Just one month ago, it looked like the glory days might finally be over for US technology stocks, as investors trembled at the prospect of a mighty artificial intelligence bubble.
Experts were increasingly urging investors to reduce their exposure by looking beyond the mega-caps towards banks, industrials, smaller companies and emerging markets.
Many needed little encouragement. They looked at the trillions being poured into AI infrastructure and wondered whether highly leveraged "hyperscalers" such as Alphabet, Amazon, Meta and Microsoft would ever get a return on that investment.
And isn’t investing cyclical anyway? The so-called Magnificent Seven and other tech hopefuls have had it their way for long enough. Surely it must be somebody else’s turn.
The great rotation?
Experts quickly came up with a name for the shift: the great rotation. But there was one problem: it wasn’t so great.
Within weeks, investors were rotating back into technology, with renewed AI enthusiasm sending the Nasdaq careering towards fresh highs. So should we all rotate back again?
This isn’t the first time investors have stepped back from big tech. The bull run may have lasted more than a dozen years, but there has been plenty of volatility along the way, notably in 2022 when valuations crashed.
Investors started looking beyond tech both at the start of this year and again over the summer, says Jason Hollands, managing director of investment platform Bestinvest by Evelyn Partners.
“Every time, enthusiasm for the hyperscalers recovered, and it has just taken off again.” The latest rebound is particularly impressive because it has come despite sharply rising bond yields, he says.
“Higher yields typically have a negative impact on the discounted cash flow models used to value the forecast earnings of growth stocks several years out.”
That should normally hit tech, because it reduces the value of future earnings. There’s another threat too: the prolific amount of debt being issued to finance AI infrastructure.
“You might assume that higher financing costs, alongside the debate about whether the speed of AI frontier models should be slowed, would dent enthusiasm but it hasn’t,” Mr Hollands says. “Investor appetite has proven resilient and corporate profit margins continue to impress and surprise.”
Dot-com parallels
But surely this can’t go on?
Those worried about the sustainability of AI mania regularly draw parallels with the dot-com boom of the late 1990s, but there’s a huge difference. The hyperscalers have deep pockets and are already making real money.
Microsoft’s second-quarter revenue jumped 18 per cent to $90 billion, while Alphabet’s second-quarter revenue rose 24 per cent to $119.8 billion and Amazon’s climbed 20 per cent to $200.6 billion.
Cloud services and AI infrastructure were core growth drivers, although Alphabet and Amazon’s net income figures were boosted by non-operating paper gains from start-up investments such as Anthropic and SpaceX.
Meta Platforms beat the lot, with revenue jumping 28 per cent to $60.8 billion as AI-optimised recommendation and targeting engines supercharged advertising, boosted by its new Muse AI assistant.
“Meta is crushing it, with Mark Zuckerberg’s AI vision finally starting to gain some traction,” Matt Britzman, senior equity analyst at Hargreaves Lansdown, says.
Despite unprecedented AI capital spending, hyperscaler valuations have actually become more attractive as underlying earnings continue to expand aggressively. Share prices have risen, but earnings have risen even faster. No wonder investors have swung back to the sector.
Risks remain
Yet there are risks. It doesn’t mean investors should stop worrying about valuations. Nobody knows how this technological revolution will play out. Yves Bonzon, group chief investment officer at Julius Baer, describes it as a “binary play”.
“Even if everyone agrees on the enormous disruptive potential of AI, nobody really knows what path it will take or where the endgame lies,” he says.
Elon Musk’s SpaceX, which floated in June, is the perfect case study. Its established Starlink operation posted a bumper operating profit of $1.66 billion in the second quarter, a 66 per cent jump.
By contrast, its cutting-edge AI division swallowed $15.83 billion in capital expenditure and posted a $1.26 billion operating loss, even as revenue surged 247 per cent.
Mr Bonzon argues that frontier AI laboratories are the weakest financial link in the AI value chain, yet one of its most important demand drivers. “Their race to develop AI agents with increasing capability is fuelling spending on chips, memory, cloud infrastructure, data centres, and power, even as their own path to profitability remains uncertain, he says.
The risk that rapid technological progress will quickly render AI chips obsolete has so far not materialised, he adds.
Older generations of processors continue to play an important role, reducing the risk that today’s AI investments become stranded assets, Mr Bonzon says. “While we continue to favour diversification, we no longer see a compelling reason to underweight the large US hyperscalers.”
Wider sell-off
But now investors have something else to worry about. October opened with a wider stock market sell-off as investors worried about rising oil prices, inflation and government debt, sending bond yields to multi-decade highs.
Investors rotated out of equities at speed, and not just technology stocks. Banks, industrials, smaller companies and emerging markets also sold off. What goes around, comes around.
Investors had been rotating out of the US dollar, but now they’re rotating back into it too, says Axel Rudolph, chief technical analyst at trading platform IG.
“The dollar just hit its highest level since April 2025, as persistent inflation, resilient US economic activity, and the hawkish Federal Reserve reinforce expectations of further rate hikes,” he says. “Meanwhile, Bitcoin continues to struggle, though at least it has gold and silver for company.”
As bond yields rise, investors are rotating into US Treasuries and government bonds, says Dan Coatsworth, head of markets at AJ Bell. “When investors can earn a relatively high return from government debt with limited risk, equities must work harder to justify their additional volatility.”
The rotation between equity sectors and asset classes will always ebb and flow, Mr Hollands says. “The more durable trend should be towards greater diversification. That is generally more robust than making large all-or-nothing bets based on the latest spurts of market leadership,” he adds.
