One small step, no giant leap
the government’s ai plan was a missed opportunity to position australia as a hub for data centre investment. by Nico louw.
First published in the MRC’s Watercooler newsletter. Sign up to our mailing list to receive Watercooler directly in your inbox.
Next year, four companies — Alphabet/Google, Amazon, Microsoft and Meta — will spend a combined $USD400 billion on data centres. Their annual investment is forecast to increase for at least the next five years, reaching over $USD500 billion in 2029 and 2030.
Put another way, spending by just these four companies on data centres for the foreseeable future will be equivalent to around 80% of the entire Australian Government Budget.
This presents a huge opportunity for Australia to be the hub for data centre investment in South East Asia. Companies should want their assets to be here. We have a highly skilled workforce, plenty of space, and strong institutions and governance.
But time is of the essence. We are already at an inflection point where global spending is booming to keep up with the expansion of AI. The small proportion of this investment that has reached Australia is already showing up in the national accounts data, which shows that capital expenditure on IT equipment and machinery rose an astonishing 91.5% in the September quarter to $2.8 billion, by far the highest quarterly result recorded.
This is the 7th quarter in a row where investment was above $1 billion. In the preceding decade, quarterly investment averaged only $400 million, reaching $500 million only four times.
Source: ABS
This increase sounds impressive, but it’s only a tiny sliver of global spending and a fraction of what’s possible. The Government should be rushing to secure our slice of the pie before it’s too late.
It’s with that in mind that there were high hopes for the Government’s National AI Plan released this week. This was our opportunity to pitch Australia to the tsunami of global capital that is ready to deploy.
Unfortunately, the so-called ‘Plan” contains almost nothing of substance. The Australian Financial Review summarised it as being “completely wrong”. The key ‘action’ in the Plan relating to data centres is to build the NBN, which already exists. The only new funding is $30 million for more public servants.
The Government did at least take one important step in the right direction. It dumped Ed Husic’s plans for an AI Act with ‘mandatory guardrails’ for AI and will instead take a technology-neutral approach and rely on existing laws and regulations. A new team in the Department of Industry, Science and Resources will coordinate the approach of regulators and other departments.
We strongly support this change of direction, since it’s exactly what the MRC argued for in our submission to the Government’s consultation process on AI regulation. It means that the Government will at least not be a hindrance to the private sector finding productive uses for AI.
This small step is good, but we needed a giant leap on data centres. What we got was a promise to "position Australia as a leading destination for data centre investment” by “developing a set of national data centre principles”. Setting aside why you’d bother to release an AI Plan before you’ve developed such principles, the language in the Plan suggests that they will largely be red tape aimed at forcing investors to build renewable energy capacity and use efficient cooling technologies.
Incentivising investment in the energy generation needed for data centres is a good idea, but the lack of any mention of gas is telling. In the US, there are multiple examples of major gas projects underway explicitly linked to data centres — so many that it has caused a global backlog of gas turbine equipment until 2030. In addition, Amazon, Microsoft and Google have all made significant financial commitments to underwrite or restart nuclear power units for their data centres.
Australia needs an urgent push to expedite planning approvals for new data centres and associated energy infrastructure. The Plan vaguely alludes to this as an opportunity the Government is “exploring” with the states. It should be their most important short-term (months, not years) priority.
The MRC is currently developing a suite of practical policies to support AI in Australia, with a strong focus on securing investment and boosting economic growth. This work aims to fill the leadership gap that has been left by the Government and will be released shortly.
