How AI Data Centres Are Showing Up in Your Cost of Living

Artificial intelligence feels weightless. You type a question, an answer appears, and nothing seems to have been used up. But every one of those answers runs on very physical things: chips, servers, cooling systems, power lines and large concrete buildings full of all of them. Building that infrastructure is now big enough to show up in Australia's economic data, and the RBA has started naming it as a source of inflation.

What the RBA actually said

In its 3 November statement, the Reserve Bank's board wrote that "AI-related investment is driving rapid growth in prices for technology-related goods." At her media conference the Governor added that investment "has been much stronger than expected, including due to data center-related investment," and that this "has added to inflationary pressures by adding to demand at a time when the domestic economy is already capacity constrained."

That is a striking thing for a central bank to say. A few years ago AI was a technology story. Now it is also a story about interest rates.

Link one: global demand for chips

The first link is global. Companies around the world are racing to build computing capacity, which means huge demand for semiconductors and the equipment around them. The RBA notes strong increases in export prices for semiconductors and other AI-related goods from parts of Asia, and that South Korea and Taiwan have been growing quickly as a result.

When demand for something runs ahead of the capacity to make it, prices rise. Some of that shows up in what Australians pay for technology goods. The RBA says these AI-related price pressures "have not broadened materially beyond electronics", but it flags a risk that they could spread further along supply chains if demand keeps outrunning supply.

Link two: building here

The second link is local. Australia is attracting a lot of data centre investment, mostly in New South Wales and Victoria. The RBA's forecast for business investment has been revised up significantly, largely because of data centres.

There is a twist here. The RBA expects around 70% of this investment to be imported: the servers, chips and specialised equipment are made overseas. That limits how much it boosts Australian GDP. But the other 30% isn't trivial. Data centres need land, buildings, electrical work and engineers, and they may need new energy and grid infrastructure as well. All of that draws on Australian workers and materials, in an economy the RBA already thinks is running near its limits.

Link three: competing for the same resources

This is where data centres meet your cost of living. Construction workers, electricians and engineering firms can only be in one place at a time. When a data centre project hires them, they aren't available to build homes, fix roads or renovate kitchens, at least not without someone paying more. More demand for the same pool of resources pushes up costs across the board.

Electricity is another pressure point. Data centres use a lot of power, and the RBA lists grid capacity as one of the factors that will decide how big the boom gets. Large new users can add to demand for generation and network investment, and some of those costs can end up in everyone's bills over time.

Link four: interest rates

The final link closes the loop. Because data centre investment adds to demand while the economy is already stretched, it is one of the reasons the RBA judged that capacity pressures were tighter than it thought. That fed into the decision to lift the cash rate to 4.85%.

So a household in Brisbane that never uses an AI chatbot may still be paying for the boom, through slightly higher tech prices, a bit more competition for tradespeople, and a mortgage rate that is higher than it otherwise would be.

The upside

None of this means data centres are bad for Australia. They bring investment, construction jobs and, potentially, a role as a regional digital hub. And if AI eventually lifts productivity, that is the one thing that can raise living standards without stoking inflation. The RBA itself mentions the possibility that AI adoption could lift productivity growth.

The catch is timing. The spending happens now, and any productivity payoff comes later. The RBA thinks most of those gains are likely to arrive beyond the next two years, at least in Australia. In between, an economy with little spare room has to absorb a big new source of demand, and that is the gap where inflation can creep in.