Movatterモバイル変換

GUEST OPINION:  The global cloud industry shows no signs of slowing down anytime soon, scaling at around 15%-19% annually and given added momentum by the rapid rise of AI workloads.

Domestically, the picture is similar, with Australian organisations using cloud as a default part of modern IT strategy across most sectors. At the same time, however, local operational pressures are intensifying, with rising energy costs and skills shortages among the various issues pushing CIOs to scrutinise cost, performance, data sovereignty and the long-term viability of their cloud choices more closely than ever.

The growth of AI adds further complexity, as new model training and inference workloads place significant pressure on compute, power and bandwidth, forcing organisations to rethink how they provision infrastructure cost-effectively. This raises some important questions around how organisations should balance cost control with operational flexibility as they refine their cloud strategies.

A changing landscape

In common with markets elsewhere, Australian organisations currently utilise a mix of public, private, and hybrid cloud models, depending on the type of workload, compliance expectations, performance requirements, and various other priorities.

Looking at the options more specifically, public cloud remains attractive for elastic workloads and rapid deployment, despite the fact that cost variability and data residency concerns are prompting increased scrutiny. As an alternative, private cloud continues to appeal to organisations managing sensitive or highly regulated data, providing them with tighter control over location, configuration, and access.

Elsewhere, hybrid cloud strategies are becoming increasingly common, not least because they enable teams to keep sensitive or latency-critical workloads on dedicated infrastructure while leveraging the scalability of public cloud for all other applications. As a result, many CIOs are now focusing less on a pure “cloud first” approach and more on aligning the right execution model with the right workload, especially given the additional cost and power demands emerging from AI.

This shift in thinking is increasingly tied to the financial realities of running cloud at scale, along with the growing importance of knowing exactly where sensitive data is located. Take the typical cloud cost base, for example, where rising operational costs are increasing the financial pressure on data-intensive workloads. Many organisations are also finding that cloud spending is rising faster than anticipated, often due to issues such as consumption-based billing, unpredictable scaling events and charges that are difficult to forecast. In addition, egress fees and complex pricing structures can make it challenging for teams to forecast long-term operational expenditure with confidence.

Yet, businesses want greater cost predictability and this issue now plays a greater role in cloud strategy discussions, particularly for sectors handling steady, high-volume workloads where volatility can become difficult to budget for. At the same time, data sovereignty is also becoming a key driver for cloud decisions, with organisations needing to keep sensitive information stored and processed within Australia, not just to comply with the Privacy Act but also to address customer expectations around the control of data.

The role of public cloud and Virtual Private Servers

So, how are these issues impacting the way technology leaders are integrating cloud into their existing tech stacks? Firstly, it’s essential to recognise that public cloud continues to suit organisations that require rapid access to scalable compute and storage without incurring the burden of managing underlying infrastructure. It is particularly useful for development environments, digital services with variable traffic and workloads that benefit from fast provisioning and broad geographic coverage.

However, the need for more predictable operating costs is prompting organisations to assess which workloads genuinely require the elasticity of public cloud and which can run more efficiently elsewhere. This is where Virtual Private Servers (VPS) and other lightweight hosted environments are gaining more interest, particularly among SMEs, digital-native businesses and teams with stable, well-understood workloads.

In particular, VPS offers a simpler, lower-cost alternative for applications that do not require the scale or complexity of full public cloud environments, giving organisations more direct control over configuration and resource allocation. For many businesses, combining a VPS for baseline workloads with a public cloud for scalable or burst requirements helps balance cost efficiency with operational flexibility, while also maintaining clarity over where workloads run and how data is handled.

Moreover, workloads with stable demand or predictable usage patterns often sit more efficiently on dedicated or VPS-style environments, while highly variable or seasonal workloads may justify public cloud elasticity (but only if consumption patterns are monitored and controlled). In contrast, AI-centric workloads require additional planning, as training, tuning and inference each impose different power, bandwidth and cost profiles.

The underlying point is that technology leaders need to identify the best execution venues for each workload category, driven by clearer governance frameworks to define how workloads are placed, moved and costed over time. This approach can help CIOs build infrastructure roadmaps that remain adaptable, even as the technology ecosystem changes at a rapid pace

Maximising Cloud Efficiency - LUMEN WEBINAR 23 April 2025