Smart Money Is Not Chasing AI, and the Market Is Moving Toward Its Risk Lines
US Stocks · Options · News · Views
The indexes are holding up, but the tape is already weak
The market did not look terrible on the surface. Nasdaq still managed to close slightly higher. But underneath the index, the tape was weak. Only about 112 names in the S&P 500 rose, while 391 fell. Out of 11 sectors, only two were up and nine were down. This was not a healthy broad rally. It was a market held together by a few large tech weights.
The Dow was the strongest part of the market the day before, then became the weakest. The S&P 500 also flipped back to more decliners than advancers. The market is not collapsing. It is becoming more selective. Groups without a new story or strong buying are getting pushed aside first.
That matters for the next read. An index that has not broken does not mean risk is gone. The real question is whether the large weights can keep carrying the market, and whether risk lines start spreading from single sectors into broader ETFs.
AVGO and NVDA are still keeping tech alive
The brightest large-cap tech names were AVGO and NVDA. AVGO rose 4.83% on volume after signing a chip development and supply agreement with Apple worth more than $30 billion. The timing was interesting because the news came near an important technical area. It shows that, even with semis shaky, the market is still willing to buy visible orders.
NVDA also rose 3.65% on volume. The driver was news that China may allow some large AI companies to buy small quantities of H200 chips, including Alibaba, ByteDance, and DeepSeek. The key is not the exact volume. It is the marginal improvement in expectations. Current NVDA forecasts do not include China sales, so any reopening there is additive.
The long-term tension in NVDA has not changed. The company is not weak. The issue is size. The bigger the base, the harder it becomes to maintain high growth. Even if AI capex growth slows, NVDA can still take a share of the spend. The market will simply debate the right forward PE multiple. The current price is not absurd, but it behaves more like a mature core tech asset than a small growth stock that can double easily.
The family-office signal is not an AI exit. It is a refusal to chase
UBS's 2026 global family office report carries a signal worth watching. The survey covered 307 family offices with average assets of about $2.7 billion. These investors tend to be conservative. They care less about extreme alpha and more about long-term asset growth and risk control.
A record 60% of respondents plan to change asset allocation over the next 12 months, compared with a previous high of 37%. Their focus is not short-term market noise. They are preparing for geopolitical conflict, debt stress, and financial-market risk. Large capital is not ignoring the AI story. It is raising the risk weight.
On AI, 65% of family offices already hold related assets, but only 7% plan to increase AI investment over the next 12 months. That does not mean they are bearish on AI. It means the low-cost allocation has already happened, and they do not want to keep chasing at higher levels.
This matters for market psychology. The biggest risk for the AI trade is not immediate collapse. It is slower growth. If the large tech companies can monetize AI and keep funding capex, semis can keep telling the story. If monetization takes too long, spending may not disappear, but its growth rate can slow. For stocks priced for perfection, slower growth is already a risk.
SOXX is not fixed just because it bounced
The easiest mistake in SOXX is misunderstanding a breakdown. A bounce after a break does not automatically cancel the break. Once support fails, that old support can become resistance. Winners and trapped buyers can both become sellers on the rebound.
So a move back above 582 only means stabilization, or a slower decline. A real upside reversal needs 621. Without 621, trapped buyers are not fully released, and momentum buyers do not have enough confirmation.
That is the trade-off between risk and reward. Chasing here can capture more upside. It can also mean taking the other side of someone else's exit. Mature trading is not only asking how much can be made. It starts with asking what happens if the trade is wrong. For SOXX, the map is simple: 582 is stabilization, 621 is reversal confirmation.
COST is still growing, but the growth rate is slowing
COST did not report bad sales on the surface. June sales hit the highest monthly level so far this year. But comparable sales growth slowed to 8.8%, the lowest since February, below 12.5% in May and 11.6% in April. Excluding gas and FX, US comparable sales rose 7.6%, Canada rose 4.9%, and other international markets rose 5.6%.
This is not a bad company suddenly turning bad. COST remains a very strong consumer leader. The problem is valuation and trend. Growth is still there, but the speed is slowing. Markets are often more sensitive to the growth curve than to the absolute growth number.
Consumption still matters for the US economy. Recently, AI investment has carried the growth narrative. If consumption slows and AI capex also slows later, the GDP support structure gets thinner. COST is still expensive, and 867 remains the support line to watch. If 867 breaks, mean reversion becomes the risk, not a simple bargain-hunting setup.
Three ETF lines: XLK, XLB, and XLY
The key risk lines now are XLK, XLB, and XLY. These are not side sectors. They are important parts of the S&P 500 structure. If tech, materials, and consumer discretionary weaken together, the broader market and QQQ will have a harder time staying insulated.
For XLK, the key line is 172. A break below that would turn the structure lower, with roughly 15% downside toward intermediate support. That potential move looks similar to the kind of drawdown risk in QQQ, and 172 is not far away.
For XLB, the key area is 48.6 to 49.7. If it breaks, the next meaningful stabilization area is around 45.4, or about 10% lower. For XLY, the key zone is 110 to 113. If that breaks, the next area is near 100, also roughly 10% lower.
One sector breaking can be absorbed by other weights. Three major groups leaning on support together is different. The point is not to be blindly bearish. It is to know what cannot break. On the way up, markets tell stories. On the way down, they trade levels.