National AI Plan commits $460m, but SMEs told to carry the load

The Albanese Government’s National AI Plan, launched on 2 December 2025, commits $460 million in existing funding across three pillars — capturing economic opportunity, spreading the benefits, and keeping Australians safe — while making plain that the bulk of future investment must come from business itself.

Published by the Department of Industry, Science and Resources, the document sets out to position Australia as a regional AI leader. Its promise is broad: accelerate adoption, build sovereign capability, and regulate the risks. But industry groups and small-business advocates warn it stops short of telling the country’s 2.5 million small and medium enterprises how they are meant to begin. This is not a detail. SMEs account for 97 per cent of all Australian businesses and employ 42 per cent of the workforce, and the plan’s SME-facing measures — a $17 million AI Adopt Program and a promise of sector-specific guidance — amount to roughly 3.7 per cent of the $460 million total. It is a ratio the Council of Small Business Organisations Australia described as moving “in the right direction” but short of long-term investment certainty.

“The load must be shared,” said Kylie Walker, chief executive of the Australian Academy of Technological Sciences and Engineering. “[The government] is clearly signalling it will provide leadership and some investment, but sees most coming from businesses.” Walker, whose academy represents more than 900 of Australia’s leading engineers and applied scientists, noted that the plan’s success depends almost entirely on whether private industry picks up the other end of the bargain.

Funding is the best-mapped section of the document. Of the $460 million, about $362 million flows through research grants administered by the Australian Research Council, the Medical Research Commercialisation Futures Fund, the National Health and Medical Research Council, and Cooperative Research Centres. A further $29.9 million funds the AI Safety Institute, set to launch in early 2026 with a remit to develop testing and assurance frameworks for high-risk AI systems, while the $17 million AI Adopt Program remains the only line item explicitly targeted at smaller firms. The Safety Institute, modelled partly on the UK’s AI Safety Institute launched in late 2023, will assess frontier models for risks including bias, misuse, and systemic harm before deployment in critical infrastructure or government services.

Private capital is already moving at scale. Government figures show $100 billion in private-sector data centre investment was announced between 2023 and 2025, and the National Reconstruction Fund made a $32 million equity investment in Harrison.ai, a Sydney-based medical imaging AI firm, in January 2025. Separately, the National AI Centre launched its AI.gov.au portal in May 2026, providing the plan’s public-facing front door with practical guidance and tools for organisations.

SMEs see little of it.

“The feedback I hear from companies with 200 employees or fewer is that they do not know where to start with AI, and the plan does not tell them how to catch the opportunities,” said Dr Sue Keay, director of the UNSW AI Institute.

William Young FCPA, co-founder of Praxio AI, put the structural problem in starker terms. “Unlike the Big Four accounting firms, small firms do not have the resources to invest in developing AI. Even finding the time to experiment and adopt AI can be challenging for smaller firms.”

Industry response has split along size lines. The Business Council of Australia — whose members include the country’s largest employers — welcomed the plan as “an important step forward.” Xero, the accounting software firm used by more than a million Australian small businesses, was blunter: the plan misses what those businesses actually need. It pointed to the absence of practical tooling guidance and sector-specific roadmaps, and noted that most small-business owners lack the time to interpret a whole-of-government strategy document unaided. Globally, Australia’s approach mirrors the pattern seen in Canada and the UK — strong on research funding and safety regulation, lighter on direct SME deployment support that countries such as Singapore have prioritised.

Assistant Minister for Science, Technology and the Digital Economy Dr Andrew Charlton, in the ministerial media release that accompanied the launch, said the plan would “enable all Australians” and pointed to the AI Safety Institute as a centrepiece of the government’s responsible-AI framework.

What comes next will test whether the plan works as a floor, not a ceiling. The Safety Institute is expected to begin operations within months. How the $17 million AI Adopt Program is actually disbursed to SMEs has not yet been detailed, and COSBOA has signalled it will watch the program’s design closely. Meanwhile, the $100 billion in private investment the plan counts on will flow overwhelmingly to firms that already have AI capabilities. For everyone else, the plan is a signal, not a manual. Whether the next federal budget raises the SME allocation will be an early test of that signal’s volume.