The artificial intelligence boom is facing mounting concerns about becoming a bubble, with South Korea's Kospi closing more than 6% lower on Thursday as AI-linked chip heavyweights SK Hynix and Samsung Electronics plunged 11% and 8% respectively. According to Moneycontrol, this follows a series of major AI-led market corrections, including IBM's worst stock market decline in nearly six decades in July 2026, falling as much as 25% and wiping out around $70 billion in market value. The SaaSpocalypse in February 2026 saw global IT and SaaS stocks face major corrections after Anthropic released powerful agentic AI tools, while Nvidia lost nearly $600 billion in market value in a single day in January 2025, marking the biggest one-day market capitalisation loss for any company in US history at the time. Fresh pressure on semiconductor and technology shares has added to those concerns, with the Wall Street Journal reporting that U.S. stock futures fell on Thursday as AI-related anxiety spread from Asian markets, where SK Hynix and Samsung Electronics dropped almost 9%. SpaceX shares slumped 2.97% Thursday, on pace for an eighth drop in nine sessions, as it wiped out $903 billion from a closing high last month, while SK Hynix Inc., which raised $26.5 billion in a record-setting debut of its own last week, dropped 12% bringing it just a few dollars above the $149 level where it sold American depositary receipts to investors.
The pullback across stocks linked to themes such as artificial intelligence infrastructure and aerospace and defense — two of the hottest sectors for newly-public companies — depressed the weighted average return for this year's US initial public offerings to 6%, lagging the S&P 500 Index's 11% return, data compiled by Bloomberg through July 15 show. The majority of US debuts over the past two months are trading below their offer price, with only 10 stocks on the Nasdaq 100 having powered the 20% rally seen by the index. Companies have already raised $157 billion through July 16, excluding blank-check companies and other financial vehicles, with Wall Street's biggest investment banks announcing they hauled in the most revenue from advising on equity offerings in the second quarter since 2021. Michael Ventura, co-head of US equity capital markets at Royal Bank of Canada, said activity in the near-term will "be less busy than we thought," with "outside of the headline names it'll be quieter."
Polymarket traders have raised the odds of an AI bubble bursting in 2026 above 17% after recently falling from 30% to 14%, with contracts using different resolution criteria placing the likelihood between 16% and 24%. According to Bloomberg, Ray Dalio has separately argued that liquidity, rather than weak technology, could break the AI boom, explaining that investors often mistake rising asset values for money they can readily spend. The Bridgewater Associates founder used private companies to illustrate the risk, noting that a business can receive a billion-dollar valuation after raising far less in actual capital, but shareholders cannot use that paper wealth without selling. Bernstein and Cummings have pointed to another pressure building beneath the boom, writing that the AI bubble was "still inflating" while technology investment had reached nearly 5% of U.S. GDP, above levels recorded during the dot-com era. Their analysis also found that large technology companies were committing enough capital to AI projects to reduce their cash reserves.
The violent rotation out of technology stocks that began in June has intensified significantly, with the Nasdaq 100 experiencing 20 of the last 26 trading sessions with price swings of 1% or more in either direction. According to CNBC TV18, the latest 1.1% upmove on Tuesday was followed by Wednesday's recovery from lows to end 0.3% lower. This compares to the earlier July 2 session where the Nasdaq 100 fell 479.92 points (-1.62%) while the S&P 500 Equal Weight Index and Dow Jones Industrial Average made all-time highs. The volatility is particularly pronounced in semiconductor stocks, with the VanEck Semiconductor ETF and iShares Expanded Tech-software ETF seeing at least 30 moves of 4% or more this year, exceeding the combined total from the last seven years. Eddie Molloy, co-head of global equity capital markets at Morgan Stanley, noted that "we had a ton of momentum with the AI theme that led to the deals trading well out of the gates but with markets trading how they are, the steam will come off that a little bit." The Philadelphia Stock Exchange Semiconductor Index has slumped 11% and a momentum basket of stocks down more than 8% over the past month, despite the S&P 500 being virtually unchanged.
Venture capitalist Chamath Palihapitiya has raised significant concerns about the financial returns of AI investments, arguing that the productivity math simply doesn't add up. According to his analysis, the S&P 493 (S&P 500 excluding largest technology companies) has produced only 9% earnings-per-share growth since generative AI entered the mainstream, with only 0% to 2% of that growth stemming from AI-driven productivity. The remainder reflects inflation-driven pricing power and aggressive share buybacks rather than genuine operating improvements. This contrasts sharply with the massive AI infrastructure spending, which has reached $37 billion in 2025 with growth exceeding 3x year-over-year. The June 2026 valuation reset hit the broader AI sector, with Alphabet falling 6%, Amazon declining nearly 5%, while Meta and Microsoft dropped around 3% as investors questioned the near-term profitability of massive AI investments. IBM warned that spending on AI infrastructure was pulling corporate budgets away from software, contributing to weaker-than-expected revenue growth, with the company's AI infrastructure spending pulling budgets away from traditional software development.
