The White House reaffirmed that the ceasefire agreement with Iran remains stable, temporarily easing geopolitical risks. Brent crude oil fell nearly 4% intraday and slipped below the $110 mark, removing the last remaining macro concern for this AI-driven frenzy.

The narrative chain has become increasingly “solid”: AI computing demand → semiconductor valuation expansion → passive ETF inflows → further gains in heavyweight semiconductor stocks.

Intel closed up 12.92%, pushing its market capitalization above $500 billion. Micron Technology rose 11.06%, lifting its market value above $700 billion. SanDisk surged nearly 12%, exceeding a $200 billion valuation and entering the world’s top 100 companies by market capitalization. The Philadelphia Semiconductor Index jumped 4.23% on the day.

But a similar curve once appeared on the Nasdaq chart in late 1999.

The “Narrowest” Rally in 25 Years

What truly makes the current market environment resemble 2000 is not just how much the indices have risen, but howthey have risen.

Goldman Sachs Chief U.S. Equity Strategist Ben Snider stated: “The index looks strong, but the median S&P 500 stock is still about 13% below its previous high.” This gap between the median stock and the index itself is one of the widest divergences seen in roughly 25 years.

Another detail worth noting: Nvidia actually fell 1% on the day. The biggest “AI icon” did not rise.

Capital instead flowed into second-tier memory chip and foundry-related names such as Micron, Intel, and SanDisk. This suggests the market has shifted from “buying high-certainty leaders” to a mentality where “anything related to AI deserves a share of the rally,” which in some ways resembles the sentiment of 2000. Back then, the market believed all internet infrastructure companies were valuable — not just Cisco and Sun Microsystems. Any company capable of telling a bandwidth story could attract speculative enthusiasm.

Combining these two observations — weak breadth among median stocks and capital rotating from market leaders into second- and third-tier names — the market is beginning to display a classic sign of late-stage bull market fatigue: fewer and fewer stocks are driving the indices higher, and even those stocks are benefiting mainly from rotational flows rather than durable fundamental consensus.

The Real Market Situation Looks More Concerning

A survey by Ernst & Young provides an even harsher reality check: 99% of surveyed companies reported AI-related financial losses, with each company conservatively estimating average losses of $4.4 million.

In other words, while consumer enthusiasm around AI remains extremely strong, almost no enterprise has yet managed to truly profit from AI investments.

So the current market situation looks like this: on one side, the Philadelphia Semiconductor Index is rallying at the same speed seen at the peak of the tech bubble 24 years ago, with gains highly concentrated in a small number of stocks; on the other side, financial returns from enterprise AI investment remain alarmingly poor.

Only the “shovel sellers” are thriving, while the “gold miners” are still losing money.

If revenue growth fails to keep pace, these companies may have almost no room left for adjustment.

That assumption may ultimately prove correct.

But just as it did 24 years ago, the path toward being “right” may first involve enormous volatility and painful corrections along the way.