By Enguerrand Artaz, Strategist, La Financière de l’Échiquier (LFDE)

It is a long-standing tradition, particularly in English-speaking countries, to indulge in a little scaremongering on the eve of All Saints’ Day. In financial markets, the effects are usually confined to a few sweets shared among colleagues and a pumpkin or two decorating the office. This year, however, markets seem to have decided to stage a Halloween encore a few days late. In the wake of the US midterm elections, in which the Democrats regained control of both houses of Congress, US equities suffered a bloodbath worthy of Stephen King, with tech stocks hit hardest. On Thursday 5 November, the Nasdaq 100 fell 6.4%, its worst session since the Covid crash, while the SOX semiconductor index plunged 12.2%, its second-largest one-day fall on record after 16 March 2020. There was no shortage of superlatives to describe the session, whose contagion then spread to markets around the world, starting with those most exposed to the AI ecosystem, such as South Korea. At the root of the rout lay a build-up of fears to which investors finally capitulated.

First, fears over data centres. The blue wave in the midterms revived concerns about a series of obstacles to building this infrastructure, which is essential to AI. Many Democratic candidates had campaigned on introducing moratoriums to curb the construction of new capacity. The issue is particularly sensitive in Texas, which accounts for 27% of the computing capacity built or planned in the United States. There, two Democrats were elected, including Gina Hinojosa, one of the party’s leading voices on data centres. The new governor wasted no time in matching words with action: the day after her election, she called for a freeze on all decisions concerning public land near SpaceX’s Starbase site, which could be transferred to Elon Musk’s company. Although no large-scale moratorium has yet been launched, decisions taken in Texas could have national implications and weigh heavily on financial markets. The consequences would be direct. Hyperscalers and AI labs could be forced to slow the pace of their investment. That in turn would lead to downward revisions in the outlook for the segments that depend on that spending, such as chipmakers, memory manufacturers and neoclouds, whose share prices have soared in recent years.

Second, fears over AI financing. The issue has been a thorn in the market’s side for several weeks, and it took a fresh turn on Thursday evening. A rumour swept Wall Street that a distressed alternative asset manager was being forced to offload a block of data-centre-related loans. Bloomberg later reported that the seller was in fact ING, which is seeking to sell around $500 million of such loans through an auction, with bids due on 19 November. The amount is modest by market standards, and there is nothing to suggest the bank is under pressure on these loans. Even so, the transaction is unusual, coming at a time of growing distrust of AI infrastructure financing.

Finally, the fear of being trapped. As the market’s slide accelerated, many investors, retail traders in particular, rushed to unwind their positions. Yet a large share of these positions were held through leveraged ETFs, which had attracted record inflows of $15 billion in October, according to Bloomberg. Panic selling of these leveraged products amplified the market’s decline, especially in AI-related segments, where retail investors own a large share.

Episodes of market capitulation are frightening because of their sheer brutality. Driven by fear, they tend to feed it, yet with hindsight they often turn out to be attractive entry points. It is far too early to tell whether that will be the case this time, particularly as the concerns behind this episode are legitimate. Either way, it is a reminder that, for the direction of markets and for asset allocation decisions alike, the future of AI remains by far the decisive factor.

Disclaimer: These data and opinions from LFDE, as well as the sectors and stocks mentioned, are provided for information purposes only and do not constitute an offer to buy or sell any security, investment advice or financial analysis. Sectors and stocks are cited as examples only, and their inclusion in portfolios is not guaranteed. Past performance is not a reliable indicator of future performance and is not constant over time.

Find all our Strategic Case articles