AI drives container growth as shadow use raises risks

AI adoption is pushing Australian enterprises towards container-based application infrastructure. At the same time, unauthorised use of AI tools and organisational silos are raising security and governance concerns, according to new survey findings from Nutanix.

Australian results from Nutanix's latest Enterprise Cloud Index research show 90 per cent of respondents said AI is accelerating their organisation's adoption of containers. The findings also highlight strong concern about data sovereignty and widespread awareness of "shadow AI", where applications or agents are deployed without IT oversight.

Containers package software and its dependencies into standard units that run across different environments. They are now a common approach for modern application development and for deploying AI-enabled workloads. The results suggest local organisations see containers as a path to faster software delivery and more consistent operations as AI use expands across business functions.

Container momentum

The research points to continued growth in container use. Some 85 per cent of Australian respondents expect container use for applications to increase over the next three years. Two-thirds (66 per cent) said they are already building both new and existing applications in containers.

Senior leadership is also driving AI deployment goals. Almost half of respondents (48 per cent) expect their organisation to have more than five AI-enabled applications within the next three years. However, confidence is lower for some deployment models: if their organisation needed to deploy AI workloads on-premises, 87 per cent said their current infrastructure was not fully ready.

Michael Alp, Managing Director for A/NZ at Nutanix, said the results show strong appetite for AI but also highlight foundational work that still needs attention.

"It is clear Australian organisations are ready to embrace AI, but this requires resilient, dependable, and well-governed infrastructure. Containerisation has emerged as a fundamental pillar in local AI and application strategies, but broader adoption requires a reimagining of the underlying infrastructure," said Michael Alp, Managing Director for A/NZ, Nutanix.

He also pointed to the operational complexity of running older and newer systems side by side.

"Rather than manage a two-speed infrastructure stack, a common operating environment to manage both containerised and traditional workloads would address key concerns like shadow IT and data sovereignty," said Alp.

Shadow AI risks

While AI uptake is driving infrastructure change, the findings underline concern about how AI tools are being introduced inside organisations. Some 72 per cent of Australian respondents said they see AI applications or agents implemented by employees in non-IT functions. Almost all leaders surveyed (92 per cent) believe unauthorised AI use introduces risk.

Respondents linked that risk to exposure of sensitive data and intellectual property. The results reflect a broader enterprise challenge: staff can access third-party AI services or build lightweight automation without formal approvals, while governance frameworks often lag behind rapid experimentation.

Internal operating models may also be compounding the problem. A large majority of Australian respondents (84 per cent) said silos between business units and IT make it difficult to execute technology initiatives effectively. These gaps can slow deployment, complicate oversight, and lead to inconsistent controls over data use and model access across the organisation.

Data sovereignty focus

Data sovereignty emerged as a particularly strong theme in Australia. Some 89 per cent of respondents said it is a high priority when making infrastructure decisions, compared with 80 per cent worldwide.

That priority appears to influence where organisations run infrastructure and place data used by AI applications. More than half of Australian leaders surveyed (60 per cent) said they feel the need to run their infrastructure domestically, either on-premises or via a local cloud region, largely due to security and data protection concerns.

Agents and productivity

The Australian findings also show strong interest in AI agents-systems that can take actions across software tools with limited human input. Most Australian IT executives surveyed (70 per cent) expect agents to improve productivity and efficiency, while 62 per cent expect them to enhance customer or employee experiences.

Some respondents also see AI agents influencing business models. The survey found 58 per cent believe agents could create new products, services, or revenue streams, even as many organisations work through governance, risk, and infrastructure constraints.

The Enterprise Cloud Index research was conducted in November 2025 by Wakefield Research. It surveyed 1,600 cloud, IT, and engineering executives (manager level or above) at organisations with 500 or more employees across multiple markets, including Australia, the United Kingdom, the United States, Japan, Singapore, Germany, France, and India.