Is there an earnings bubble? AI boom fuels debate over US stock valuations

As US stocks continue to trade near record highs, investors are once again debating whether markets have entered bubble territory.

Elevated valuations and persistent geopolitical tensions have fueled concerns that equities may have become disconnected from fundamentals.

Increasingly, however, market strategists argue that the more important question is not whether there is a stock market bubble, but whether an earnings bubble is forming.

The distinction is significant because corporate profits have continued to exceed expectations.

Earnings forecasts have risen sharply across sectors, helping justify higher valuations even as some investors warn that current estimates may prove difficult to sustain.

Earnings growth has accelerated beyond previous recoveries

Wall Street analysts are forecasting approximately 25% earnings growth for 2026 and nearly 18% growth for 2027, according to Bloomberg data.

Several investors note that the pace of earnings upgrades is among the strongest since the post-pandemic recovery.

Technology has led the upgrades, with earnings forecasts rising by more than 30% this year.

Communication services have also seen upgrades exceeding 20%, while energy earnings expectations have increased for different sector-specific reasons.

Importantly, analysts say earnings growth is broadening beyond a handful of mega-cap technology companies into other industries.

The rapid increase in earnings expectations has also helped keep equity valuations from expanding as quickly as stock prices.

US stocks currently trade at roughly 20 times forward earnings estimates, according to Bloomberg data.

While elevated, that multiple remains below levels reached during the 2020 recovery and well below valuations recorded during the dot-com bubble.

"We are in the middle of the strongest earnings upgrade cycle since the commodity supercycle," said Arun Sai, senior multi-asset strategist at Pictet Asset Management in a Financial Times report.

Some investors warn an earnings bubble may be emerging

Not everyone believes current earnings expectations are sustainable.

Ben Inker, co-head of asset allocation at GMO, said forecasts for the coming years have risen unusually quickly.

Forecasts for the next year's profits have increased by almost 20% in just six months, representing the fastest rise since 2021.

"What we are due for, in the market, is the eventual realisation that they will not come true," Inker said.

Capital Economics also warned this week that "AI-related equity markets may be approaching a point where earnings expectations and capital expenditure assumptions become difficult to sustain" and that any correction could "trigger a broad equity market pullback".

Sarah Ketterer, chief executive of Causeway Capital Management, also noted that low valuation multiples may not necessarily indicate attractive buying opportunities if companies are approaching peak earnings.

Semiconductor cycle remains central to the debate

Much of the current discussion centers on semiconductors, where extraordinary demand driven by artificial intelligence has fueled record earnings.

The industry has historically been highly cyclical.

Companies such as Micron Technology have previously traded at very low earnings multiples during peak earnings periods because investors anticipated future oversupply.

This cycle appears different in the near term because supply constraints remain significant.

Taiwan Semiconductor Manufacturing Co. has outlined approximately 40% growth in capital expenditure, while Samsung Electronics plans to invest $73 billion (approx. R 1,2 trillion) in capital spending and research and development.

SK Hynix and Micron are also expanding production capacity.

However, much of that additional manufacturing capacity is not expected to become operational until 2027 or 2028.

As a result, current supply shortages are expected to continue supporting earnings over the next 12 to 18 months.

Memory industry revenue illustrates the scale of the expansion.

The sector generated roughly $200 billion (approx. R 3,4 trillion) in revenue during 2025, with forecasts pointing to approximately $600 billion (approx. R 10,3 trillion) in 2026 and nearly $800 billion (approx. R 13,7 trillion) in 2027.

Even so, some analysts believe long-term earnings projections may already reflect excessive optimism.

AI spending is supporting broader corporate earnings

Beyond semiconductor companies, analysts point to another important development.

Large technology companies including Alphabet, Meta Platforms and Microsoft are no longer simply accumulating cash through share buybacks and balance sheet expansion.

Instead, they are deploying substantial capital into AI infrastructure, particularly data centers.

Analysts estimate that data center investment now represents more than 2% of US gross domestic product through new capital expenditure.

Those investments create broader economic activity by generating demand for construction, electrical work, logistics, and industrial materials.

The resulting economic multiplier has contributed to improving earnings across industries beyond technology.

This has created an unusual backdrop in which consumer spending remains pressured by higher borrowing costs while corporate earnings continue to strengthen.

Concentration and valuations remain warning signs

Despite improving earnings, concerns remain about market concentration.

According to Bank of America, the "AI Big 10" now account for approximately 41% of the S&P 500, a concentration similar to technology and telecommunications companies during the dot-com era.

The group includes Nvidia, Microsoft, Alphabet, Amazon, Meta Platforms, Apple, Tesla, Broadcom, Micron and Advanced Micro Devices.

The Nasdaq Composite rose 21.4% during the second quarter of 2026, marking its strongest quarterly performance since the post-pandemic rebound.

The gains were fueled largely by continued investment in AI infrastructure, semiconductor companies and enthusiasm surrounding SpaceX's public listing.

Some strategists also point to traditional valuation measures as evidence of stretched markets.

The cyclically adjusted price-to-earnings ratio (CAPE), popularized by economist Robert Shiller, has climbed above 40 for the S&P 500.

Analyst Joachim Klement of Panmure Liberum argues current conditions differ from the dot-com bubble because today's leading AI companies generate substantial profits.

However, he also warns that investors may now be paying premium valuations on earnings that are themselves unusually elevated.

Key risks investors are watching

Despite strong earnings momentum, investors continue monitoring several potential risks.

Inflation remains a primary concern. Any renewed acceleration in price pressures could prompt the Federal Reserve to keep interest rates higher for longer, increasing financing costs while slowing consumer demand.

Oil prices also remain an important variable. Rising energy costs would support profits for energy producers but could pressure consumers and businesses across much of the economy.

The biggest market risk, however, remains a slowdown in earnings growth itself.

If earnings growth falls materially below current expectations while investors simultaneously reduce the valuation multiples they are willing to pay, markets could face both earnings downgrades and multiple compression at the same time.

Investors are also closely watching whether earnings strength continues expanding beyond technology and whether AI-related capital spending remains sufficient to justify current expectations.

For now, corporate earnings continue to provide support for elevated equity valuations.

Whether those expectations prove sustainable may ultimately determine whether today's market represents justified optimism—or the early stages of an earnings bubble.