As money pours into artificial intelligence around the world, a senior Reserve Bank of Australia official has begun asking a careful question: how much of it is worth what investors are paying? In an interview on the ABC's 7.30 programme on 8 September 2026, Deputy Governor Andrew Hauser reflected on the difficulty of valuing AI companies and cautioned that some will not survive.

This is a valuation and investment-risk story, distinct from the separate question of whether the AI boom adds to inflation. Mr Hauser was not predicting a crash. He was raising the kind of prudent questions a central banker asks when a rapid Investment wave meets uncertain returns, and pointing to who is bearing the risk.

His comments carry weight because they touch on financial stability as well as the economy, at a time when the cash rate sits at 4.35 per cent and the Bank is monitoring a range of upside risks.

What the RBA Said

Mr Hauser recounted a visit to a Melbourne data centre, describing it as "all funded from overseas". That detail matters for how any losses might ultimately be distributed.

He also relayed a question he heard from banks grappling with AI lending and investment: "how do we value the future income stream?" It is a deceptively simple question that goes to the heart of valuation, because a company's worth rests on the profits it is expected to earn in future, and those profits are especially hard to estimate for a young, fast-moving technology.

On the prospects for individual firms, Mr Hauser was blunt but measured: "Several of them won't...they'll go by the wayside." The framing is important. He was making the ordinary observation that in any investment boom some ventures fail, not asserting that AI as a whole will collapse.

The Key Numbers

Business investment: double-digit growth over 9–12 months. One Melbourne data centre "all funded from overseas". Headline CPI: 3.5%; cash rate: 4.35%.

What Could Happen Next

The questions Mr Hauser raised are unlikely to be resolved quickly. Whether today's AI valuations prove justified will depend on whether the expected income streams actually materialise, something no one can yet confirm.

From a policy standpoint, the Reserve Bank's interest is twofold. It watches the investment boom for its potential to add to inflation, and it watches the financing of that boom for any build-up of risk in the financial system. The overseas funding of projects like the Melbourne data centre currently points to risk sitting largely offshore.

For the wider economy, the practical implication is that a shakeout among AI firms, should it come, need not be a systemic event if Australian exposure is limited. That is analysis drawn from Mr Hauser's own emphasis on foreign funding, not a prediction of events.

Conclusion

Mr Hauser's remarks are best read as a central banker doing his job: asking hard questions about a boom before, rather than after, any reckoning. He has drawn attention to the difficulty of valuing AI income streams, the likelihood that some firms will not make it, and the fact that at least one major local project is funded entirely from overseas. None of that amounts to a forecast of failure. It amounts to prudent scepticism about valuations at a moment when investment is accelerating and returns remain unproven.