Financial institutions are facing a growing data challenge, yet many are still failing to prioritise the infrastructure needed to turn that data into meaningful AI outcomes.
New research from Hitachi Vantara reveals a sector grappling with rapidly expanding data volumes, increasing regulatory demands and growing pressure to adopt AI. However, the findings suggest many organisations remain focused on managing immediate challenges rather than building the foundations required for long-term success.
The global study of 100 decision-makers across banking, payments and investment firms found that managing data growth is now the top storage priority for financial institutions. More than a third (35%) cited data growth as a key concern, making it the highest-ranked issue across the survey.
Yet despite the industry’s growing focus on AI, only 10% of respondents said enabling AI-ready storage and data platforms was a priority. Even fewer (9%) identified centralised data hubs for governance, reporting, AI and data reuse as a strategic focus.
The findings highlight a growing disconnect between AI ambition and data reality.
While organisations are investing heavily in AI initiatives, many appear to be overlooking the data infrastructure required to support those ambitions at scale. Without stronger foundations, the risk is that AI projects become constrained by fragmented data environments, inconsistent governance and limited visibility across the organisation.
The challenge is compounded by increasing regulatory and sovereignty requirements. The survey found that ensuring data sovereignty, regulatory compliance and policy-driven governance is now the second-highest priority for financial institutions, cited by 30% of respondents.
These concerns are already having a tangible impact on AI strategies.
Almost every respondent (99%) said data sovereignty requirements influence where they run AI workloads, while 19% reported that sovereignty concerns are significantly limiting AI scalability or performance.
This highlights a reality often overlooked in discussions around enterprise AI. Success is not simply determined by access to compute resources or advanced models. Organisations must also determine where data can be stored, how it can be accessed and whether it can be used across different jurisdictions without creating compliance risks.
According to Octavian Tanase, Chief Product Officer at Hitachi Vantara, many organisations recognise that data management is becoming more complex but have yet to fully address the implications.
As data volumes continue to grow, financial institutions require platforms capable of managing block, file and object storage environments in a unified way. Reducing fragmentation, improving visibility and enforcing consistent governance are becoming increasingly important as organisations seek to support both traditional operational workloads and emerging AI use cases.
Cost pressures are also shaping decision-making.
When selecting object storage platforms, nearly two-thirds (65%) of respondents identified cost or total cost of ownership as the most important consideration. This was almost 20 percentage points higher than the next highest factor, data resilience and availability.
The findings suggest that while organisations recognise the need for modernisation, immediate budget considerations continue to influence infrastructure decisions. For many financial institutions, balancing cost efficiency with long-term investments in governance, accessibility and AI readiness remains a difficult challenge.
Data sovereignty requirements add further complexity. Nearly a quarter (23%) of respondents restrict AI workloads to specific regions, while 21% train models centrally but keep data local. A further 16% split training and inference activities across locations to meet sovereignty obligations.
Together, these approaches demonstrate how organisations are attempting to balance innovation with regulatory control, often creating increasingly complex operating environments in the process.
The research also reveals an industry progressing at two different speeds.
While 35% of financial institutions have implemented object storage across multiple workloads and teams, 36% remain in pilot or early-stage deployments. This suggests some organisations are already building scalable, integrated data foundations, while others are still taking the first steps towards modernisation.
As AI and analytics initiatives continue to expand, the importance of data infrastructure is only likely to increase.
The organisations that succeed will not necessarily be those with access to the most advanced AI models. Instead, they may be the ones that have invested in creating resilient, governed and accessible data environments capable of supporting both today’s operational requirements and tomorrow’s AI ambitions.
The research suggests that for many financial institutions, the biggest challenge ahead may not be AI itself. It may be getting their data foundations right first.