🚨 SK Hynix plunges 14%, South Korean stock market circuit breaker triggered! Tom Lee: AI bubble will eventually come, but definitely not now.

$SK Hynix(SKHY.US)

The AI semiconductor sector is experiencing its most severe adjustment this year. On July 29, SK Hynix plummeted approximately 14% during trading hours, and the South Korean KOSPI index fell over 10%, triggering a circuit breaker. Panic quickly spread across global AI semiconductor and memory sectors.

Just as markets worried that the AI rally was peaking, Tom Lee, co-founder of Fundstrat who has long been bullish on tech stocks, offered a starkly different view in his latest podcast: This plunge resembles forced deleveraging driven by liquidity and leverage issues, rather than a turning point for AI fundamentals.

In recent quarters, leveraged ETF funds surrounding hot targets like SK Hynix expanded rapidly. They amplified stock prices during rallies but triggered forced liquidations when prices dropped, creating a negative feedback loop. Although this adjustment followed earnings releases, the true driver of the widened decline was the concentrated exit of leveraged capital, not the earnings reports themselves.

Tom Lee emphasized that no one can precisely bottom-fish. Waiting for all indicators to turn positive before entering often means buying back in at higher levels. The real test is never finding the lowest point, but whether you can stick to your long-term judgment when the market is most pessimistic—money is made by holding positions, not by frequent trading.

Regarding the heated discussion about the 4.5x expected P/E ratio, he reminded that the storage industry is highly cyclical; P/E ratios are often lowest when industry profits peak because stock prices reflect future expectations. What truly matters is not how low the current P/E is, but whether future earnings forecasts can continue to be revised upward.

He pointed out that the biggest current risk lies in order authenticity: The AI supply chain is long, and concerns over terminal demand price hikes may trigger downstream duplicate ordering and excessive hoarding. If real demand cannot keep up with order growth, the industry may face inventory adjustments—this is also the hardest part to judge in every storage cycle.

On the topic of an AI bubble, Tom Lee does not deny it will eventually appear, but believes we are far from that stage. He cited Cisco as an example: From 1993 to 2000, its stock price accumulated a gain of about 100 times, with at least four drawdowns exceeding 40%. Each time, the market thought the bull run was over, but the true top only appeared when valuations approached 200x P/E and demand forecasts severely detached from reality. He believes the AI industry currently resembles the early-to-mid stages of the internet bull market, with capital expenditures and profitability still expanding. This adjustment is more like normal volatility within a bull market.

Notably, the GRNY ETF managed by Tom Lee is not heavily weighted in semiconductor hardware, but leans more towards large tech companies, AI software firms, and crypto assets. It performed better than most funds heavily invested in semiconductors during this pullback.

He believes there are only two questions the market truly needs to answer: Has global AI capital expenditure decreased? Have real demands for HBM, GPUs, and AI servers slowed down? As long as the answers remain negative, this adjustment is merely deleveraging at the capital level, not a reversal of industrial trends.