Over $1.7 trillion wiped from the largest AI-exposed stocks in two weeks. The question investors are asking is not whether AI works. It is whether it pays.
Between 10 and 24 July 2024, Meta, Amazon, Microsoft, Alphabet, Apple, Tesla, and Nvidia saw more than $1.7 trillion wiped from their collective market capitalisation. The immediate catalyst was quarterly results from several of those companies. The underlying driver, though, was something more significant: investors are no longer willing to take AI's commercial potential on faith. They want to see returns on the billions that have been poured into the technology, and they want to see them soon.
That is not irrational. It is how capital markets are supposed to function. The narrative around generative AI has been extraordinary in its reach and pace, but narrative is not revenue. At some point, the connection between AI investment and financial performance has to be demonstrated clearly and at scale. The July selloff was a signal that patience has limits.
What comes next is an interesting question, and not one with a simple answer. One possibility is that investors begin to rebalance towards a wider cohort of stocks, reducing the concentration risk that comes with so much capital allocated to seven names. Another is a continued increase in allocations to hedge funds and alternative investment strategies, which offer downside protection, diversification from market direction, and risk-adjusted returns that listed equity concentration tends not to provide.
I spent 18 years at HSBC Alternative Investments managing the infrastructure behind exactly these kinds of strategies. The argument for alternatives as a complement to concentrated equity exposure is not new, but the conditions of mid-2024 made it more compelling. When a handful of technology stocks account for a disproportionate share of index returns, any correction in those names will cascade through passive portfolios in ways that many investors had not fully priced.
The broader point is that AI's integration into financial services, and into the economy more generally, is real and will be sustained. The question the market is now asking is one of timing and distribution: which firms will generate the first genuinely material commercial returns from their AI investment, and over what period? The selloff did not signal a loss of faith in AI. It signalled that the evidence threshold for that faith has risen, and that is probably healthy.