reflects a fundamental difference in how governance is treated as a constraint versus a capability. Optimized organizations have embedded governance into their operations, moving from ad hoc data location awareness to strategic, automated management across the entire data estate. That shift unlocks speed, scale, and analytics for operational data while maintaining control, isolation, and auditability for sensitive data.
The research makes the cost of getting this wrong concrete. In fact, 46% of failed AI projects cite insufficient data as a cause, 41% point to organizational silos, and 29% name infrastructure limits. These figures represent governance failures masquerading as technology failures — and they’re preventable with the right foundation.
Hitachi Vantara’s data governance capabilities and VSP 360 are built around this reality — automating policy enforcement, ensuring compliance across hybrid environments, and giving IT leaders the visibility to govern data throughout its complete lifecycle.
Octavian Tanase, Chief Product Officer at Hitachi Vantara, framed it clearly in the State of Data report: “In order to be relevant in this market, you have to act with agility in a way many companies have not experienced before. Time to market, constant innovation and the reality that no one vendor can do it all are critical.”
That’s the infrastructure conversation that matters right now. Not which AI model to buy, but whether the foundation underneath it is built to deliver. Maturity is not correlated with budget, company size, or data volume. It’s a strategic choice driven by leadership prioritization.
The organizations that treat storage modernization and data infrastructure investment as a strategic imperative — rather than a maintenance cycle — will find themselves with a meaningful advantage as AI becomes the primary driver of enterprise differentiation. Those that wait will find the gap increasingly difficult to close.
Learn more about VSP One Block High End and download the 2025 State of Data Infrastructure Global Report to see where your organization stands — and what it takes to move forward.
Jay Subramanian is General Manager, Core Storage Platform, at Hitachi Vantara
