Is it 1999 again? What the dot-com comparison gets right and wrong

The US sharemarket is more concentrated than at any time in six decades, and the bubble talk is everywhere. The comparison with the dot-com era holds on some measures and falls apart on others.

The ten largest companies in the S&P 500 now account for nearly 40 per cent of the index's value, the highest concentration since 1965, according to The Motley Fool. At the peak of the dot-com boom in 2000, the top ten made up about a quarter. Most of today's leaders, from Nvidia and Microsoft to Apple, Alphabet and Meta, are tied together by one theme: artificial intelligence.

"Unlike past periods when the top 10 spanned unrelated industries, today's leaders are closely linked by a common theme," wrote Tyler Frawley of RBC Wealth Management in January 2026. That theme is AI.

Even Reserve Bank of Australia (RBA) Governor Michele Bullock has raised the question. "It might not be a bubble, but it might be," she told a Committee for Economic Development of Australia event on 22 September 2026, adding that if it unwinds in a disorderly manner, "that could have implications."

So how close is the parallel?

Some of the warning signs look familiar. Concentration is higher now than in 2000. The largest companies have outperformed the rest of the market by a margin RBC says is approaching the late-1990s peak, and the big-company index now trades at a roughly 30 per cent premium to an equal-weighted version of the same index, up from about 13 per cent before the pandemic. Sentiment swings violently: in mid-September 2026, a call by AI leaders to slow development wiped more than US$600 billion off the Nasdaq 100 in two days, before a product launch helped add US$3 trillion back.

The biggest difference is profit. Many of the dot-com darlings had little revenue and no earnings. Today's leaders are highly profitable. RBC calculates that at the end of 2025 the ten largest companies made up about 41 per cent of the S&P 500 by value and about 32 per cent of its earnings. That is a gap, and it means investors are paying up for future growth, but it is nothing like the chasm of 1999 and 2000.

The honest risk lies in expectations. Those built into the leaders' share prices are very high, and so much of the market now rides on a single theme. If AI spending slows or returns disappoint, the fall would hit a large slice of the market at once.

For Australians, this matters more than the local sharemarket suggests. The ASX has relatively few big technology companies, but most super funds and popular exchange-traded funds hold large amounts of US shares, and through them, a heavy weighting to the same handful of AI names.

This is not 1999, but it rhymes. The profits are real in a way the dot-com era's were not, which is why the bears have been wrong for years. The concentration is also real, and it is higher now than at the dot-com peak. We do not expect a crash on the scale of 2000 to 2002, but we do expect sharp, sentiment-driven corrections to keep hitting the AI leaders as long as a single theme drives so much of the market. The sensible response is to know your exposure. Check how much of your super and ETFs sits in the same few US names, and decide whether that is a bet you mean to make.