RBA's Verdict on AI: No Productivity Lift Yet, But Rates Are Rising Because of It

Australia's central bank has raised interest rates five times this year to an 18-year high of 4.85%, and Reserve Bank of Australia Assistant Governor Sarah Hunter says the AI data centre boom is one of the reasons why. Speaking at a UBS conference in Sydney on Monday, Hunter delivered a sober assessment that cuts against the wave of AI-will-change-everything narratives: there is, so far, little hard evidence of an economic payoff.

Her exact words: there is "limited evidence that artificial intelligence had meaningfully lifted productivity and caused widespread job losses." That's a central banker's way of saying the productivity revolution is still a promissory note. Australia's economy is spending like the revolution has arrived, but the output numbers haven't caught up.

What does this change for businesses and investors?

For practitioners, the RBA's position sets a useful benchmark. Hunter did not dismiss AI's long-term potential — her framing was explicitly conditional: "If AI delivers broader productivity gains, economies will be able to produce more with the same inputs." Note the "if." The RBA treats productivity improvement as a future possibility, not a current fact, and it will watch for evidence rather than promises.

What's already real is the capital spending. A surge in domestic data centre investment has added demand pressure to the Australian economy at a moment when supply is constrained. Hunter identified that pressure as one driver behind the five rate hikes that took the cash rate to 4.85% — a level not seen in 18 years. In plain terms: AI is currently inflationary in Australia, not productivity-enhancing.

The inflation math, and why it shifts over time

Hunter laid out a two-phase picture of how AI investment hits prices:

• Near term: AI investment boosts demand and adds to price pressures while the economy remains supply-constrained.
• Later: "Over time, supply will catch up with demand as data centre capacity and related sectors expand, easing current bottlenecks," she said.
• If productivity gains materialise: the initial inflationary pressures would dissipate.

For anyone pricing Australian assets or planning capital expenditure, that sequencing matters. The inflation hit comes first; the offsetting supply response and any productivity dividend come later, on an uncertain timeline.

Tentative labour market signals, not mass displacement

On jobs, Hunter pointed to "tentative signs of adjustment" rather than large-scale displacement. The evidence she cited: "weaker job ad posting and employment outcomes in more AI-exposed occupations, and stronger growth in areas where the technology has had less of an impact, particularly for younger workers."

The nuance is important for hiring managers and workers alike. The RBA is not claiming AI has zero labour market effect. It's saying the effect shows up in hiring patterns within exposed occupations — fewer postings, softer outcomes — rather than in economy-wide unemployment. Younger workers appear to be feeling the shift most directly.

A financial stability warning worth pricing in

Hunter also flagged a risk that goes beyond inflation arithmetic. Asset prices, she warned, could face "sharp repricing or even disorderly adjustments if AI investment returns fall short of expectations." That's the RBA explicitly acknowledging that current valuations may be leaning on AI returns that haven't been delivered. If those returns disappoint, the correction may not be gentle.

The bottom line for practitioners

The RBA's assessment boils down to four concrete points:

• AI investment is real and large enough to influence interest rates.
• Productivity gains have not yet appeared in the numbers.
• Labour market effects are early, uneven, and visible in hiring patterns, not aggregate job losses.
• The main immediate macroeconomic impact is demand-side inflation pressure from capital spending.

What remains unknown is the timeline. Hunter's "if AI delivers" formulation leaves open whether productivity gains arrive in years, decades, or not at all. Until the data shifts, the RBA's stance gives professionals a defensible reference point: treat AI's macroeconomic story as a capital-spending story today, and a productivity story only when the statistics say so.