Nasdaq Takes the Hit: Chip Selloff Splits the Tape as Energy Catches a Hormuz Bid

The Nasdaq 100 is one of the cleanest tells in the opening snapshot. While the Dow Jones Industrial Average slipped less than 1% and the energy sector caught a geopolitical bid, the tech-heavy QQQ ETF fell 4.2% on the week — a divergence that puts the entire AI infrastructure thesis under simultaneous stress from capex anxiety, a potential rate hike, and a Strait of Hormuz oil shock that nobody saw coming four weeks ago.

Here is the week in five numbers:

The split is not subtle. The Dow’s relative stability masks a semiconductor correction that erased more than $1 trillion in chip-sector market value and pushed the Philadelphia Semiconductor Index down roughly 11% — its steepest weekly decline in over a year.^[1]

TSMC posts record earnings — and its stock falls

Taiwan Semiconductor Manufacturing Co. reported second-quarter revenue of NT$1,270.38 billion, up 36% year-over-year, with net income surging 77.4% to NT$706.56 billion.^[2] By any conventional measure, this was a blowout. Yet TSMC shares closed down more than 3% on Friday^[3], and the selloff spread across the chip complex: Applied Materials fell 5.5%, Intel dropped 2%, and AMD slipped more than 1%.^[3]

The disconnect is capex. TSMC’s announcement of higher capital expenditure than previously forecast rattled investors who are already questioning whether the massive AI infrastructure buildout can generate returns fast enough to justify the spending. Nvidia, the sector’s bellwether, lost nearly $200 billion in market value over the week and closed Friday at $202.81.^[1]

This is not a demand problem — TSMC’s numbers prove AI chip demand is alive and well. It is a cost-of-capital and return-on-investment problem. And it arrives at a dangerous moment for the cost side.

The rate-hike specter

Futures markets now price a roughly 52% probability of an interest rate hike — not a cut — at the September Federal Reserve meeting.^[3] That is a striking shift. The Fed funds rate sits at 3.63%, CPI inflation at 3.46% year-over-year, and the 10-year Treasury yield at 4.55%.^[4] The yield curve is positively sloped at 37 basis points (10-2Y), which historically signals expansion, not contraction.^[4]

But the inflation pulse is not dead. Consumer inflation expectations for the next year stand at 4.2%, down from 4.6% in June but still well above the Fed’s comfort zone.^[5] Real GDP grew 2.66% year-over-year as of the latest reading^[4] — solid, not overheating. Yet the combination of sticky inflation expectations and an oil price shock pushing through gas pumps could force the Fed’s hand in the wrong direction for equity valuations built on cheap capital.

The macro backdrop resembles mid-2006 more closely than any other period in the database — unemployment near 4.2–4.7%, CPI in the mid-3% range, a Fed pausing but not yet cutting.^[4] That analog ended with a housing-led recession 18 months later. The parallel is imperfect, but the pattern is worth flagging: a Fed that pauses while inflation stays sticky is not a green light for duration risk.

The Hormuz bid

Renewed hostilities between the United States and Iran over the Strait of Hormuz sent Brent crude surging roughly 5% to around $80 per barrel early in the week.^[6] President Trump reinstated a U.S. naval blockade on Iranian shipping and declared that all non-Iranian cargo transiting the strait must pay a toll — reportedly a 20% charge.^[6] Tanker traffic through the strait has collapsed, with no large vessel crossing via the U.S.-coordinated route.^[6]

The energy sector absorbed the shock directly: XLE rose 4.7% on the week.^[7] This is the inverse of the chip selloff — an old-economy sector catching a geopolitical risk premium while the new-economy sector sheds one. For a market that has spent two years buying AI and selling oil, the rotation is jarring.

The oil shock also feeds back into the inflation loop. If Brent holds near $80 and the Hormuz disruption persists, the consumer sentiment bounce documented this week could reverse quickly.

The sentiment head-fake

The University of Michigan’s preliminary Index of Consumer Sentiment rose to 54.4 in early July from 49.5 in June — a 10% jump and a five-month high, driven by falling gas prices.^[5] It was the second consecutive month of a 10% improvement.^[5]

But the survey was conducted before the latest escalation in US-Iran hostilities and the subsequent oil price spike. The same report flagged that consumers expect prices to rise at an annual rate of 4.2% over the next year.^[5] If gas prices climb back on Hormuz fears, the sentiment improvement is likely to prove transient — and the inflation expectations that accompany it could harden.

This is the pattern to watch: a sentiment indicator that bounces on a variable (gas prices) that is simultaneously being shocked by a geopolitical event the survey did not capture. The July read may be a local peak, not a trend.

What to watch next

• Chip earnings calendar. TSMC set the tone; the rest of the semiconductor complex reports through August. If capex guidance from NVDA, AVGO, and others mirrors TSMC’s upward revision, the capex-anxiety selloff could deepen. If guidance is disciplined, the correction may exhaust itself.
• Hormuz tanker traffic. The oil bid is entirely geopolitical. Any de-escalation or resumption of shipping through the strait would reverse the energy move quickly. Continued disruption pushes Brent toward $90 and feeds the inflation loop.
• September Fed meeting. With futures pricing a 52% chance of a hike^[3], the next CPI print and the August jobs report become high-conviction events. A hot CPI would lock in the hike scenario; a soft one would give the Fed room to hold.
• Consumer sentiment revision. The final July Michigan reading will reflect post-escalation survey responses. If the preliminary 54.4 is revised sharply lower, the sentiment bounce is confirmed as a head-fake.
• Small-cap resilience. Small caps have held their 2026 advantage even as the Nasdaq faltered^[8] — a rotation signal worth monitoring if the chip selloff extends.

The cleanest read on this market is the QQQ-versus-XLE spread. When the AI trade and the oil trade move in opposite directions at this magnitude, something is breaking under the surface. It may be as simple as profit-taking after a long run. It may be the first crack in a capex narrative that has carried the market for two years. The data does not yet tell us which — but the divergence itself is the signal.