‘The bigger risk may be an earnings bubble rather than a valuation bubble’

🔗 원문 전체 보기 → The Irish Times

요약

Top 1000: The AI boom looks bubbly. But should investors really panic? Proinsias O’Mahony reports

본문

It is not hard to see why some investors are getting nervous about the stock market.

Elon Musk’s money-losing SpaceX recently went public in the biggest initial public offering (IPO) in history, with its valuation quickly topping $2 trillion and, at times, exceeding 100 times annual sales – an extraordinary multiple for a mega-cap company. Meanwhile, an even bigger AI flotation is in the pipeline, with Anthropic reportedly considering a $2trillion-plus IPO.

And just look at 2026’s high-fliers. The 10 best-performing stocks in the S&P 500 this year are all beneficiaries of the AI infrastructure boom, notes Creative Planning strategist Charlie Bilello. SanDisk is up almost 600 per cent, Dell has almost quadrupled, Micron has surpassed a $1 trillion market capitalisation after more than tripling, while Intel, AMD and others have all bagged triple-digit percentage gains.

Then there is the spending. The world’s biggest technology companies are pouring trillions of dollars into data centres, chips and other AI infrastructure, with investors waiting to see how much of that money will ultimately translate into profits.

READ MORE

It all looks a little feverish, exuberant, bubbly – doesn’t it?

A casual glance at the headlines – “The AI boom could trigger the next crash, central banks warn”, “US tech stock correction likely, warn ECB economists”, “Norway wealth fund warns of AI-driven stock market bubble” – might suggest as much.

That last headline refers to Nicolai Tangen, head of Norway’s $2.4 trillion sovereign wealth fund – the world’s largest – warning that an extreme market collapse could halve the value of its portfolio. Tangen has been surprised by how resilient markets have proved. “If you went back two years and told me this is going to happen with the Hormuz strait, trade barriers, geopolitical tensions, and so on, I would never have thought that the market would be as resilient as it is,” he said.

All of which sounds distinctly alarming. Before reaching for the panic button, however, there are a few things worth considering.

Extraordinary earnings

For one, there is something underneath the extraordinary share-price gains: extraordinary earnings. S&P 500 earnings are on course to rise around 50 per cent year on year, notes Bilello, with second-quarter earnings beating expectations by a record 29 per cent. Analysts expect earnings to grow 32 per cent in 2026, more than twice what they expected at the beginning of the year.

Clearly, this is not the late 1990s, when companies with no earnings routinely commanded extraordinary valuations.

Indeed, the valuation picture is more complicated than the headlines suggest. The US technology sector has risen around 65 per cent since the start of 2025, says Ritholtz Wealth Management’s Ben Carlson, but its forward price-earnings ratio has fallen from around 29 to 22. “Earnings have been growing so fast that even though the stock market keeps increasing, valuations have actually fallen,” notes Carlson.

That same point is stressed in a recent Goldman Sachs report. Indeed, industrial stocks trade at higher valuations than tech stocks, says Goldman, with the technology sector now trading in line with its 20-year average.

And this idea is also emphasised in a recent Schroders report. Far from ignoring the risks, markets are “wary”, says Schroders, with “valuations for many of the big tech companies” at or near “their cheapest levels for a decade”.

Secondly, while there has been much angst about indices being overly reliant on high-flying AI giants, the market has in fact been broadening. The equal-weighted S&P 500, which gives roughly the same weight to each company rather than weighting them by market value, has outperformed the conventional index. Most of the magnificent seven stocks have underperformed the index in 2026. Smaller companies, value stocks, and non-US firms have outperformed.

In other words, it would be wrong to look at a few spectacular share-price gains and conclude that the entire stock market has become detached from reality.

Concerns

Still, there are causes for concern. The obvious one is whether the huge growth in earnings can continue, and whether the enormous sums being spent on AI will ultimately pay off.

The hyperscalers – Alphabet, Amazon, Microsoft, Meta and others – are spending at an extraordinary rate on data centres, chips and electricity, while increasingly turning to debt, leases and other forms of financing to fund the buildout. The Financial Times recently noted that Goldman Sachs estimates the major hyperscalers have accumulated some $1.5 trillion of lease commitments, around $1 trillion of which relates to leases that have not yet commenced and therefore do not appear on their balance sheets. Morgan Stanley has identified a further $982 billion of future purchase commitments for computing power, chips, equipment and electricity across Alphabet, Microsoft, Amazon, Nvidia and Oracle.

The tech giants remain highly profitable and cash-rich, but investors are increasingly asking whether the returns from AI will justify the enormous sums being committed to it. The negative share-price reaction to Meta’s increased spending plans, for example, is a reminder that investors are certainly not cheering every dollar committed to AI.

There is another complication: some of the earnings growth is coming from gains on companies’ holdings in AI businesses such as SpaceX and Anthropic. Acadian Asset Management’s Owen Lamont calculates that unrealised gains on holdings by Alphabet, Amazon, and Nvidia added $69 billion to S&P 500 earnings in the first quarter. Strip them out, he says, and earnings growth fell from 28 per cent to 16 per cent. Amazon alone reported $16.8 billion of pretax gains on its Anthropic investment, helping to push first-quarter net income to $30.3 billion. These are genuine accounting gains, but they are rather different from recurring operating profits.

As Goldman Sachs and others have cautioned, the bigger risk may be an earnings bubble rather than a valuation bubble, with complacent investors assuming that today’s exceptional AI-driven profit growth can continue indefinitely.

Some investors might thus be tempted to diversify away from the tech-heavy US, but it’s not as straightforward as it sounds. Emerging-market indices are themselves increasingly dominated by technology giants such as Samsung, TSMC, and SK Hynix, which together account for around 30 per cent of the MSCI Emerging Markets index. South Korea’s recent extreme volatility has highlighted the risks of concentrated markets, and it’s increasingly clear that investors can move around the world without necessarily escaping the AI trade.

Timing the market

However, while there might be good reasons for caution, none of this tells investors when to sell, as evidenced by the long history of warnings from central banks. The European Central Bank issued warnings in 2018 (“very high valuations by historical standards”); in 2019 (“stretched” valuations “may be subject to further adjustment”); in November 2020 (risk of an “abrupt equity market correction remains elevated”); weeks after the bottom in November 2022 (valuations “still appear stretched”); in 2023 (it warned of “disorderly corrections”); and in 2024, 2025, and now 2026, with recurrent warnings about high valuations, market concentration, and AI-related risks.

Regulators must highlight vulnerabilities and potential sources of instability, but such warnings are not actionable signals for long-term investors – that much is clear.

And there is a revealing contrast between the alarm in some of the headlines and the advice of the aforementioned Nicolai Tangen. Despite his warnings about the potential for a market collapse, the Norwegian fund will not try to take profits or move out of the market. “I would say the way to make money is, one, be very, very long term – don’t change your strategy – and be well diversified,” he said.

That’s echoed by Ben Carlson. “Will this last forever?” he asks. No. Will it end badly at some point? Probably. Can anyone forecast when? No. So what should the long-term investor do? Stay invested, appropriately diversified, and accept that periods of unpleasantness are unavoidable.

Markets rise and markets fall. “On average each year, at some point the market falls by 15 per cent and rises by 23 per cent”, notes Schroders. Volatility is normal, and long-term investors must accept it rather than fearing it.

The age-old advice about sitting tight and owning the market in good times and bad still applies. That may be less exciting advice than predicting the next crash, but it is also more useful – “there are reasons to worry” is not an investment strategy.