The rapid growth of artificial intelligence investment has created huge opportunities for investors, but a sharp sell-off in AI-linked stocks in July has highlighted the risks of using borrowed money to chase the boom, according to Butterfield.
In its August Investment Views, the bank said stocks in chipmakers and other semiconductor companies bore the brunt of the sell-off, with the Philadelphia Semiconductor Index falling almost 30% from its peak to trough during July.
However, it said the decline needed to be viewed in context – the index was still up about 60% between the start of the year and the end of July, with a recovery emerging in late July and early August.
The more interesting story, Butterfield said, was what happened beneath the headline figures.
While some of the market’s biggest AI winners fell sharply, six of the 11 major stock market sectors recorded positive returns in July.
Energy and financial stocks performed particularly well, helping to keep overall market volatility relatively contained.
Butterfield attributed much of the AI sell-off to market trading dynamics rather than a fundamental weakening of the AI sector.
Hedge funds have been heavily involved in AI-related stocks, with some borrowing money to amplify their returns.
As semiconductor shares began falling, leveraged investors were forced to reduce their positions to limit losses, adding to the downward pressure.
The Situational Awareness fund made headlines as it offloaded a large amount of AI-linked stocks, contributing to a rapid fall in prices.
The fund is managed by 25-year-old German artificial intelligence researcher and investor Leopold Aschenbrenner and grew from US$225 million in 2024 to a peak of around $45 billion.
Butterfield said Aschenbrenner had made “some very good investment calls” but added that the episode was “a reminder of the power of leverage”.
The fund subsequently fell to around $10 billion as leverage magnified losses.
The sell-off was also fuelled by concerns about the amount of debt being used to finance the construction of data centres, the emergence of cheaper open-source AI models and the potentially high computing costs being incurred by companies using AI.
But investors with cash stepped in as prices fell, helping to stabilise AI-linked stocks.
The report’s wider message was the importance of diversification rather than relying too heavily on a single area of the market.
Butterfield said the performance of other sectors helped cushion the impact of the technology sell-off. The MSCI World Index ended July up 0.5% in US dollar terms, although it fell 1.1% in sterling terms.
The bank also pointed to strength in corporate earnings.
US companies were reporting earnings about 29% above estimates, according to FactSet, which Butterfield said represented the largest quarterly earnings beat since tracking began in 2008.
When gains linked to companies’ holdings in other businesses were excluded, underlying S&P 500 earnings growth was still about 31%.
Financial companies were among the stronger performers, with global financials recording earnings growth of 19% over the previous year.
Butterfield said financials had benefited from a solid economic backdrop, stronger trading revenue and increased merger and acquisition activity and initial public offerings.
The bank’s analysis also pointed to opportunities outside the US technology sector. Chinese equities rose 8.7% in July, while European earnings growth reached 22% in the second quarter.
The report concluded that the sharp fall in AI-linked shares could ultimately prove a healthy development after their strong run, while the performance of other sectors demonstrated the value of holding diversified portfolios.
Butterfield publishes its Investment Views monthly, covering the global economy and financial markets.
