The structure of the week
The market held together this week, but only barely and only where the largest weights carry the index. The S&P 500 rose 0.47% and the Nasdaq 100 rose 0.42%, both modest gains that keep price above the rising stack of moving averages and inside the upper half of the volatility bands. Underneath that, the equal-weighted version of the same index fell 0.44%, and small caps fell 1.40%. That gap is the whole story of the week: cap-weighted benchmarks advanced while the average stock did not. The broad structure remains upward — successive higher lows have not been threatened and the swing low from the spring correction is far below current price — but the participation supporting it thinned noticeably. A structure that is intact and a structure that is healthy are not the same thing, and this week the distinction became visible.
The rotation
Leadership narrowed to a single theme. Software rose 5.93% and cybersecurity rose 3.51%, far ahead of anything else, with communication services up 1.43%, technology up 1.30% and artificial intelligence up 1.25% filling out the top of the table. Financials rose 1.08%, the only non-technology sector in the leading group. The lagging end was crowded and cyclical: aerospace fell 4.22%, crude fell 3.67%, gold fell 3.42%, memory fell 3.21%, semiconductors fell 2.20%, biotechnology fell 2.16% and health care fell 1.98%. Industrials fell 1.73% and energy fell 1.51%. This is not defensive rotation — utilities fell 0.09% and staples fell 0.63%, so money did not move to safety either. It is breadth narrowing: capital concentrating into software and adjacent growth while everything cyclical, commodity-linked and small was sold. Narrow leadership can persist for a long time, but it removes the cushion that broad participation provides.
One week against four
The one-week versus four-week comparison sorts the noise. Software is a genuine continuation: up 5.93% on the week on top of 15.78% over four weeks, with price pushing above the upper volatility band and momentum at the top of its range — extension, but extension in the direction of the existing trend. Cybersecurity is the same shape, 3.51% on the week over 8.19% for the month. The contradictions are more informative. Biotechnology fell 2.16% this week yet is up 12.09% over four weeks; nuclear fell 1.52% against a four-week gain of 10.57%; memory fell 3.21% against 10.84%; gold fell 3.42% against 10.05%; health care fell 1.98% against 5.30%. These are pullbacks inside monthly uptrends, a loss of pace rather than a break — the trend still leads, the week merely lagged it. The genuinely weak names are the ones negative on both horizons: China tech, down 1.28% on the week and 7.62% over four weeks; retail, down 0.93% and 3.22%; utilities, down 0.09% and 3.65%; industrials, down 1.73% and 1.50%; real estate, down 1.33% and 1.31%. And semiconductors sit awkwardly in between — down 2.20% on the week against just 0.74% over four weeks, meaning a month of work was largely undone in five sessions while price still trades well beneath its longer smoothing line.
The crypto divergence
Crypto splits from the equity read on timing rather than direction. Bitcoin fell 0.38% on the week but is up 22.02% over four weeks; ether fell 1.23% against 29.27%; the bitcoin proxy rose 0.50% on the week with 23.18% behind it over four weeks. Solana was the exception, rising 8.39% on the week on top of 40.85%. So crypto delivered by far the strongest four-week move of anything we track, then went quiet while equities were rotating — a pause after a vertical advance, with price stretched far above its longer-term smoothing and the shorter averages still below the longer ones from the prior decline. Rates fit the same picture: long treasuries rose 1.01% on the week and 0.77% over four weeks, a small bid that did not come with defensive equity buying, and high-yield credit rose 0.16% on the week and 0.33% over four weeks — flat, not deteriorating. Credit is not confirming stress.
What would change this read
What would invalidate this read. First, the equal-weight index and small caps would need to break their recent swing lows rather than simply underperform; as long as they hold higher lows, narrow leadership is a quality problem, not a structural one. Second, credit would have to turn: high-yield rolling over from flat into sustained weakness would remove the main piece of evidence that this is rotation and not risk reduction. Third, the leaders themselves would have to lose their footing — software is currently trading above its upper volatility band, and an unwind that carried it back beneath its rising medium-term averages would leave the index without the engine that produced this week's gain. Fourth, semiconductors: they are already below their longer smoothing line despite a positive four-week number, and a decisive loss of the levels where recent lows formed would turn a lagging group into a broken one. Until one of those changes, the read stands as it is — an upward structure with thinning participation, led by a single theme, with the majority of the pullbacks still sitting inside intact monthly uptrends.