‘People doing dumb things’: Wall Street boss Jamie Dimon warns of AI frenzy

“There’s always a surprise in a credit cycle,” JPMorgan Chase’s Jamie Dimon told his investors this week.

“This time around, it might be software, because of AI [artificial intelligence].”

The Wall Street boss made his comments as the US sharemarket experienced a sharp sell-off – which stabilised overnight – as fears from earlier this month resurfaced about the potential of AI to wipe out software-as-a-service companies.

That bout of turbulence coincided with a different set of concerns about the potential dangers lurking within the private credit sector after one of the bigger private credit lenders, Blue Owl Capital, was forced to freeze payouts from one of its funds after a wave of redemptions.

Bizarrely, another factor in Monday’s market jitters was a Substack post at the weekend in a popular finance blog that outlined a fictional scenario under which AI leads to mass unemployment and declines in wages, causing a steep fall in consumer spending, with companies increasing their investment in AI in what it described as “a feedback loop with no natural brake.”

In turn, that would precipitate a massive crisis for private credit and mortgage lending and a market crash late next year that wipes out 57 per cent of the S&P 500’s value, the post outlined.

‘My own view is people are getting a little comfortable that this is real: these high asset prices and high volumes, and that we won’t have any problems.’

Jamie Dimon

The scenario is, of course, hypothetical. What it did, however, was provide extra focus on the vulnerability of software companies and private credit providers.

Dimon’s comments, while centred on AI, were broader.

“Unfortunately, we did see this in ’05, ‘06 and ’07, almost the same thing – the rising tide was lifting all boats, everyone was making a lot of money.” he said.

“My own view is people are getting a little comfortable that this is real: these high asset prices and high volumes, and that we won’t have any problems,” the influential banker warned.

“I don’t know how long it’s going to be great for everybody. I see a couple of people doing some dumb things.”

Inevitably, he said, the economic cycle would turn and there would be a wave of borrower defaults that would broadly affect lenders and impact industries few people expected.

Given that this particular sharemarket cycle has been running for more than three years and has been largely driven by the excitement about AI – it really kicked off when OpenAI launched ChatGPT in late 2022, with the market rising more than 40 per cent since then – it isn’t surprising that investors are becoming more cautious and more sensitive to perceived threats, or that a banker like Dimon, who led JPMorgan through the 2008 global financial crisis, is conscious of the potential for a repeat episode.

What’s particularly interesting about the events at the start of this week is that they bring together different strands of an AI-focused story, one latent with threats to the stability not just pf the sharemarket, but both non-bank and bank balance sheets.

Dimon said he was starting to see parallels to the era ahead of the 2008 financial crisis, with some lenders making riskier loans to boost their incomes.

If there were to be another financial crisis, it would probably have an AI epicentre. That’s because the tremors within private credit relate largely to the sector’s significant exposures to not just software companies, but increasingly to the AI developers and their infrastructure that are threatening the software sector.

The software and data centre risks also come via leveraged loans to private equity firms, which brings another strand of the financial system into play.

It’s also the case that banks have been growing their lending to private credit firms and private equity firms, so there are conduits into the regulated core of the financial system.

The speed at which the AI sector has been developing has caught investors and lenders by surprise, unprepared for the disruption that AI might cause, so early in its development, to existing businesses.

The Blue Owl fund’s experience is instructive. Private credit funds, and the exchange-traded funds that have grown their exposures to private equity credit, hold illiquid assets.

That doesn’t faze institutional investors, who invest for the long term and welcome that the assets’ values aren’t – unlike listed assets – marked to market in real time.

The growth in retail funds with private equity and credit exposures – funds that promise liquidity despite the illiquidity of the underlying assets – has, however, created the potential for mismatches between what their investors expect and what they can deliver, and therefore the potential for the kind of run that led to Blue Owl’s cutting of redemptions from its fund.

It’s not just the potential for disruption of existing funds and businesses, however, that might be a threat to financial stability.

The sums involved in AI are unprecedented. Companies, some – like the “hyperscalers” – valued at trillions of dollars and others many hundreds of billions, are investing around $US700 billion ($990 billion) in AI this year and plan to spend even more on training, chips and data centres next year and beyond.

Where the first phase of AI was funded with equity, some from private equity and venture capital firms, the companies are increasingly tapping debt markets and private credit to fund projects whose returns and timing are extremely uncertain.

At this phase of AI’s development, it is a heavily cashflow-negative sector, even if some of the biggest spenders – companies like Meta, Amazon, Google and Microsoft – are subsidising their investments from cashflows from their legacy operations.

It would take very little to trigger an implosion in AI valuations, cutting off access to both equity and debt markets, and triggering meltdowns in equity values and in the value of existing loans.

It is the overlaps between the AI companies, their investors and lenders and the multitude of companies engaged in building the massive data centres and energy and water infrastructure to support them that could turn what might appear to be a sectoral issue into one with threatening systemic and economic implications.

Anthropic’s CEO Dario Amodei has described the risk-reward equation confronting AI companies by saying that the escalation in the costs of the computing power now required for AI could be “ruinous” if companies got their timing wrong – if they invested too much too early and weren’t generating the revenue required to support their investments. It was Anthropic’s release of its Claude Code tools earlier this month that ignited the sell-off in software stocks.

Amodei’s comment applies not just to those AI companies building their large language models and the infrastructure to train them, but to all those exposed, directly or indirectly, to their activity.

The sheer scale of the numbers involved means that a relatively small mismatch in the timing of the spending and revenues could trigger contagion within the sector, which would then infect the other segments of the system with AI exposure.

There doesn’t have to be any judgement of the merits and eventual worth of the investments in AI and its infrastructure to recognise that investors’ capitalisation of the sector’s prospects has been so aggressive that it has stepped up its vulnerability to any misstep or disappointment.

Add to that the concentration of exposures to the sector from apparently diverse segments of the financial system – private equity, private credit, institutional and retail investors and the banking system – and you could have a threat to systemic stability developing just as the Trump administration’s deregulation of finance and banking encourages more risk-taking.

It may not come to that. Hopefully, it won’t.

But there is now an increasingly fine line between AI success and failure. There is a lot riding on which side of that line we end up – including the fates of investors and lenders to the sector, those threatened by AI, and the stability of the entire financial system.

The Business Briefing newsletter delivers major stories, exclusive coverage and expert opinion. Sign up to get it every weekday morning.