Published on the 30/10/2025 | Written by Heather Wright

Exporting intelligence and big SMB benefits…

Australia’s small businesses are poised to be the big winners in AI, with productivity gains projected to be 22 percent higher for them compared to large enterprises. But the opportunity doesn’t stop at home, with a new report touting a potential $10.6 billion annual export market and the chance to become a leading regional hub for AI in Asia Pacific.

“If Australia develops a domestic AI product/service at similar rates to domestic software development, Australia will capture most of the benefits in this layer.”

Australia’s AI Opportunities, which was developed in collaboration with organisations including the Australian Computer Society, Australian Information Industry Association and the Business Council of Australia and major data centres, and funded by OpenAI, looks at three key opportunities for Australia around AI: AI adoption to increase labour productivity, doubling down on building new AI sovereign capabilities, and becoming a regional hub and exporter of AI products and services.

If – and it’s a big if – Australia captured all three opportunities, the report says it could add $142 billion per annum to the economy by 2030 – with $112 billion of that coming from gains through adoption. The 60-page report – which comes complete with a comprehensive outline of the methodology behind the figures – notes that AI is already adding $21 billion a year through increased labour productivity via using AI in areas including proofing documents, ideation or more efficient search.

Small business, big gains

It’s in AI adoption that small businesses stand to be big winners, the report claims, suggesting small businesses could see productivity lifts of up to 22 percent greater than for large businesses. Unlike previous productivity booms which required significant capital investment, AI tools are increasingly delivered via scalable cloud-based platforms. This AI-as-a-service model enables small businesses to access powerful technologies without the need for expensive infrastructure (or at least needing potentially less expensive infrastructure). At the same time, SMBs are heavily concentrated in labour-intensive, lower-productive industries where AI has the greatest potential to automate routine tasks and boost efficiency.

The report shows small business skewed industries stand to reap a 7.1 percent productivity increase, versus the 5.8 percent for large business skewed sectors.

A 2025 Salesforce survey cited in the report found that 86 percent of Australian SMBs already report revenue boosts from AI use, underscoring the sector’s readiness to adopt and scale AI solutions.

Major service export industries, including higher education, professional services and tourism could be significant beneficiaries too,

Health and social assistance too, will experience greater benefits and accompanying wage growth as AI improves the speed at which workers can complete tasks and unlock potential for new types of work such as classifying items in large datasets to create new analytical opportunities.

On the other hand, traditional Australian strengths such as resources are proportionally less likely to benefit from AI, given they have already adopted the technology and are highly capitalised meaning labour productivity improvements have smaller relative impacts.

Exporting intelligence: A regional AI play

Developing Australian AI capabilities presents an $18 billion a year opportunity by 2030, the report claims, while export opportunities present a potential $11 billion per annum windfall.

Australia will need to either import or develop homegrown AI capabilities in compute and AI product development, with the report calling on the country to aim to domestically host the required compute – worth an estimated $6 billion per annum. Developing user-ready AI products and services using AI inferences as inputs could add the additional $12 billion.

“This layer of the supply chain will be competitive, but if Australia develops a domestic AI product/service at similar rates to domestic software development, Australia will capture most of the benefits in this layer.”

The third economic opportunity requires Australia to develop sufficient domestic capabilities to become a regional tech hub, providing compute to Asia-Pacific neighbours for regular business use of compute for global training loads for frontier model developers.

That’s a model akin to that touted earlier this year by Atlassian co-founder Scott Farquhar, who suggested Australia should provide ‘digital embassies’ for the wider region.

The report suggests Australia has many existing strengths which lend themselves both export potential and becoming a regional hub, with a robust data centre industry, globally competitive software companies like Canva, Atlassian and Lorikeet, and a rapidly growing AI education sector with international enrolments in AI-specific degrees.

The country’s geographic and political stability also make it attractive for regional compute services and access to renewable energy, available land and low-latency infrastructure position Australian data centres well to serve APAC neighbours for training and inference workloads.

The report suggests Australia could capture five percent of Asia Pacific inference demand – worth $2 billion – and 2.5 percent of global training demand, worth a further $1 billion – but investment of around $14 billion in data centre infrastructure – and a 50 percent increase in compute capacity – would be required along with investment in supporting infrastructure including energy and network transmission.

Barriers to AI growth

Data centre infrastructure investment is just one area where work is needed, with the report highlighting four key other barriers, on top of compute capacity, to realising the benefits of AI.

That includes a need to develop clarity around which capabilities the country wants to cultivate and upskilling and improving AI literacy.

Addressing a decline in AI research and development, with Australia under-performing its Asian neighbours in model development is also highlighted, with AI research declining by 46 percent since 2022 and the country accounting for just 0.2 percent of large models trained since 2022, along with building public trust and confidence in safe, responsible and democratic AI tools and use cases.

The report warns that while Australia is well-positioned to realise each of the three opportunities outlined, concerted action will be required.

“With coordinated action across government, industry and education, Australia can position itself as a trusted, capable and regionally significant AI economy by 2030, securing long-term productivity and inclusive growth.”