Nutanix has published research on AI adoption in financial services, highlighting growing concern about the unsanctioned use of AI tools across the sector.
Its Financial Sector Enterprise Cloud Index found that 66% of IT executives in financial services encounter AI applications or agents introduced by employees outside IT, while 86% believe the use of AI tools and agents without official oversight creates business risk.
The findings point to a gap between enthusiasm for AI and the systems needed to manage it. Governance, infrastructure and operational readiness are lagging as financial institutions try to expand AI use across their organisations.
When asked what was slowing efforts to scale AI, respondents ranked process issues and organisational barriers ahead of technical limits. Some 38% cited process complexity as the main obstacle, while 34% pointed to leadership and skills issues. By contrast, 28% identified technical limitations as the chief barrier.
Another recurring issue was the relationship between business units and IT teams. The research found that 83% of respondents said silos between those groups make it harder to execute technology initiatives effectively.
Infrastructure strain
The report suggests many firms still lack the computing environment needed to support broader AI deployment in-house. Around 68% of respondents said their infrastructure is not fully equipped to support AI workloads on-premises, and 64% said they rely on third-party providers to close that gap.
That dependence sits alongside concerns about where data is held and processed. While 79% said data sovereignty is a priority, 62% reported running containerised workloads in the public cloud, creating what Nutanix described as a growing tension between policy priorities and deployment choices.
The company called that mismatch "Sovereignty Debt", arguing that firms are under pressure to meet compliance expectations while continuing to use cloud environments that may complicate control over sensitive data.
Containers rise
Containerisation emerged as a central part of how firms are preparing for AI use. The research found that 90% of respondents said AI is increasing container adoption, while 89% expect containerisation to grow further.
The trend reflects a broader effort across financial services to run applications more portably and consistently across different environments. For firms operating across on-premises systems and public cloud services, containers are increasingly seen as a practical way to manage workloads as AI use spreads.
The study was based on a survey of 1,600 cloud, IT and engineering executives at manager level or above from organisations with 500 or more employees. Respondents came from 15 countries, including Australia, the United Kingdom, the United States, Japan, India, Germany and Singapore.
In Australia, the findings come as regulators and institutions pay closer attention to the risks linked to newer AI systems. The Australian Prudential Regulation Authority has warned about security concerns tied to agentic AI tools, noting that they can widen the attack surface and introduce new vulnerabilities.
That backdrop helps explain why shadow AI has become a concern for financial institutions, where governance, oversight and auditability are closely tied to regulatory compliance. Unapproved use of AI tools by staff can create risks around data handling, decision-making and security, especially if those tools are integrated into daily operations without central review.
Jay Tuseth commented on the regional implications of the findings.
"Across APJ, the race is no longer just about who has the most advanced AI models, but who can scale them securely and responsibly. As financial institutions navigate growing risks around sovereignty and unsanctioned AI use, success requires a shift toward flexible, containerized platforms that unify workloads across hybrid environments. The winners won't just have the biggest compute budgets - they will be the ones who successfully align their infrastructure with regional compliance and data sovereignty demands," said Jay Tuseth, Vice President, APJ, Nutanix.
The data suggests financial services groups are moving beyond early experimentation with AI into a phase where internal controls, architecture and operating models matter more. For many institutions, the challenge appears less about whether to adopt AI than how to do so within existing risk frameworks.
For IT leaders, the survey indicates that the immediate pressure is not only to support new AI workloads but also to regain visibility over how employees and business units are already using the technology. In a sector where governance failures can quickly become regulatory issues, that may prove as significant as the underlying infrastructure gap.