When new innovations emerge that become highly sought after by the market, and if supply and demand become misaligned, the consequences can impact the entire economy. Currently, the most closely watched phenomenon in the market is artificial intelligence or AI which is attracting massive capital inflows, pushing U.S. stock markets to repeatedly hit new all-time highs.
Throughout 2025, the S&P 500 index has risen by 16%, led by AI market winners such as Nvidia, Alphabet, Broadcom, and Microsoft, which have captured significant investment. At the same time, concerns have grown alongside the hundreds of billions poured into the market. It is expected that AI investments by big tech firms like Microsoft, Alphabet, Amazon, and Meta will collectively increase by over 34% to reach $440 billion by next year.
Besides the major players, OpenAI plans to invest more than $1 trillion in AI infrastructure, an astonishing figure given the company is still unprofitable. Another concern is the cyclical investment loop between OpenAI and large U.S. publicly traded tech companies, raising questions about systemic risks.
Historically, it is common for massive capital to flow into new technologies that transform society, as seen with the advent of electricity, railroads, and the internet. Expertsbelievethis time will be no different from the past.
"Sometimes building infrastructure takes longer than the short-term needs of the economy, but eventually the railroad was completed and the internet became an integral part of our lives,"said Brian Levitt, global chief investment strategist at Invesco.
However, questions arise about how much further the market can rise. The S&P 500 recently hit new all-time highs, valuations are becoming stretched, and if AI fails to meet expectations, how much value could be lost? Stocks like Nvidia, Microsoft, Alphabet, Amazon, Broadcom, and Meta together account for nearly 30% of the S&P 500. This means that if AI stocks are sold off simultaneously, the overall market index would be severely impacted.
Bloomberghas analyzed and compared historical bubble events with the current AI market situation as follows.
One of the simplest ways to assess whether AI-driven tech stock gains have gone too far or too fast is to compare them with past bull markets.
Data from a study of 10 stock bubbles worldwide since 1900 by Michael Hartnett, a strategist at Bank of America, found that past bubbles lasted on average about two and a half years and delivered an average trough-to-peak return of 244%.
Compared to the current cycle, the AI-driven market rally is entering its third year, with the S&P 500 up 79% since late 2022, while the Nasdaq 100 technology index has surged 130%.
Although these numbers do not definitively prove a bubble, Hartnett warns investors not to rush to exit the market even if they believe it is in a bubble phase. The reason is that the late stage of bull markets often sees the strongest stock price gains, and missing this final phase could mean losing significant returns.
For risk-averse investors, Hartnett recommends hedging strategies by diversifying into undervalued assets such as UK stocks or energy sector shares, which are considered value plays in the current market.
Currently, the top 10 stocks in the S&P 500 make up about 40% of the index, an unprecedented level of concentration not seen since the 1960s. This concentration has led some investors to pull back from the market.
Previously, Ed Yardeni, a legendary senior Wall Street analyst, commented in December that"It no longer makes sense to recommend investors overweight technology stocks."
However, financial market historians note that while today's concentration seems extreme to modern investors, similar cases have occurred historically.
Paul Marsh, a professor at London Business School who studied global asset returns over 125 years, pointed out that the proportion of large-cap stocks dominating U.S. market value was similar during the 1930s and 1960s.
By nature, bubbles are difficult to detect while forming and become clear only after they burst. According to Dario Perkins, an economist at TS Lombard, this is because bubble debates often revolve around "fundamentals," and the valuation metrics investors use can evolve with time and perspective.
For example, compared to the dot-com bubble era, today's major AI companies have significantly lower debt-to-earnings ratios than past tech firms like WorldCom. Moreover, companies like Nvidia and Meta have reported concrete profit growth from AI, unlike 25 years ago when stock prices soared on expectations but revenues and earnings did not justify valuations.
Despite stronger fundamentals, credit risk in the AI market has started to unsettle some investors. A recent example is Oracle, which issued $18 billion in bonds last September, causing its stock to drop 5.6% the next day and declining over 37% since then.
Furthermore, Societe Generale estimates that in 2026 alone, Meta, Alphabet, and Oracle will need to raise a combined $86 billion to invest in AI, indicating that
while AI companies are more profitable and financially stable than during the dot-com era, debt burdens and financing costs could become vulnerabilities for the AI trend in the future.
Looking at U.S. stock market valuations, the S&P 500 is now valued near the all-time highs seen in the early 2000s during the dot-com bubble.
This evaluation is based on the CAPE Ratio (Cyclically Adjusted Price-to-Earnings Ratio), developed by Nobel laureate economist Robert Shiller, which divides stock prices by the inflation-adjusted 10-year average earnings to reflect true long-term value.
Optimistic investors argue that although market values have risen due to tech and AI stocks, the pace of increase is clearly slower than during the dot-com era. For example, in 2000, Cisco Systems traded at over 200 times trailing 12-month earnings, while today Nvidia, central to the AI trend, has a P/E ratio below 50.
This difference suggests that although AI stock prices are high, they remain reasonable and not like the late-stage dot-com bubble.
Richard Clode, a fund manager at Janus Henderson, explained that stock bubbles usually occur when prices become completely detached from earnings growth. He believes that ongoing discussions about valuations, risks, and price rationality indicate the market has not yet entered an irrational frenzy, which typically signals the final stage before a bubble bursts.
Throughout the past year, debate over whether the stock market is facing a "bubble" has been intense, especially in November and December, fueled by warnings from prominent investors like Michael Burry and signals from the Bank of England.
Bloomberg data shows that in November alone there were over 12,000 news stories mentioning "AI bubble," nearly matching the total from the previous ten months combined.
Investor sentiment also reflects growing concern. A December Bank of America survey found that investors consider an AI bubble a low-probability tail risk but one with potentially severe portfolio impact.
More than half of respondents said that the "Magnificent Seven" stocks represent the most crowded investments on Wall Street today, indicating heavy capital concentration in a few names.
This atmosphere contrasts sharply with the dot-com bubble era, when the market was filled with unquestioned excitement and no fear that the internet would revolutionize everything.
Venu Krishna, head of U.S. equity strategy at Barclays, explained that investors are now more questioning whether AI investments will deliver worthwhile returns, especially as tech companies increasingly issue bonds and take on more debt. However, Krishna views this questioning as a positive sign for the market.
"Overall, close scrutiny is good because skepticism helps prevent extreme market moves that could lead to collapse,"he said.
Source:Bloomberg
Follow the Facebook page: Thairath Money at this link -https://www.facebook.com/ThairathMoney