The limits of hedging the AI boom

As the spotlight on the AI trade intensifies, even the staunchest bulls are starting to sweat. After last week’s sharp sell-off in global tech stocks, a relief rally now appears to be taking hold. But with investment in chips, data centres and computing infrastructure showing few immediate signs of slowing down, investors should brace for a bumpier ride — and a growing need to hedge their bets.

Indeed, until the vast sums behind the AI spending boom begin to make more sense, the trade will remain hostage to sudden swings in sentiment. For measure, the consultancy Gartner estimates that AI data centre owners will need to generate $8.5tn in revenues by the end of 2030 to cover their capital and financing costs and generate healthy returns. Yet, worryingly, it projects that consumer and corporate spending on AI services that would flow to those providers would reach “no more than $3.6tn”.

In turn, there is no shortage of potential triggers for another sell-off. Poor quarterly earnings reports from a tech giant, a high-profile AI safety failure or growing political opposition to data centres, which has become an important issue in America ahead of next week’s midterm elections, could all conceivably shake investor confidence. The sheer financial demands of sustaining the boom are also becoming a source of instability in their own right.

A looming wave of equity issuance will test investors’ appetite for AI at already frothy valuations. On Monday, Anthropic unveiled ambitious plans for an initial public offering that could value the maker of Claude at up to $1.75tn. OpenAI is expected to follow next year, while Big Tech firms are planning significant share sales to fund capital expenditure. Credit markets are showing signs of strain, too, as AI companies turn increasingly to debt. Recently, bonds issued by specialised cloud computing providers such as CoreWeave have sunk into distressed territory.

As a result, the search for AI hedges and diversifiers is intensifying. Cash is one option, although today’s elevated inflation levels weaken the appeal. Gold, a traditional safe haven, has proved volatile. Instead, many investors are pinning their hopes on government bonds, which might benefit if an AI-led slowdown prompts central bankers to cut interest rates. (That is assuming concerns over sovereign creditworthiness ease.) Others are turning to less tech-heavy equity markets and value stocks. Even Apple is emerging as a favoured Wall Street hedge given the iPhone developer’s relatively limited involvement in AI spending.

Sophisticated investors have more hedging opportunities: options and other derivatives offer insurance, while some are looking to hedge funds and private assets. The very well-heeled are exploring stakes in AI-resistant sectors from live entertainment to football clubs.

None of these options is foolproof. Even equity markets in Europe — an AI laggard — are exposed via the broader tech supply chain through companies such as ASML and other specialist equipment makers. As for derivatives, this month, volatility has pushed up the cost of buying credit default swaps on many AI-linked companies to record highs. And the innovative market for compute derivatives, tied to GPU rental prices and capacity, is still nascent. Hedge fund and private asset investments, meanwhile, are often opaque and hard to exit.

Imperfect protection may be the best investors can hope for. As the AI boom draws in more capital and reaches deeper into the global economy, exposure to the technology is spreading across assets and borders. Hedging against falling tech stocks is easier than escaping a broader AI-led downturn. The bigger the boom becomes, the fewer places there will be to hide.