The advent of AI, especially generative AI, has allowed businesses to fast-track much of that.

“We’re starting to de-risk the things people didn’t do before because they were cost prohibitive or too complex,” noted Walls. “That also helps de-risk modernisation as you modernise applications and architectures.”

Hybrid cloud strategy

Datacom sees cloud as a model, not just as a data centre. This means first working out the problem customers are trying to solve. The next step is to find the right platform and technology for that particular problem or workload — making sure not to put all your eggs in one basket, said Walls.

“You can then work out the right platform from a cloud-native perspective, rather than just lifting and shifting and not realising the benefits because it can’t scale or be flexible,” he said. “A hybrid cloud strategy enables the ability to move and connect workloads between multiple clouds — on-premises, private, or public.”

However, moving those workloads doesn’t happen automatically. The technologies in hybrid aren’t always compatible, vendor lock-in doesn’t make it easy to move out, and supportability is a challenge.

“What we’ve seen — and what the report highlights — is a significant increase in the use of containerisation technologies,” explained Walls. “First, workloads ran on hardware. Then came virtualisation with VMs.

“Cloud took that further, but often still ran VMs. Containerisation takes it further still, packaging workloads into a more portable mechanism.”

Walls said it’s a more flexible technology that allows workloads to be delivered more efficiently.

“That’s where people are reducing the complexity of hybrid multi-cloud and moving towards the portability goal, as containerisation has really taken off in the last 12 to 18 months,” he said.

AI and sovereignty

This reduces the complexity of the back end and makes it easier with the acceleration of AI workloads for high-performance, low-latency infrastructure.

Datacom’s cloud report shows a large percentage of organisations are intending to keep their data processing in-country.

However, over half of this year’s respondents expressed concern about in-country infrastructure capacity.

“It’s clear that AI is not only reshaping compute and storage requirements but also intensifying focus on data sovereignty and local infrastructure investment,” said the report.

According to Walls, the other interesting thing that shows up throughout the report and in customer conversations is the increasing geopolitical changes and tensions around the world.

“Vendor lock-in is one of them, but you also look at tariffs, fluctuating exchange rates, and challenges around resiliency of your platform, like access to your data — if an internet cable gets chopped, for example,” he said. “We’re seeing challenges around internet cables being attacked or supposedly attacked around the world.

About 68 per cent of Australian respondents interviewed said they wanted local data processing in place.

“That fed into AI workloads being put in private cloud—around 50 per cent preferred that platform,” said Walls. “This links to concerns about political tensions and data sovereignty.

“Generally, people use hyperscale public cloud to experiment and test because it’s flexible: you can turn it on and off, and all the technology exists there, so it’s accessible.”

Once users understand their use cases and want to get into full production of running LLMs full-time, it’s often perceived to be safer and potentially more commercially viable to run in private cloud.

“It’s always on and productionised at that point,” said Walls. “Again, that lends itself to a hybrid strategy: use public cloud for experimentation, innovation, and customer-facing digital omnichannel elements.

“Run your core knowledge work in sovereign or private cloud because of those concerns –using both effectively is still a good strategy.”

Risk and resilience

From a technology priority perspective, security was rated as a top technology priority. From a business priorities perspective, risk management ranked sixth.

Walls said security strategy and hybrid cloud strategy should work together and coexist.

“It’s a Venn diagram: security strategy, cloud strategy (hybrid), and cloud security strategy as a subset,” he said. “Cloud security strategy is about protecting your cloud assets and ecosystem.

“Hybrid strategy is about protecting your wider ecosystem at the same time. It’s all connected.”

From a security perspective, only 43 per cent of those surveyed had a formal cloud security strategy in place, and only 50 per cent had mature security tools.

“There’s still a maturity challenge in just protecting cloud,” he said. “There’s even less maturity in understanding that hybrid cloud strategy provides part of your security and GRC [governance, risk and compliance] strategy too.

“People are starting to realise now, and we’re seeing it show up in customer conversations. AI has forced many to think about this, as have geopolitical tensions.”

This also plays into the need for locally skilled staff as a top criterion for choosing technology partners, for government procurement.

It reflects both skills shortages and geopolitical concerns, as well as the need for enablement alongside technology delivery, added Walls.