The AI Trade Is Entering Its Next Phase: Prove It

Global markets sold off on Tuesday, but this wasn’t a typical risk-off day driven by geopolitics or fears of an economic slowdown. Instead, the pressure was concentrated in one area: AI.

South Korea’s Kospi fell nearly 10%, a drop steep enough to trigger a circuit breaker and halt trading on the exchange. AI memory-chip giants SK Hynix and Samsung Electronics both fell more than 12%. Japan’s Nikkei 225 lost 3.6%, and when US markets opened, the Nasdaq was down roughly 2%, led lower by Micron, which fell around 9% ahead of its earnings report.

The scale of the move grabbed headlines. The reason behind it matters far more.

What sold off

The selling was heavily concentrated in AI-linked semiconductor stocks.

In Asia, investors rushed to sell memory-chip names that have been among the biggest beneficiaries of the AI boom. SK Hynix and Samsung Electronics led losses in South Korea, while Japanese and Hong Kong markets also finished lower.

Europe held up better, although semiconductor stocks still came under pressure. Germany’s Infineon fell more than 5%, helping drag the DAX lower.

The weakness then spread to the US, with Micron, Western Digital and Qualcomm all opening sharply lower.

What is interesting, however, is what did not sell off.

The Russell 2000, which tracks smaller US companies, actually rose. Defensive names and software companies such as IBM and Accenture also gained. That suggests investors weren’t abandoning equities altogether. They were rotating out of one of the market’s most crowded trades.

What is driving the sell-off?

At its core, this move is about a simple question: Have AI valuations run too far ahead of reality?

For the last two years, investors have rewarded almost any company connected to the AI ecosystem. The thesis was straightforward: AI demand is exploding, hyperscalers are spending aggressively on infrastructure, and suppliers throughout the supply chain will benefit.

Now the market is beginning to ask for evidence. The AI trade is entering its next phase: prove it. Memory chips became the market’s favourite trade The immediate trigger appears to be the sell-off in memory-chip stocks. Companies such as SK Hynix and Samsung have enjoyed extraordinary gains thanks to soaring demand for AI-related memory products. Advanced AI models require huge amounts of high-bandwidth memory, making these companies some of the biggest winners from the AI boom.

But when a trade becomes crowded, expectations become extremely difficult to satisfy. After months of strong performance, investors appear to be taking profits and reassessing how much future growth is already reflected in share prices. This doesn’t necessarily mean the long-term story is broken. It simply means investors are becoming less willing to pay ever-higher multiples without fresh evidence that earnings can continue to justify them.

The second driver is Micron’s earnings report. Micron sits at the centre of the AI infrastructure story because it produces high-bandwidth memory (HBM), a critical component used in AI servers. As a result, investors see Micron as one of the clearest indicators of whether AI-related demand remains as strong as markets currently expect. Those results are due after the US market close today.

The challenge is that expectations are already incredibly high. When valuations become stretched, companies don’t just need to report good numbers. They need to exceed expectations and provide guidance strong enough to justify future growth assumptions. Micron’s results could therefore influence sentiment across the entire AI ecosystem, not just its own stock.

There is also a broader shift happening beneath the surface. Earlier in the AI rally, investors were happy to buy almost anything with an AI narrative attached to it. Today, the market is becoming more selective. Investors are increasingly separating companies with clear earnings visibility from those whose valuations rely primarily on optimism and momentum. This is a healthy development.

Bull markets eventually move from excitement to execution. The market appears to be entering that stage now.

The geopolitical backdrop

One of the more interesting aspects of Tuesday’s sell-off is that geopolitics doesn’t appear to be the main culprit.

Over the weekend, investors worried that tensions involving Iran could escalate and potentially disrupt shipping through the Strait of Hormuz, one of the world’s most important oil routes. A closure of the strait would likely have pushed oil prices significantly higher, creating renewed concerns about inflation and interest rates. Instead, the news flow moved in the opposite direction. President Trump said Iran had agreed to nuclear inspections and indicated that the Strait of Hormuz would remain open. Oil prices continued to fall, while gold also moved lower. Both are signs that markets were becoming less concerned about an immediate geopolitical shock.

That distinction matters. If geopolitics were driving the sell-off, we would likely have seen broad-based weakness across sectors and a stronger move into traditional safe-haven assets. Instead, the heaviest selling was concentrated in AI and semiconductor stocks. That points much more towards a valuation and positioning reset than a geopolitical event.

Rates are the other pressure

Geopolitics aside, the rate backdrop is doing some of the work too.

Inflation has proven sticky, and the market is increasingly worried that the Federal Reserve may need to raise rates again this year rather than cut. When rates are expected to stay higher for longer, the companies that suffer most are the ones whose valuations depend on profits far out in the future, which describes much of the AI complex. Thursday’s US PCE inflation print, the Fed’s preferred gauge, is the next thing to watch on that front.

My thoughts

Back in May, I wrote that markets had stopped rewarding AI capex announcements and had started demanding proof that all this spending would eventually generate returns (Link to the article is here). Tuesday’s sell-off feels like the next stage of that process.

The companies that benefited most from the AI narrative are now facing a higher bar. Investors still believe in AI, but belief alone is no longer enough to support every valuation.

Importantly, nothing that happened on Tuesday suggests the underlying AI investment cycle has suddenly changed. Hyperscalers are still increasing capital expenditure. Demand for AI infrastructure remains strong. The long-term growth story remains intact. What appears to be changing is investor expectations.

For now, this looks like a valuation and positioning reset rather than a deterioration in fundamentals.

The early reaction supports that narrative. As I write, the Asian markets have steadied today, with the Kospi recovering a good part of Tuesday’s fall and Samsung clawing back much of its drop. That is the kind of snap-back you tend to see after a positioning shake-out, not a genuine change in the fundamentals.

The risk worth watching is concentration. When a handful of stocks account for a large portion of index gains, any reassessment in one corner of the market can spread remarkably quickly through the rest.

That’s exactly what we saw on Tuesday.

And with Micron reporting tonight, the next test arrives quickly.

Sources: Yahoo Finance, CNBC, CNN, Reuters, Bloomberg, NBC News, The Economic Times and Investor’s Business Daily (23–24 June 2026 market coverage).

Disclaimer: The views expressed in this article are my own and do not represent those of my employer or any affiliated organisation. This article is for informational and educational purposes only and should not be considered financial, investment or trading advice. Investors should conduct their own research and consider their individual financial circumstances before making any investment decisions.