According to AP, on June 23, U.S. tech stocks and the AI supply chain collectively declined, with the Nasdaq index closing down 2.2% and the S&P 500 down 1.4%. This pullback was not due to issues at a single chip company, but rather a simultaneous pressure on the most crowded AI hardware trades over the past year: one being a sudden surge in expectations for Federal Reserve rate hikes, and the other being investors beginning to question when the cloud providers’ continued heavy capital expenditures on AI will translate into clearly defined profits.

The most direct pressure fell on the hardware chain. Market data showed that NVIDIA (NVDA) dropped about 4% on Tuesday, with its market capitalization falling below $5 trillion. Micron plunged 13.2%, Qualcomm declined about 8%, and SanDisk and Western Digital also suffered significant losses. The simultaneous weakness in memory, storage, AI chips, and smartphone chips indicates that the selling pressure is not confined to a single segment.

Asian markets also came under pressure. On June 23, South Korea’s KOSPI index fell nearly 10%, with both SK Hynix and Samsung Electronics posting double-digit declines. For months, tight supply of HBM and memory chips had supported South Korean tech stocks, but this time, the market chose to book profits first.

The AI hardware chain was sold first.

The order of this downturn is highly significant. Investors did not first pull back from software or internet platforms; instead, they initially targeted chips and memory stocks, which had previously benefited the most from AI-related capital spending.

NVIDIA remains the core asset in the AI boom. Its GPUs have nearly defined this cycle of data center expansion and have become the primary outlet for market risk appetite. A market cap falling below $5 trillion does not alter the company’s industrial standing, but it serves as a prominent price signal on the trading front. When both interest rates and return cycles are questioned, assets with the largest gains and most crowded positions are often sold first.

Micron’s decline was steeper, partly due to its upcoming earnings report. The company announced it will release its third-quarter fiscal year 2026 results and hold an earnings call on June 24. The market had already bet on sustained strong demand for high-bandwidth memory driven by AI servers. If guidance is weak, investors fear the prior price gains lack new earnings catalysts; even if guidance is strong, the company must demonstrate that high memory prices and AI demand are not merely short-term buying surges.

The market reaction in Korea further amplified these concerns. SK Hynix and Samsung, both key players in the global memory and HBM supply chain, posted double-digit declines, indicating that this correction has spread from U.S. tech leaders to the global AI hardware supply chain.

Previously, Broadcom's AI revenue guidance falling short of the most optimistic expectations triggered a sell-off in chip stocks. Tuesday's market movement resembled a concentrated release of these concerns. Demand for AI remains strong, but the market is no longer willing to pay ever-higher prices solely for the promise of "huge potential in the future."

Expectations for rate hikes have turned hawkish, increasing pressure on overvalued tech stocks.

The trigger at the macro level comes from changes in expectations regarding Federal Reserve policy.

According to a Federal Reserve announcement, Kevin Warsh was sworn in as Chair of the Federal Reserve on May 22. Citing a Bank of America forecast, Reuters reported that the Fed may raise interest rates by 25 basis points in September, October, and December 2026, for a total of 75 basis points for the year, citing resilient labor markets and persistent inflationary pressures.

This is particularly unfavorable for technology stocks. The valuations of AI leaders rely heavily on long-term growth expectations; rising interest rates increase the discounting pressure on future cash flows and make low-risk assets like U.S. Treasuries more attractive again. Recently, U.S. Treasury yields have remained elevated, and futures markets have clearly intensified bets on rate hikes this year, causing market expectations for policy path to adjust rapidly.

The market isn't suddenly doubting the existence of AI; it's recalculating a more realistic question: If the cost of capital is higher and future profits are further away, how much are we willing to pay for AI assets today?

This is also why adjustments in chips, memory, and high-growth tech stocks have been so synchronized. They previously benefited together from the combination of “sustained explosive demand for AI” and “eventual interest rate declines.” Once one of these pillars weakens, the segments with the largest gains and highest valuations come under pressure first.

Cloud providers are still spending money, while investors are beginning to ask about returns.

Another source of pressure comes from AI capital expenditures themselves.

Major hyperscale cloud and AI investors such as Alphabet, Amazon, and Meta continue to maintain intensive data center construction. Over the past year, such spending has been viewed by the market as a guarantee of demand for NVIDIA, memory chips, power equipment, and data center assets. As long as cloud providers continue to invest heavily, the hardware supply chain will sustain its revenue.

But now the question is, can this money ultimately be recovered?

Training and inference for AI models require massive computational power, electricity, and server investments. Cloud providers can monetize through enterprise customers, advertising tools, developer platforms, and consumer subscriptions, but it remains unproven whether service pricing can fully cover capital expenditures. The market is now scrutinizing AI product pricing, customer usage intensity, and whether enterprises are willing to pay high long-term costs for generative AI.

This is also why the "sell the heavy spenders" trade has gained popularity. Investors are not only selling chip stocks but also becoming more cautious toward internet and cloud computing giants that continue to increase their AI budgets. The more aggressive their prior spending, the more scrutiny they face regarding profit margins and free cash flow.

Volatility in overvalued assets is amplifying this sentiment. According to Axios, SpaceX's stock fell more than 16% on Monday after its IPO, wiping out approximately $400 billion in market value. While not the primary driver of the recent chip stock decline, it illustrates that highly hyped, overvalued assets are now facing stricter market scrutiny.

It’s too early to say the bubble has burst—Micron and inflation data will provide the answers.

This correction is more accurately described as a concentrated pullback following significant gains in AI trading, rather than a confirmed bubble burst.

Demand for AI hardware remains strong, and cloud providers have not halted their capital expenditures. The fundamentals of companies like NVIDIA, Micron, and SK Hynix are still closely tied to data center construction, HBM supply, and AI server shipments. The real question is whether current stock prices have already priced in too many positive developments.

The first key checkpoint is Micron’s earnings report. The market will focus on three things: whether demand for memory driven by AI servers remains strong, whether price increases can be sustained, and whether management’s guidance for upcoming quarters is sufficient to justify prior gains. If the earnings are strong, the chip sector may get a reprieve; if guidance falls short of expectations, selling pressure could spread further across more AI supply chain companies.

The second checkpoint is interest rates. Whether the Fed under Walsh truly raises rates starting in September will depend on inflation, employment, and energy prices. If inflationary pressures remain stubborn, growth stock valuations will continue to face pressure; if the data shows cooling, the market may reprice expectations toward a policy pivot, providing room for tech stocks to recover.

The current market divergence lies in whether this is merely a normal profit-taking within the AI bull market, or the beginning of a shift by investors from “buying growth at all costs” to “demanding tangible returns.” Tuesday’s decline at least indicates that the AI narrative remains strong, but it can no longer alone offset the pressure from higher interest rates and a longer path to profitability.