Semiconductors broke. This week tests whether the damage spreads

Last week broke the momentum in semiconductors. This week tells us whether the damage spreads.

Here is the full picture as I see it going into the week: what broke, what held, what could change the answer and what I am doing about it.

What broke

The S&P 500 fell 1.6% last week and the Nasdaq dropped 2.9%. Those index numbers hide the more important event underneath.

The Philadelphia Semiconductor Index recorded its worst week in more than a year and finished more than 20% below its late-June record—bear-market territory for the sector that carried the first half. The index is still up more than 60% in 2026, which shows how extraordinary the preceding run had been. Reuters

After that kind of advance, this was not a routine wobble. The momentum trade that powered the first half has, for now, stopped working.

The underlying question is familiar: can the enormous spending on AI infrastructure convert into revenue and cash flow quickly enough to support the valuations investors were willing to pay in June?

A new Chinese model added pressure to that debate. Moonshot AI released Kimi K3, a 2.8-trillion-parameter open-weight model that the company says competes with leading U.S. systems. Its release does not prove that it caused the U.S. semiconductor selloff. It does reinforce the question investors are already asking: if capable models become cheaper and more widely available, how durable are the margin and capital-spending assumptions embedded across the AI supply chain? Reuters

The market is no longer rewarding AI exposure automatically. It is asking for evidence.

What held—and why it matters

Here is the reason I am not calling this a broad market problem yet.

While semiconductors entered a bear market, Apple reached a fresh record and reclaimed its position as the world’s most valuable company. The Russell 2000 lost only 0.5% for the week and remains up roughly 19% this year. Energy benefited as oil rose, and the damage outside the market’s former chip leadership was materially lighter. Reuters AP

That does not prove the selloff is contained. Friday’s weakness broadened beyond chips, and that deserves respect.

But the evidence still looks more like a rotation under pressure than a wholesale exit from equities: capital moving away from the most expectation-loaded part of the market and toward businesses with more visible present cash flows and less demanding assumptions.

Rotation was my central first-half theme. Last week was its sharpest expression yet.

This week tests whether that interpretation survives.

Four reports—and guidance over results

Alphabet and Tesla report Wednesday. They are very different businesses carrying the same burden: large expectations and heavy spending tied to the future.

For Alphabet, I will be watching cloud growth, advertising demand, capital expenditure and evidence that AI investment is translating into durable revenue.

For Tesla, the important questions are automotive margins, deliveries, energy growth, free cash flow and the cost of its robotaxi, AI and Optimus plans.

The semiconductor sector faces its more direct test through Texas Instruments on Wednesday and Intel on Thursday. Texas Instruments Intel Tesla

After the selloff, guidance will matter more than the quarter that just ended. The market already knows that recent demand was strong. What it needs now is evidence that future demand, margins and capital spending can support what investors were paying in June.

The reaction to strong results from Samsung and TSMC already showed the problem: when expectations are priced close to perfection, “great” can still disappoint.

The wildcard underneath: Iran and oil

The conflict intensified through the weekend, adding another layer of risk beneath the earnings story.

Brent finished last week near $88, briefly traded above $90 on Monday and then moved back toward $88. The exact intraday price matters less than the persistence of the pressure. Reuters

Sustained higher oil affects inflation, yields, consumers and corporate margins at the same time.

That is why oil is the variable most capable of turning a contained sector rotation into a broader market problem. It attacks the “everything else” side of the market that is currently helping absorb the weakness in chips.

One spike does not establish a new inflation trend. Persistent pressure would matter.

What I am doing

Two trades stopped out Friday: AMD and JPM.

Both setups were valid when entered. Both failed. The losses were taken, documented and kept small at the portfolio level. That is what the risk rule is designed to do: make any single failed idea manageable.

The broader portfolio ended the week close to flat, and most of my eToro account remains in cash.

That cash matters more this week than usual.

If earnings disappoint and the selloff deepens, companies I actually want may be pushed toward clean demand zones. That is not automatically a threat. It may create the kind of setup I have been waiting for.

But I am not buying semiconductors simply because they have fallen.

A 20% decline from a peak does not make a stock cheap, and it does not create demand by itself. Price still has to reach a defined level. The structure still has to make sense. The stop still has to be acceptable.

If nothing reaches a valid zone, I do not force a trade.

If a setup appears, the cash is ready.

Rotations create opportunities, but only for investors who retain enough patience and liquidity to use them.

Cash ready. Levels marked. Watching earnings.

This is a record of my process and opinions, not investment advice. I hold positions in securities mentioned or related. Copy trading involves risk, including loss of capital. Past performance is not an indication of future results.