There is a peculiar habit on Wall Street. When something new appears, investors immediately start combing the archives for the last time it all went wrong. Today, that instinct has settled on artificial intelligence. Every mention of vast spending on chips, data centres and digital infrastructure seems to summon the ghost of the dot-com era. Soaring investment, rising enthusiasm, lofty expectations: surely we have seen this film before.

History, though, is rarely so obliging.

Profits, not just promises

As we move through the second half of 2026, equity markets continue to be supported by something investors should never take for granted: earnings. Global equities advanced strongly in the first half of the year, but the more telling story is that corporate profits kept outperforming expectations. First-quarter and second-quarter earnings growth beat forecasts by a wide margin, and analysts have spent much of the year revising estimates upwards rather than downwards, an unusual pattern in recent history. Veteran US market strategist Ed Yardeni captured it well last week, noting that consensus earnings expectations for this year and next have “never” risen so quickly, producing what he called “an earnings-led melt up in the stock market to record highs” (Yardeni Research, 12th August 2026).

That distinction matters. Speculative bubbles are usually marked by share prices rising well ahead of company profits. Today’s environment looks different: earnings have been doing much of the heavy lifting, particularly among businesses tied to AI infrastructure, which JP Morgan estimates have contributed roughly half of S&P 500 earnings growth this year (10th August 2026).

The demand is already here

Scepticism about all this spending is healthy. Investors are right to ask whether the sums being committed to AI infrastructure are justified, since spending alone creates neither value nor returns. What makes this cycle interesting is that the evidence increasingly points to demand that already exists, not demand investors are simply hoping will arrive. AI “token” usage across major model networks, a direct measure of real-world activity rather than theoretical potential, has risen sharply this year (OpenRouter, 10th August 2026), and North American data centres are running at utilisation rates above 99%, a stark contrast to the excess fibre-optic capacity that defined the technology boom of the late 1990s (Alger Capital, July 2026). Put simply, these are not empty buildings waiting for customers to arrive. The machines are busy.

That does not mean the road ahead will be smooth. Inflation remains stubborn, central banks remain cautious, and higher interest rates and oil prices will inevitably test investor conviction. The situation in Iran remains volatile, US equity valuations are elevated, and the bond market is watching America’s growing indebtedness with concern. Markets will almost certainly move through periods of volatility as expectations ebb and flow.

Stronger foundations than the dot-com era

The other crucial difference is the quality of the companies financing this build-out. Twenty-five years ago, much of the internet’s infrastructure was funded by heavily indebted businesses dependent on continual access to capital markets; when financing conditions tightened, many simply ran out of road. Today’s leading AI infrastructure providers sit in a far stronger position. Microsoft, Amazon and Alphabet generally carry robust balance sheets, strong cash generation and modest net debt, and, according to Goldman Sachs (11th August 2026), much of their AI investment is being funded internally rather than through speculative borrowing, though the market is watching closely as the spending eats into their cash flow.

For long-term investors, that distinction is crucial. The future will undoubtedly contain winners, losers, disappointments and excesses, as every technological revolution does. But step back from the daily noise and the picture is one of rising profits, measurable demand and financially resilient businesses investing in infrastructure that looks increasingly essential to the modern economy. That story is also broadening well beyond a handful of technology giants, into healthcare, industrial infrastructure, energy and advanced manufacturing. As Microsoft’s Satya Nadella put it at the company’s recent earnings update:

“We are only at the beginning phases of AI diffusion, and already Microsoft has built an AI business that is larger than some of our biggest franchises.”

Profitability, not just concentration

Two weeks ago, cloud infrastructure company CoreWeave posted strong results and made the point plainly:

“AI is no longer confined to frontier model labs. It is becoming embedded in software, industrial systems, financial markets, enterprise workflows and national security missions. We see that breadth in our backlog, in the new commitments we have signed and in the utilisation and pricing environment across our platform.”

Crucially, this breadth of demand is showing up in profitability, not just prices. The “Mag 7’s” relative valuation has compressed to its lowest level in ten years on a forward price-to-earnings basis, a sign of healthy investor scepticism that we view as a positive market factor (Goldman Sachs, 10th August 2026). And while the top ten US stocks now make up around 40% of the S&P 500, they also generate roughly 38% of the index’s profits (Goldman Sachs, 7th August 2026), a sharp contrast to the relativities seen during the dot-com bubble of 1999 and 2000, when concentration in price ran well ahead of concentration in earnings.

The quiet work beneath the surface

In our view, today’s equity markets look less like another dot-com bubble and more like what a genuine technological transformation looks like when the customers arrive before the capacity does. That distinction matters most for those investing with a long horizon. The past eighty years have contained wars, recessions, inflation shocks, banking crises and countless predictions of economic decline, yet corporate earnings continued to grow through the cycles, and patient investors were ultimately rewarded. Wealth is rarely built through perfect market timing; more often, it is built through owning productive assets and having the discipline to stay invested through periods of uncertainty.

Artificial intelligence may be the most visible symbol of this era, but the deeper story is not technology itself. It is productivity, businesses investing in their own future, and innovation creating new industries while transforming old ones. As has so often been the case throughout history, these are the forces that ultimately drive earnings, wealth creation and long-term investment returns.

The machines may indeed be busy. But for investors, it is the quiet work being done beneath the surface, in productivity, profitability and human ingenuity, and the patience to let it play out, that will matter most.

Written by Pramit Ghose, Global Strategist