News

The AI boom is increasingly being seen as a credit-fuelled investment bubble with potentially greater systemic risks than the dot-com crash of 2000. Investors must therefore remain vigilant. “Free cash flow and balance sheet quality are coming into focus,” says Thorsten Fischer, Managing Director and Head of Portfolio Management at Moventum AM.

The market is changing its view of the AI boom. For a long time, comparisons were drawn with the dotcom bubble, the bursting of which in 2000 had only a relatively minor impact on the economy as a whole. Now, however, the 2008 financial crisis is coming to the fore as a point of reference – and with it, a potentially greater systemic risk. “The dotcom analogy falls short,” explains Fischer. Unlike many internet companies in 2000, today’s AI giants such as Nvidia, Microsoft, Alphabet and Meta are already generating high turnover and profits running into the billions.

That might be reassuring. Yet the real danger of the AI boom lies not in overvalued shares, but in a credit-fuelled investment bubble. “The AI race is creating pressure to invest,” says Fischer. “Hyperscalers are investing regardless of short-term profitability, as nobody wants to risk falling behind technologically.” As a result, massive investments in data centres, chips, energy supply and network infrastructure are increasingly being financed through debt.

Overcapacity thus becomes a real risk. Should demand for AI services fall short of expectations, these investments in the future could turn out to be bad investments. “A credit bubble would be far more dangerous than a stock market bubble,” says Fischer. For whilst share price losses primarily affect shareholders, loan defaults can put a strain on banks, corporate financing and, ultimately, the entire real economy.

Credit default swaps (CDSs) therefore serve as an early indicator of a bubble. Rising CDS spreads signal higher credit default risks and, at the same time, make it more expensive for companies to refinance. The US corporation Oracle is seen as a warning sign; its CDS spreads have recently risen sharply, which is usually an indication that market participants are pricing in a significantly higher probability of default. “This reflects concerns about the aggressive financing of AI infrastructure,” said Fischer.

The major software companies are also significantly increasing their debt. Google’s parent company, Alphabet, plans to spend around 200 billion US dollars on AI infrastructure this year alone. As a result, the share price has recently come under pressure on the stock market, despite the company having significantly exceeded forecasts in the second quarter: the high level of AI investment is putting investors off. The situation is similar at Meta: despite solid revenue growth, the share price fell by more than seven per cent at one point following the announcement of the quarterly results.

According to Fischer, AI differs fundamentally from traditional platform models. Unlike earlier internet companies, every additional AI query incurs running costs for computing power, energy, chips and cooling. Free cash flow is therefore becoming increasingly important: investors are paying ever greater attention to whether the enormous investments will generate sufficient cash inflows in future, rather than assessing revenue or profit growth alone. “Rising capital expenditure is putting a significant strain on free cash flow,” explains Fischer. Furthermore, rising interest rates are exacerbating the situation. Higher financing costs and rising risk premiums are increasing the pressure on highly indebted companies and making the capital requirements of the AI push significantly more expensive.

“The key question for investors is no longer whether AI shares are expensive, but whether the massive infrastructure investments will pay for themselves financially,” says Fischer. The sustainability of the AI boom will depend more on free cash flow and return on capital in future than on revenue growth or technological leadership.

Should demand for AI applications fall short of expectations, the current investment momentum could give rise to a credit-fuelled bubble with systemic risks. For investors, therefore, balance sheet quality, debt ratios, refinancing costs and cash flow generation are becoming increasingly important factors in the valuation of AI companies.