Hitachi Vantara has published research showing that only 10% of financial institutions prioritise AI-ready storage and data platforms, highlighting a gap between AI plans and storage investment in financial services.
The study surveyed 100 decision-makers across banking, payments and investment firms worldwide. It found that 35% of respondents ranked managing data growth as their top storage priority, while 30% cited data sovereignty, regulatory compliance, and policy-led governance as leading concerns.
The figures suggest many firms are focusing first on immediate operational and regulatory pressures rather than the systems needed for broader AI use. Only 9% said they prioritise centralised data hubs for governance, reporting, AI and machine learning, and data reuse.
Data sovereignty emerged as a near-universal factor in AI planning. Almost all respondents (99%) said sovereignty concerns influence where they run AI workloads, while 19% said those concerns significantly limit AI workload scalability or performance.
Storage buying decisions also appear to be shaped heavily by budgets. The research found that 65% of financial institutions cite cost or total cost of ownership as the most important factor when choosing object storage, compared with 46% who chose data resilience and availability.
Cost pressure
That focus on cost reflects the competing demands facing banks, payments groups and investment firms as they expand data estates while staying within regulatory boundaries. The survey indicates that short-term spending pressures can outweigh longer-term work on data accessibility, governance and storage structures suited to AI systems.
Different approaches to sovereignty add to that complexity. Among respondents, 23% said they restrict AI workloads to specific regions, 21% said they train models centrally while keeping data local, and 16% said they split training and inference across locations because of sovereignty rules.
These patterns point to a fragmented operating environment across the sector. Rather than following a single model for AI deployment, firms are making trade-offs based on regulation, geography, data location and cost.
Split market
The findings also indicate a divided market in object storage adoption. Some 35% of organisations reported enterprise-wide deployments across multiple workloads and teams, while 36% said they remain at an early stage or in pilot phases.
That split suggests some institutions are further ahead in integrating storage across business functions, while others are still testing how to build broader data management structures. In a heavily regulated sector, differences in maturity can affect how quickly firms move from isolated AI experiments to wider operational use.
Object storage has become more important as companies seek to manage large volumes of unstructured data for analytics and machine learning. Financial institutions still rely on block storage for core databases, transaction systems and other operational workloads, making the relationship between older and newer storage approaches an important issue for technology planning.
Octavian Tanase, Chief Product Officer at Hitachi Vantara, said the results show many firms have not yet aligned data management priorities with their infrastructure needs.
"Financial institutions clearly recognise that data management is becoming more complex, but many are not yet fully addressing what their environments require," said Tanase.
"As data volumes grow, organisations need unified data platforms that can span block, file and object storage to reduce fragmentation, improve visibility and support consistent governance for the mission-critical data that financial institutions depend on. This includes unstructured data used for analytics and AI, as well as the mission-critical databases and transaction systems that support core operations. That activation depends on the data availability and resilience needed to keep information accessible, protected and ready for use."
The research adds to a broader debate in financial services over whether AI strategies are outpacing the infrastructure beneath them. While firms have invested heavily in data, cloud and analytics in recent years, the survey suggests many are still balancing foundational work with pressure to show near-term returns.
Tanase said storage systems need to support both established operational demands and newer data models.
"Financial institutions are being asked to manage more data, in more places, while maintaining control, resilience and compliance," he said. "Modern storage platforms, such as Hitachi Vantara's Virtual Storage Platform One, feature block, object and file storage systems that can help organisations address that complexity by supporting scalable, governed and accessible data environments that serve both traditional retention needs and emerging AI-based demands, including data lakehouse architectures supported by open table formats and native S3 Tables capabilities."