Australia AI spending rises but returns lag, report says

Cloudera has released a report on enterprise AI adoption that highlights a gap between spending and returns in Australia. The findings are based on a global survey of 1,500 IT leaders and decision-makers.

Australian respondents reported substantial spending on AI projects, but fewer said those investments were delivering net value. The survey found that 76% of Australian organisations invest at least USD $500,000 a year in AI, while only 21% said AI currently generates more value than it costs.

The figures suggest a cautious phase in corporate AI spending. Only 43% of Australian respondents said their AI spending had increased from a year earlier, even as businesses continue moving projects from trial stages towards wider use.

Cost pressures also appeared to be shaping deployment decisions more broadly. Globally, 89% of respondents said they had limited or redesigned AI use cases because of rising token consumption, while 45% had moved AI workloads in response to higher or more volatile costs.

Those concerns sit alongside operational and regulatory constraints. In Australia, 82% of respondents said governance and compliance were limiting their ability to scale AI, and 74% said data residency and sovereignty requirements influenced where AI workloads were deployed.

The same themes were evident globally. The survey found 75% of respondents cited governance and compliance as a constraint on scaling AI, while 84% said data residency, sovereignty or jurisdiction-specific requirements influenced where workloads ran.

Governance pressure

The findings come as AI policy and oversight tighten in Australia. The federal government has been pursuing broader AI adoption while also developing stronger governance measures, including Australian standards for AI and new privacy transparency requirements covering some forms of automated decision-making.

That policy backdrop helps explain why deployment location has become a board-level issue for many organisations. Data sovereignty rules can determine whether workloads remain onshore, move between cloud environments or require redesign before they can be deployed more widely.

The report suggests businesses are finding that technical performance alone is not enough to justify expansion. Respondents also reported reliability and execution problems: 98% globally said they had pulled back an AI deployment because of failure, hallucinations or unexpected behaviour.

For some companies, the lesson appears to be that the basics were not addressed early enough. Asked what they would do differently if restarting their AI efforts, 49% of global respondents said they would strengthen governance from the beginning, 46% would invest in data quality earlier, and 45% would modernise infrastructure sooner. Another 45% said they would invest more in employee training.

Shift in priorities

The survey also points to a shift in how executives are judging AI programmes. Rather than rewarding experimentation for its own sake, organisations appear to be looking more closely at costs, reliability and the practical hurdles involved in putting systems into production.

That is particularly visible in Australia, where investment growth lags ongoing spending commitments. Businesses may still be allocating large sums to AI projects, but the lower share reporting positive returns suggests many are under pressure to prove commercial value before increasing budgets further.

Organisations operating AI across multiple environments cited resilience, sovereignty and residency requirements, cost reduction and proximity to data among the main reasons for doing so. The figures suggest where AI runs is becoming a strategic issue shaped as much by regulation and economics as by computing needs.

There were also signs that moving workloads is not straightforward. The report found 58% of global respondents said shifting workloads to meet residency or sovereignty requirements required moderate or major redesign, suggesting compliance changes can create significant operational work.

Brian Martin, Chief Technology Officer of Applied AI at Cloudera, said the economics and rules around enterprise AI were changing quickly. "Scale is exposing a new reality for enterprise AI: what works today may not make economic or operational sense tomorrow. Costs change, models evolve, regulations shift, and the right place to run a workload changes with them. Leaders need an AI strategy built for that reality, one that gives them the freedom to adapt without giving up control of their data, governance, or spend," Martin said.

For Australia and New Zealand, Cloudera argued that the focus was turning towards tighter discipline around use cases and measurement. "While Australia trails the broader Asia-Pacific region in AI investment growth, demand remains strong as organisations pivot from experimentation to accountability," said Keir Garrett, Managing Director, Cloudera Australia and New Zealand.

"Success is no longer about spending more, but about investing with precision: backing high-value use cases, fortifying data governance, and defining clear metrics from day one. As companies move AI from pilot to production, the focus is firmly on measurable business value - driving productivity, optimising costs, and accelerating revenue growth. Ultimately, AI at scale is not about deploying technology everywhere, but about scaling what demonstrably works," Garrett said.