From Seoul to Silicon Valley: Chip Stocks Tumble Again as AI Boom Faces Leverage Scrutiny

How large was the chip stock sell-off?

The sell-off in semiconductor shares was broad and sharp. According to United Daily News (money.udn.com), U.S. equities extended their decline on July 17, 2026, as a chip-stock sell-off intensified selling pressure, dragging the tech-heavy Nasdaq down 1.4% and the S&P 500 down 1% (E1). The Central News Agency (CNA) reported the identical index moves, confirming the scale of the broader market drag (E12 references the same Philadelphia Semiconductor Index figures).

The damage was concentrated in chip names. Per money.udn.com, the Philadelphia Semiconductor Index fell 1.6% on the day, bringing its weekly loss to roughly 10% — the steepest one-week decline in more than a year. The index closed more than 20% below the record high it set at the end of June, a threshold that confirms the index has entered a bear market (E3). CNA's report carried the same set of figures verbatim, underscoring how the story was covered in parallel across both money.udn.com and cna.com.tw (E12).

Does the sell-off reflect real fundamental deterioration?

Despite the bear-market confirmation, market watchers pointed to positioning rather than a fundamental collapse. Chuck Carlson, chief executive of Horizon Investment Services in Indiana, told reporters, "I don't think this has much to do with fundamentals. It's mainly just portfolio rebalancing, along with profit-taking in some stocks that had risen too far" (E6).

That view is bolstered by the index's longer-run performance: according to money.udn.com, the Philadelphia Semiconductor Index is still up more than 60% year-to-date even after the week's rout (E5) — a gain that dwarfs the roughly 10% weekly and 20% peak-to-trough declines cited above.

How did new AI model releases deepen capex concerns?

A specific catalyst named across both outlets was a new AI model from China. Moonshot AI, a Chinese startup, unveiled its latest model, Kimi K3, which it claims has caught up to some of the most advanced U.S. AI models. Both money.udn.com (E2) and CNA (E13) reported that the release deepened investor concern over whether massive AI capital expenditure is justified, and that it contributed to declines in shares of Intel, Applied Materials, and AMD.

Adding to the unease, Bloomberg News reported — as cited by money.udn.com — that Alphabet's Google is running months behind schedule on the release of its flagship AI model, Gemini 3.5 Pro (E7). The delay, paired with Kimi K3's emergence, fed a narrative that AI development timelines and payoffs are less certain than markets had priced in.

Is leverage buildup a hidden risk behind the drop?

Beyond fundamentals and product timelines, sources pointed to leverage as an amplifying factor. Tony Pasquariello, head of hedge fund coverage at Goldman Sachs, wrote in a client note that signs of leverage buildup were visible across the market — including rising retail margin balances, growing assets under management in leveraged ETFs, and increasing trading volume in short-dated options (E9).

Walter Todd, chief investment officer at Greenwood Capital in South Carolina, was more direct: "People have overextended their positioning in these stocks. Over the past few months, many assumed these stocks would only go up. If they borrowed money to buy in, they could now be facing forced margin calls" (E10).

How did the sell-off spread to global tech markets?

The pullback was not confined to Wall Street. South Korea's KOSPI index had already fallen into a bear market the prior week, even though it is still up nearly 62% year-to-date — almost matching the Philadelphia Semiconductor Index's own 60%-plus annual gain (E5, E8). Japan's Nikkei 225 fell into correction territory on July 17. In Europe, tech shares — which had posted their best quarterly gain since 2001 in June — became one of the worst-performing sectors of the week (E8).

How do market professionals interpret the pullback?

Toni Meadows, investment chief at BRI Wealth Management, framed the decline as a convergence of two forces: "This pullback reflects profit-taking, along with heightened scrutiny of the sustainability of AI capital spending. Semiconductor valuations had already priced in near-perfect demand. This is a cyclical industry, so after such a rapid rally, these stocks are prone to correction at some point" (E4).

What this means

The figures assembled here point to a tension rather than a single narrative. The Philadelphia Semiconductor Index's bear-market confirmation (a 20%-plus drop from its June peak, per E3/E12) sits alongside a year-to-date gain of more than 60% (E5) — the same pattern seen in South Korea's KOSPI, which is in a bear market despite a 62% annual advance (E8). Meanwhile, the immediate triggers cited — Moonshot AI's Kimi K3 (E2/E13) and Google's delayed Gemini 3.5 Pro (E7) — are product and competitive developments, not disclosed earnings misses. Layered on top, Goldman Sachs' Tony Pasquariello (E9) and Greenwood Capital's Walter Todd (E10) both flagged leverage — margin debt, leveraged ETFs, short-dated options — as a mechanism that could turn ordinary profit-taking into a sharper move. Taken together, the evidence assembled by money.udn.com and CNA suggests a market working through elevated valuations and leveraged positioning at the same time it is digesting new information about AI competition and spending, rather than a sell-off driven by a single confirmed fundamental shock.