AI Sell-Off Goes Global: Korea Drops 10.5%, SpaceX Sheds $600B

The AI trade that powered markets to record highs just got its first serious reality check, and it rattled trading floors from Seoul to New York.

Global markets tumbled on June 23 as investors started questioning whether the AI boom that lifted indexes to records has run out of room. The selling hit hardest in South Korea, where the KOSPI fell 10.5% and tripped a 20-minute trading halt, while the tech-heavy Nasdaq dropped 2.2%.

Picture a trading floor in Seoul on Tuesday morning. The KOSPI is sliding so fast the exchange slams on a 20-minute timeout to let everyone breathe. Chipmakers that doubled in value this year are suddenly in free fall. Retail investors who piled in for the easy ride are watching the ride get a lot less easy. This was the day the AI trade got asked a hard question: how high is too high?

What Happened

The damage was global and it was fast. South Korea’s KOSPI led the world down with a 10.5% drop, its two giant memory-chip makers, Samsung Electronics and SK Hynix, both plunging more than 12% in a single session. Japan’s market fell 3.6%. Taiwan and Hong Kong each dropped more than 1%. In Europe, the Stoxx 600 slipped nearly 1%, with semiconductor names like ASML, Infineon, and STMicroelectronics posting sharp declines.

Back in the US, the S&P 500 closed down roughly 1.4% and the Nasdaq fell 2.2%. Even the year’s biggest market story took a hit: SpaceX has shed around 20% of its value in the past week, erasing roughly $600 billion in market value after its post-IPO pop. The stock actually ticked up nearly 1% on Tuesday in choppy trading, landing around $156 a share, but the trend is unmistakable.

The Backstory

To understand the drop, you have to understand the run. South Korea was the world’s best-performing stock market since the start of 2025, and the fuel was simple: AI needs memory chips, and Samsung and SK Hynix make the best of them. Both stocks more than doubled this year. As they soared, retail investors crowded in, which is exactly the kind of dynamic that produces giant, unpredictable swings in both directions.

That same pattern played out everywhere. A small cluster of AI and chip companies have an outsize grip on the major indexes, so when they rally, the whole market looks unstoppable. The flip side is that when they wobble, everyone feels it. For months the market had bigger things to worry about, like oil prices and the war in Iran. This week, the spotlight swung back to one question: is the AI story priced for perfection?

The Core Development

The selling shifted the conversation from “how high can AI go” to “what happens if the AI engine stalls.” A strategist at BNY in London warned that if Tuesday’s action signals AI exhaustion, worries will grow about whether the global economy can generate growth from anywhere else, especially with funding tightening. That is the real anxiety here. Strip out the AI names and the growth narrative gets thin in a hurry.

The timing added a spooky historical rhyme. Alan Greenspan, the former Federal Reserve chairman who coined “irrational exuberance” to describe the dot-com mania, died on Monday, instantly reviving comparisons to the late-1990s boom and bust. Goldman Sachs analysts offered a more measured read: unlike the 1990s, strong profit growth has mostly kept the worst imbalances from building. But they flagged a catch. Outside of AI, the broader economy looks weaker than it did back then, which makes markets unusually fragile to anything that dents the optimistic AI story.

The Business Model Angle

Here is the lesson for operators, and it has nothing to do with timing the market. This is a case study in concentration risk at the scale of an entire economy. When a handful of companies tell one story, and that story drives the index, your portfolio (and a lot of the growth narrative) is really a single bet wearing a diversification costume.

CLSA’s chief equity strategist, Alexander Redman, captured the mood by calling the market “very, very frothy,” noting that a 12% one-day drop used to cause panic and now barely registers as unusual. That normalization of wild swings is the tell. The strategic takeaway for founders is the same one that applies to a startup’s revenue: if 80% of your growth comes from one customer, one product, or one narrative, you do not have a strong business. You have a fragile one that happens to be working. Diversification is boring right up until the day it is the only thing that saves you.

The Risk

Now the honest counterpoint, because forced doom is as lazy as forced optimism. One bad day is not a crash. Samsung and SK Hynix more than doubled this year, so a 12% drop is a dent, not a collapse. Redman himself admitted he could not say whether Korean shares would bounce back fast or whether this was “the beginning of the end.” Nobody ringing the alarm actually knows which it is yet.

Goldman’s point cuts both ways too. Profits have been real, not just vibes, which is a meaningful difference from the dot-com era. The risk is not that AI is fake. The risk is that the price assumes everything goes right, and markets that price in perfection have a long history of punishing the first disappointment hard. The smart move is neither panic nor denial. It is watching the fundamentals underneath the narrative.

Quick Questions

The Bottom Line

A market that treats a 10.5% national index drop as a normal Tuesday is telling you something. The lesson for founders and operators is not to fear AI or chase it. It is to notice when an entire system has stacked its growth on one story. Concentration feels like momentum on the way up and like freefall on the way down. Build your business, and read your market, with the diversification that froth makes you forget you need.

Read the original story at The New York Times. For the deeper valuation debate driving all this, see our breakdown of why skeptics are worried about SpaceX’s $1.77 trillion price tag.