Citadel Securities: AI trading has no margin for error, capital expenditure becomes a "double-sided risk"

Market bets on artificial intelligence trading are now deeply embedded in the economy and capital markets. If huge expenditures fail to deliver expected returns, there is little room for adjustment.

Nohshad Shah, Head of Fixed Income Sales for Europe, the Middle East, and Africa at Citadel Securities, noted in a research report that the recent sharp sell-off in tech stocks suggests that investors are becoming increasingly cautious about whether AI’s massive capital spending can ultimately translate into sufficient profits.

He warned that if AI investment and financing continue to retreat, the growth momentum and household wealth that previously helped the economy withstand high cost pressures would be threatened.

Currently, market signals are becoming increasingly tense. The rapid unwinding of AI-related positions last week dragged down tech stocks significantly. Signals of risk aversion in the credit market have risen to the highest level in over a year, and the combination of rising borrowing costs and multiple global risks is making the financing environment more complex.

Meanwhile, AI model developer Anthropic PBC submitted an IPO application this Monday, becoming the first company focused on large-scale AI model development to seek a public listing.

In the report, Shah succinctly highlighted the core dilemma:

As economies and stock markets become increasingly reliant on AI trading, with other risks gathering, there is virtually no margin for error.

He further described capital expenditure as a "two-way risk": cutting spending would raise doubts externally about the sustainability of the entire investment cycle, whereas continued increases in spending bring deeper concerns about financing capacity, execution, and the ultimate rate of return.

This assessment echoes comments made last week by Citadel Securities founder Ken Griffin—Griffin noted that U.S. companies are struggling to balance the scale of investment in AI technology with the value they can extract, warning of bubble risks.

The debate around the pros and cons of AI trading continues to intensify among top investors. Blackstone Group President Jon Gray holds a completely different position, arguing that current demand is growing "almost parabolically" and there is no need for concern.

Pablo Salame, Co-Chief Investment Officer at Citadel Securities, previously stated in a rare public interview that the enormous capital demands of the AI revolution are in direct conflict with lenders whose risk appetite is trending conservative.

Shah pointed out that tariff policies, fiscal stimulus effects, and persistently high energy prices have made financing and operating costs for AI investment rise significantly.

At the same time, growing public opposition to data center construction may further delay the progress of projects and increase associated costs.

So far, the U.S. economy has been able to absorb these pressures mainly because corporate profits remain strong and market confidence is relatively stable.

However, Shah warned that if AI investment and financing experience a prolonged contraction, the growth momentum and household wealth accumulation that previously acted as an economic "buffer" will face a substantial threat. In that case, the economy's overall ability to withstand a high-cost environment will be greatly reduced.