Nasdaq's 5.5% Plunge Exposes the AI Trade's Concentrated Core

The Nasdaq Composite fell 5.5% last week in its biggest wipeout since the COVID crash, and the selling revealed a structural problem rather than a passing correction. Nvidia lost 9.3% in the rout. AMD fell 17.3%. The damage did not stay in the United States.

Japan's Nikkei and South Korea's Kospi each dropped roughly 6% the following day. Investors moved into government bonds as concern about AI spending triggered a broader retreat from risk, according to TradingView's market commentary that catalogued the moves.

The numbers matter for the semiconductor supply chain because the selloff showed how deeply AI has embedded itself in global markets. When analysts looked under the hood, the concentration was hard to miss.

How diversified is a diversified portfolio?

Consider a standard allocation. An investor holds an S&P 500 ETF, Nvidia shares, a global technology fund and Asian semiconductor stocks. On paper, that looks diversified across companies, funds and countries.

In practice, all four positions depend on one driver: continuous AI spending. Nvidia, Microsoft and Alphabet carry enormous weight in US indices. Taiwan manufactures the advanced chips. South Korea supplies memory. European and Japanese companies produce essential semiconductor equipment.

When expectations for AI investment change, businesses across that entire chain can suffer simultaneously. The commentary defines this plainly as concentration risk: too much exposure to a single source of losses, even when investments appear spread out. That definition deserves attention from any portfolio that feels safer than it is.

What about the debt behind the boom?

Much of the AI infrastructure expansion runs on borrowed money. CoreWeave, the Nvidia-backed cloud provider, has accumulated approximately $50 billion in debt while expanding its computing capacity.

The company reported a $104.2 billion revenue backlog in the second quarter. That figure deserves scrutiny. It represents contracted future business, not cash available today, and the distinction matters when interest payments are due now.

Higher borrowing costs make future expansion more expensive. If construction slows while interest payments continue, companies may struggle to generate sufficient cash to cover obligations. The commentary sketches a chain reaction: expensive financing delays data centers, delayed data centers reduce equipment orders, and weaker growth expectations make investors reluctant to finance further expansion.

Chipmakers face a different version of the same exposure. Nvidia does not depend on borrowing the way indebted infrastructure operators do. But its revenue growth depends partly on those same customers continuing to buy its products. A customer crisis becomes a supplier problem with a lag.

Why did other markets fall too?

When investors lose confidence in a dominant theme, they cut exposure across whole portfolios. Selling spreads to unrelated stocks as funds raise cash, reduce leverage or rebalance positions.

Government bonds benefit from demand for safer assets. Currencies react to shifting risk appetite. Copper responds to expectations for infrastructure investment. Each market has its own drivers, but AI has grown so large that one sentiment shift now moves multiple asset classes at once.

That is the uncomfortable lesson for traders who assume different asset types provide protection. Last week suggested the assumption fails precisely when it is needed.

What does the selloff actually prove?

Two vulnerabilities, per the TradingView analysis. First, too much capital concentrated in one beloved theme. Second, too much of that theme's expected growth dependent on expensive, debt-financed infrastructure.

CoreWeave illustrates the point sharply. The neocloud is the picks and shovels of the AI boom, yet its stock trades down nearly 80% from its record. Nvidia stock, for its part, is down 25% from its own peak.

AI demand remains strong, and the technology could deliver substantial economic benefits. That does not guarantee every company in the chain will generate attractive returns. Investors must weigh how much future growth is already priced in, whether companies can finance expansion, and how quickly infrastructure spending converts to cash.

The commentary leaves the question open: whether last week marks a long-term reshuffling or a buy-the-dip opportunity. Until debt-financed capacity starts producing cash rather than backlogs, the concentration risk stays on the table.