June 23, 2026 (Eastern Time), the US stock market experienced a sharp sell-off centered on AI-related stocks. The Nasdaq closed down 2.21%, while the Nasdaq 100 plunged 3.3%. The Philadelphia Semiconductor Index dropped 7.87% in a single day, marking its largest one-day decline since June 5. This sell-off was not an isolated event; it was the result of multiple structural pressures converging after three years of AI industry valuation expansion.
How Broad and Deep Was the Latest AI Stock Sell-Off?
Chip stocks bore the brunt of this sell-off. SanDisk (SNDK) plummeted over 13%, falling below the $2,000 mark. Micron Technology (MU) dropped 13.18%. ARM fell more than 10%. Qualcomm and Western Digital both dropped over 8%. TSMC and Intel fell more than 6%. AMD declined over 5%.
NVIDIA (NVDA), the benchmark for AI computing power, closed down 4.15%, with its market cap falling below $5 trillion. Tesla tumbled 5.79%. Among major tech stocks, only Microsoft bucked the trend, rising 1.8%.
The impact extended far beyond US stocks. During Asian trading hours on June 23, Korea’s KOSPI Index triggered a circuit breaker after plunging 8%, ultimately closing down 9.99%. Samsung Electronics dropped 12.31%, and SK Hynix fell 12.47%. The A-share AI computing power sector collectively adjusted, and Hong Kong tech stocks also declined. From Asia-Pacific to North America, global capital is visibly retreating from high-beta tech assets.
What Changed in SpaceX’s Valuation Logic as Over $600 Billion Was Wiped Out in Three Days?
SpaceX stands out as the most symbolic case in this sell-off. The aerospace and AI company completed a record-setting IPO on June 12, and within just two weeks, it went from frenzy to panic.
In its first week after listing, SpaceX’s stock hit a short-term high of $225. Over the next three trading days, shares fell consecutively, with a cumulative decline of about 23% and over $600 billion in market value erased. On June 22 alone, the stock plunged 16.4%, wiping out roughly $400.8 billion in market cap—the second-largest single-day loss in US corporate history. On June 23, SpaceX’s stock price fell below $150, matching its opening price on IPO day.
One direct trigger for the crash was the company’s announcement of a $20 billion bond issuance plan. SpaceX aimed to raise capital for AI infrastructure expansion, but the market’s reaction was negative. For a company that has yet to turn a profit and is expected to remain cash-negative until 2029, large-scale debt financing only intensified concerns about liquidity and repayment ability.
Supply-side risks are even more critical. Currently, only about 5% of SpaceX shares are in circulation, with the remaining 95% locked up. As the concentrated lock-up expiration window arrives in August and September, insiders could potentially sell up to 44% of total shares, expanding the float nearly ninefold. With demand momentum clearly cooling, the looming supply shock is fundamentally reshaping the stock’s valuation structure.
Does NVIDIA’s Market Cap Falling Below $5 Trillion Signal a Turning Point for AI Hardware?
NVIDIA’s decline is equally indicative. As the biggest beneficiary of the "AI computing power scarcity" narrative over the past three years, NVIDIA’s market cap briefly returned to $5 trillion in April 2026. However, the stock has faced persistent pressure since June.
This is not just a routine stock correction. Goldman Sachs strategists noted in a June 23 report that the market is witnessing a widening structural divergence: hyperscale cloud providers continue to ramp up capital expenditure commitments, yet their stock prices lag behind the broader market; meanwhile, AI hardware stocks like NVIDIA and TSMC previously defied the trend and rose. This divergence itself is a sign of market mispricing.
Morgan Stanley portfolio manager Andrew Slimmon commented, "This sell-off is mainly concentrated in stocks benefiting from the AI theme. I don’t think these companies are overvalued, but the trades have become overcrowded. AI has become the era’s momentum play. When an investment theme gets too crowded, sharp corrections like this are inevitable."
How Large Has AI Infrastructure Debt Financing Become?
To understand the deeper logic behind this sell-off, we must examine the financing structure underpinning AI infrastructure development.
According to Morgan Stanley’s "AI Debt Financing Tracker," by the end of May 2026, global AI-related bond issuance had reached $236 billion, up 357% from the same period in 2025. Morgan Stanley projects total AI debt issuance for the year will surpass $570 billion. In April alone, issuance exceeded $74 billion, with project finance for data center construction accounting for 85% of high-yield bond supply.
America’s four tech giants—Google, Amazon, Meta, and Microsoft—are expected to spend about $650 billion in capital expenditures in 2026. Combined capital spending by Amazon, Microsoft, Alphabet, and Meta will reach $725 billion. Global AI capital expenditures may approach $800 billion in 2026.
What’s more concerning is the off-balance-sheet debt. Morgan Stanley estimates about $982 billion in long-term procurement commitments, over $800 billion in unactivated leasing contracts, and tens of billions in supplier financing arrangements, together amounting to roughly $1.8 trillion in off-balance-sheet exposure. These liabilities don’t appear on balance sheets, but they lock in future cash outflows.
The overall gross leverage ratio for hyperscale cloud companies has surged from 0.9x in Q3 2025 to 1.8x currently. Morgan Stanley predicts Amazon and Meta’s free cash flow will approach zero or turn negative in 2026, making incremental financing almost entirely reliant on new debt.
Why Are Crypto Assets Under Pressure During the AI Sell-Off?
The plunge in AI stocks is not isolated within the tech sector—crypto assets are feeling the strain as well.
As of June 24, 2026, according to Gate market data, Bitcoin (BTC) was priced at $62,595, down 2.1% in 24 hours; Ethereum (ETH) was at $1,662, down 3.7%. Leveraged long positions faced large-scale liquidations.
Bitcoin’s correlation with the Nasdaq remains around 0.45, higher than its 10-year average. This means that during systemic risk events, Bitcoin still struggles to decouple from tech stocks. The market widely attributes Bitcoin’s pullback to cooling risk appetite—after AI stocks gave up gains, investors became more cautious toward high-volatility assets.
A notable structural shift: some Bitcoin mining companies are pivoting to AI data center hosting, holding long-term power purchase agreements. Investors are now evaluating these firms using AI infrastructure valuation logic, focusing on installed power capacity, data center assets, and customer contracts. This means the crypto mining sector is being incorporated into the AI infrastructure narrative—when AI hardware stocks are sold off, this transmission chain also impacts crypto assets.
From Computing Power Rental Prices to Tightening Budgets: How Is the AI Industry Chain Being Stress Tested?
Over the past three years, the AI industry has followed a straightforward logic: the scarcer the computing power, the more justified the capital expenditure; the larger the capital expenditure, the higher the valuation; the higher the valuation, the easier the financing. But this self-reinforcing cycle is being disrupted by multiple forces.
Upstream, the spot prices and forward contract prices in the computing power rental market have diverged unusually. Midstream, tech giants that once spent lavishly are now tightening AI budgets. At a deeper level, physical constraints like power supply and engineering delivery are becoming tougher bottlenecks than chip manufacturing.
Goldman Sachs strategists warn that the AI market is like a stretched rubber band—the market’s persistent disregard for negative signals will eventually reach a breaking point. If any major tech giant cuts AI spending first, the entire sector’s valuation logic will be fundamentally reshaped.
TS Lombard analyst Dario Perkins points out that capital expenditure of this scale is extremely rare in tech history. Historically, any tech bubble bursts not because the technology fails, but because the money runs out—or the funders lose patience.
From AI Stocks to Crypto Assets: How Does the Risk Transmission Chain Work?
This sell-off reveals a clear risk transmission pathway:
First layer: AI hardware stocks (NVIDIA, chip makers) face direct valuation pressure. Second layer: AI+ narrative stocks (SpaceX) suffer more severe corrections due to high leverage and low float structure. Third layer: global risk assets (including crypto assets) come under pressure as overall risk appetite declines.
The deeper logic: roughly $750 billion in annual AI capital expenditures, combined with about $1.8 trillion in off-balance-sheet exposure and over $570 billion in annual debt issuance, create a fragile structure highly dependent on ongoing financing. When the market begins to question whether AI commercialization can support such massive capital consumption, the vulnerability of the entire financing chain becomes apparent.
For the crypto market, this means Bitcoin’s "digital gold" safe-haven narrative has not fully materialized during systemic risk events. In an environment of tightening liquidity and falling risk appetite, crypto assets tend to move in tandem with high-beta tech stocks, rather than serving as independent hedges.
Conclusion
Between June 23 and 24, 2026, AI stocks experienced their most intense collective sell-off since the latest AI boom began. SpaceX lost over $600 billion in market value in three days, NVIDIA fell 4.15%, SanDisk and Micron dropped more than 13%—behind these numbers lies a collective market repricing of the sustainability of AI infrastructure capital expenditures.
About $750 billion in annual capital spending, over $570 billion in AI-related debt issuance, and roughly $1.8 trillion in off-balance-sheet exposure together outline an industry ecosystem highly reliant on continuous financing. As computing power rental prices fall, tech giants tighten budgets, and leverage ratios climb, the ecosystem’s fragility is coming to light.
For crypto assets, Bitcoin’s correlation with the Nasdaq at around 0.45 means it still struggles to escape the shadow of tech stock volatility. Whether the AI stock sell-off is over or just the beginning of a larger adjustment depends on whether AI commercialization can deliver enough returns before capital consumption hits a critical threshold. The market is shifting from the "computing power scarcity" narrative to a new phase focused on "ROI validation."
Frequently Asked Questions (FAQ)
Q1: What is the core reason behind the latest AI stock sell-off?
The main reason is a concentrated outbreak of concerns about the sustainability of massive AI infrastructure capital expenditures. About $750 billion in annual capital spending, over $570 billion in AI-related debt issuance, and roughly $1.8 trillion in off-balance-sheet exposure have led the market to question whether AI commercialization can support such large-scale capital consumption.
Q2: Why did SpaceX’s decline become so severe?
After its IPO, SpaceX saw a rapid short-term surge, then faced triple pressures: liquidity concerns triggered by a $20 billion bond issuance plan, anticipated supply shock as 95% locked shares approach mass unlocking, and the reality that the company is not yet profitable and is expected to remain loss-making through 2029.
Q3: Why are crypto assets affected by the AI stock plunge?
Bitcoin’s correlation with the Nasdaq 100 Index remains around 0.45, higher than its 10-year average. During systemic risk events, crypto assets still struggle to decouple from tech stocks. Additionally, some Bitcoin mining companies are shifting toward AI data center operations, bringing the crypto mining sector into the AI infrastructure valuation framework.
Q4: How large is AI capital expenditure?
America’s four major tech companies (Google, Amazon, Meta, Microsoft) are expected to spend about $650–$725 billion in capital expenditures in 2026. Global AI capital spending may approach $800 billion. As of the end of May 2026, global AI-related bond issuance had reached $236 billion, with the full-year figure expected to surpass $570 billion.
Q5: Does the decline in AI stocks mean the AI bubble is bursting?
It’s too early to say the AI bubble has burst, but the market is transitioning from the "computing power scarcity" narrative to the "ROI validation phase." Goldman Sachs strategists note that if any major tech giant cuts AI spending first, the entire sector’s valuation logic will be fundamentally reshaped. The current correction should be seen as a response to overcrowded trades, not a rejection of AI’s long-term trajectory.
