Alphabet fell 6% on Monday — its worst day in over a year. The Nasdaq sank 2.1% on Tuesday. SpaceX wiped out its post-IPO gains before recovering. Semiconductor stocks took the worst beating. The question driving all of it: hyperscalers are spending $452 billion on AI. When does the revenue catch up?
The AI stock market sell-off is the second significant correction in four months — and it is driven by the same question that sparked the February episode. The Nasdaq Composite sank roughly 2.1%, leading Wall Street down. The S&P 500 lost 1.4%, while the Dow Jones Industrial Average fell just below the flat line. The sell-off began a day earlier. The sell-off began on 22 June when Alphabet stock plummeted 6%, marking its worst day in over a year, driven by investor anxiety over the company’s massive AI infrastructure buildout and talent departures to rivals. The concerns are specific and grounded. “Some of the news lately about AI raises questions about all the spending that’s being done and the capex and ramping of the capacity for semiconductors,” said Thomas Martin, senior portfolio manager at Globalt. The market is not rejecting AI. It is demanding evidence that AI generates returns proportionate to its costs.
What’s Happening & Why It Matters
The Capex Numbers That Frightened Investors
The AI stock market sell-off of 23 June 2026 centres on a specific and extraordinary set of capital expenditure figures. Alphabet guided 2026 capital expenditures to a range of $175 billion to $190 billion, while Amazon flagged about $200 billion in capex across the company, bringing combined hyperscaler capex to over $452 billion for the year. To put that in context, the combined GDP of Norway and the Netherlands is approximately $1.1 trillion. The two hyperscalers are spending nearly half that figure on AI infrastructure in a single calendar year.
By contrast, the cash flow picture tells a darker story. Alphabet’s Q1 2026 free cash flow fell 47% year-over-year to $10.12 billion, while Amazon’s trailing free cash flow collapsed 95% to $1.2 billion. Additionally, concerns over hyperscalers’ debt-funded AI spending have contributed to the selloff. Elon Musk’s SpaceX, which debuted this month, has joined a list of megacaps tapping the bond market to raise capital. Companies spending $200 billion on infrastructure while their free cash flow contracts by 95% are making a very specific bet that the revenue arrives before the debt is a problem.
Semiconductor Stocks: The Worst Single Day of the Sell-Off
The sector that took the largest is on 23 June was semiconductors — the physical infrastructure layer of the entire AI bet. Korean markets were the worst after stocks of Samsung and its competitor SK Hynix fell 12% apiece. Intel and Advanced Micro Devices were both off around 6%. Micron took the biggest beating, mostly over nervousness ahead of the company’s results expected on Wednesday.
The semiconductor sell-off is a specific concern. Chipmakers have been scaling production capacity in anticipation of sustained AI demand. Micron’s stock has skyrocketed in value in the past year — up close to 800% — on soaring demand for memory chips from the AI build-out. As Tim Cook told the Wall Street Journal, AI data centres have quadrupled memory chip costs. As TF covered in its Apple price increases article, that memory cost surge is already hitting consumer product pricing. Any slowdown in hyperscaler capex directly threatens chipmakers whose production lines are calibrated to the current pace of AI investment.
SpaceX: From Post-IPO High to Three-Day Losing Streak
The most dramatic individual stock story in the sell-off belongs to SpaceX. Earlier, investors piled into SpaceX stock after the company’s initial public offering, sending it above $200 within days. But the stock has retreated since June 17 as investors fret over whether the company can justify its valuation exceeding $2 trillion. On Monday, SpaceX shares plunged 16%. As TF covered in its SpaceX Cursor acquisition article, that $211 peak produced the stock currency to pay for a $60 billion acquisition. The subsequent 16% single-day fall demonstrates how quickly post-IPO sentiment can reverse.
By contrast, SPCX recovered partially on Tuesday. SpaceX briefly wiped out all of its post-IPO gains but then turned higher in afternoon trading, ending the day slightly in the green. The intraday recovery is the specific dynamics of a new public company — large short positions, index-inclusion mechanical buying, and retail investor buying-the-dip behaviour all creating volatility in both directions.
The Revenue Gap — Only 3% of Bank of America Customers Pay for AI
The fundamental concern beneath the AI stock market sell-off is commercial, not technical. Although apps like OpenAI’s ChatGPT and Anthropic’s Claude have made a splash among consumers, the vast majority of those users employ free versions and other AI tools. New data from the Bank of America Institute show that only about 3% of its customers — mostly households with more than $125,000 in annual income — pay for AI services.
That 3% figure is the data point that makes the $452 billion capex most exposed. Additionally, the Shiller price-to-earnings ratio for the US market exceeded 40 for the first time since the dot-com crash, signalling extreme overvaluation according to some analysts. By contrast, bulls argue that AI monetisation is at the same stage internet monetisation was in 1998 — a delay before explosive commercial adoption, not an indication that the commercial opportunity is absent. “The market is trying to kind of digest all this and saying, ‘Are we going to start to see returns?'” said Mark Vena, CEO of SmartTech Research.
TF Summary: What’s Next
Micron reports its quarterly results on Wednesday — the most closely watched single earnings event in the sell-off context. A strong Micron result demonstrating sustained AI memory demand could partially reverse the semiconductor decline. The Federal Reserve‘s June projections indicated a possible rate increase before year-end, adding a macro headwind to the AI spending concern. OpenAI and Anthropic are both preparing IPOs in the environment — the sell-off directly affects both companies’ public market ambitions.
MY FORECAST: The AI stock market sell-off of 23 June 2026 is a correction, not a crash — and Micron’s Wednesday earnings will determine whether it extends or reverses. The underlying concern is legitimate and will not be resolved by one quarterly report. Hyperscalers are spending $452 billion on AI infrastructure in 2026. The revenue that justifies that spending must arrive in 2027 and 2028 to validate current valuations. By contrast, the 3% paying-customer figure is the tell. When that proportion reaches 10%, the commercial model is proven. When it reaches 20%, the current valuation levels are defensible. The market will oscillate between optimism and anxiety until the paying-customer curve bends sharply upward — or definitively does not. That answer does not arrive yet.
Related Stories
- Apple iPhone Price Increases — Memory Crunch and Intel’s US Chips Are Why
- Alphabet Raises $80 Billion for AI
- SpaceX Acquires Cursor’s Parent Anysphere for $60 Billion