AI's next phase hinges on data, governance & infra

Hitachi Vantara executives warn that the next phase of artificial intelligence adoption will depend more on data quality, governance and infrastructure than on ever-larger models. The shift comes as enterprises move from early AI experiments to broader deployment across core business processes.

Corporate AI strategies are now under pressure from regulators, communities and boards, senior leaders at the data storage and management company said. Attention is moving away from headline-grabbing models and toward the less visible foundations of how data is stored, moved and secured.

Simon Ninan, Senior Vice President of Business Strategy at Hitachi Vantara, said many customers have already trialled generative AI and agentic systems. Those projects have exposed weaknesses in data pipelines and governance structures that struggle under real-world conditions.

"The headroom is lower on large language models: better models won't necessarily drive more powerful business outcomes. Rather, the next few months are going to see increased attention towards data governance, security, sovereignty, and responsible AI approaches. Governance must get to the heart of data value by addressing the foundations of how data is created and managed throughout its lifecycle. Security is an increasingly urgent concern, as agentic AI comes with significant risks both from AI-empowered malicious actors as well as autonomous systems that can make enterprise decisions with minimal human oversight. Sovereignty, or the need to manage data within strict local boundaries, has become the top requirement in geographies such as Europe that are trying to contain both geopolitical and cyber risks. And, responsible AI is about refocusing AI approaches and narratives in the face of rising community opposition to AI and data centers; what is needed is a greater emphasis on resource-efficiency, ethical operations, transparency, and ecosystems that can deliver net-positive outcomes for society at large. These are fundamentals, but it is these foundations that will eventually turn into market access, investor confidence, and the ability to move faster than competitors," said Simon Ninan, Senior Vice President of Business Strategy at Hitachi Vantara.

Ninan's comments reflect a wider industry view that performance gains from frontier models are delivering diminishing returns for many enterprise use cases. Companies instead face growing scrutiny over how they govern sensitive data, comply with localisation rules and explain automated decisions.

Energy use is also emerging as a central concern as AI training and inference workloads expand. Hitachi Vantara Chief Executive Officer Akinobu Shimada framed infrastructure choices as a key factor in whether AI growth can align with corporate sustainability commitments.

"As organizations embed AI across more business processes, the volume of data, compute, and infrastructure demands can dramatically increase energy consumption and resource use. As a result, the future of AI depends not only on smarter models, but on smarter infrastructure that makes AI possible in a way that is more energy- and environmentally efficient. At Hitachi Vantara, we're helping customers build the trusted, high-performance data foundation AI requires with systems that reduce the environmental footprint of the workloads it creates. We believe organizations shouldn't have to choose between accelerating AI innovation and advancing sustainability," said Akinobu Shimada, Chief Executive Officer of Hitachi Vantara.

Shimada's focus echoes growing investor pressure on large technology buyers over emissions from data centres and high-density compute clusters. It also highlights the role of storage architectures and workload placement in managing power demand.

Inside many enterprises, attention is shifting from proof-of-concept AI projects to integrating models into production systems. That transition is exposing fragmentation across on-premise, private cloud and public cloud environments.

Sunitha Rao, Senior Vice President and General Manager of Hybrid Cloud and Software Defined Storage, said organisations often underestimate the integration work required between data sources and AI tools.

"The biggest obstacle isn't the model but the data. Many organizations underestimate how difficult it is to make enterprise data accessible, trusted, secure, and performant across hybrid environments. AI can only deliver meaningful outcomes when the right data reaches the right models at the right time. Data infrastructure is no longer back-office technology; it's the foundation of enterprise AI," said Sunitha Rao, Senior Vice President and General Manager of Hybrid Cloud and Software Defined Storage at Hitachi Vantara.