All Eyes on the Silicon
Ten years ago, semiconductors were three percent of the S&P 500. Today they are nearly a fifth of it, and on track to supply roughly 44% of the index’s second-quarter earnings growth. That concentration was earned, not inflated: the sector grew into its weight through profits. But it also means the index now inherits the sector’s temperament. When chips wobble, the market wobbles. In July, they wobbled hard, with more than $2 trillion coming off the complex since its late-June peak. This month’s spotlight walks through how we got here, what the leadership actually costs, and how to read the correction now.
The instinct is to treat the semiconductor trade as the NVIDIA trade. The data says otherwise. Across the 52 stocks in the sector, the median one is up 79% over the trailing twelve months, even after July’s sharp sell-off, and 49 of 52 remain positive. NVIDIA, at +11%, ranks near the bottom of its own sector. Leadership has migrated down the AI supply chain into memory, networking, and equipment, where names like Micron and Intel posted the kind of returns usually reserved for turnaround stories.
Critically, this rise was not investors simply paying up for the same story, what analysts call multiple expansion. Profits paid for it. Semiconductor earnings are growing at 133% year over year against 26% for the broader index, per LSEG. The sector’s rise from roughly 3% of the S&P a decade ago to nearly 20% today tracked its rising share of index profits. The weight is warranted. That is precisely what makes it consequential.
For two decades, semiconductors have swung two to three times as hard as the market, more on the way up and more on the way down. That ratio currently sits near 2.9x, toward the top of its two-decade range and within reach of the 2024 record just above 3x. Citadel’s trading desk made the point plainly this summer: semis are no longer an industry event, they are an index event.
The math compounds through earnings season. A sector responsible for roughly 44% of index earnings growth, and swinging at nearly three times the market’s volatility, turns every major chip earnings report into a market-wide event. This is the trade-off investors have quietly accepted in exchange for the earnings power. July was a reminder of the terms.
The correction gathered force through the month and accelerated into its final week. The Philadelphia Semiconductor Index fell into bear market territory, off more than 20% from its June peak, with more than half its members down at least 25%. Memory sat at the epicenter: SK Hynix fell roughly 30% from its high, Samsung and Micron more than 25%, pressured by the blockbuster Shanghai debut of Chinese memory maker CXMT and fresh questions about whether producers would hold the line on supply. A powerful new open-source AI model out of China added to the unease, and NVIDIA lost its title as the world’s most valuable company after reports it was in talks to backstop $250 billion of OpenAI’s financing, raising concern that AI’s biggest players are increasingly funding one another’s spending, a loop that can unwind fast if demand disappoints.
That last point matters because chips are where everyone else’s spending shows up first. The capital budgets of the hyperscalers (the giant cloud operators like Amazon, Microsoft, and Google) along with Oracle’s borrowing and the broader wave of corporate bond issuance, all converge on chip revenue, paid up front. So the stock-market wobble reached the credit market: hyperscaler bond spreads, the extra yield lenders demand to hold their debt, widened meaningfully over two months, and Oracle credit default swaps, the contracts that pay out if it defaults, traded at levels last seen in 2009. For the first time this cycle, the equity chart and the credit chart told the same story.
Semiconductors do not lead quietly and they do not exit gracefully. The sector’s history against the S&P is a sequence of multi-year regimes: a decade of dominance into 2000, a twelve-year give-back, a decade of quiet outperformance, and the current AI-era run since mid-2022 compounding at more than 30% a year over the market. Corrections inside a leadership regime are normal; the question is always whether the regime itself has turned. Morgan Stanley’s framing is that this is a midcycle adjustment rather than the end of the cycle, the fourth rotation between hyperscalers and semis since ChatGPT.
One valuation trap is worth flagging, because it runs opposite to instinct. With deeply cyclical stocks, a cheap-looking valuation is usually a warning rather than a bargain. These are companies whose profits swing hard with the economic cycle, and that inverts the usual signal. Their price-to-earnings multiple is lowest at the earnings peak, when the market has already begun pricing the coming decline, and highest at the bottom, when profits have fallen faster than the share price. Peter Lynch’s rule was blunt: buy a cyclical after several years of record earnings, once the multiple looks cheap, and you can lose half your money in a hurry. Memory has run this exact script. Micron hit a record high in May 2018 at four and a half times earnings, its most profitable stretch to that point, and prices for DRAM, the fast working memory in every PC and server, rolled over within months. The stock lost 57% by December, and revenue fell 23% the following year. The takeaway: a low multiple on a memory stock is often a peak signal, not a discount.
The band is well established. Memory producers tend to trade at four to eight times expected earnings when the cycle peaks, and in the high teens to low twenties once it breaks. Single-digit multiples across the complex today sit in the first band. Cheap semis have usually meant late semis.
But there is a real bull case, and it flies directly in the face of that trap. It turns on how this cycle is built. In past downturns, memory prices fell in the spot market the instant supply caught up with demand, and earnings fell with them. This time, most high-bandwidth capacity is locked into multi-year contracts, some running as long as five years, so that revenue does not drop just because spot prices do. The cushion is large enough to matter: even Bank of America’s bear case, which assumes DRAM and NAND prices (NAND being the flash memory that keeps data when the power is off) fall 30% and 40% in line with past downturns, still leaves memory earnings at roughly eight times the last cycle’s peak. On those numbers, today’s low multiple is a mispricing, not a warning. The catch is that the multiple looks cheap whether the bulls or the bears are right, so valuation alone settles nothing.
The honest frame from here is simple. Semiconductors are the index’s center of gravity on market cap, earnings, and risk, and what happens to them happens to everyone. The story will be judged through one lens: whether the earnings hold up, powered by hyperscaler spending and demand for computing. TSMC’s record quarter argues the profits are still there; the credit market’s July flinch argues the financing behind them now draws scrutiny. For the next two quarters, the chip earnings calendar is the market’s calendar.
The RQA Economic Forecast Model advanced to 0.34 in July, its strongest reading since December 2024. The spring plateau proved to be exactly that, a pause rather than a turn, and the model has resumed the climb it began in January. The signal holds comfortably in positive territory, and the improvement is broad: consumer spending, output surveys, and credit conditions have all contributed, with labor the one channel still subtracting. The level remains short of what a mature expansion produces. But the trajectory since January is unmistakable, and the message has moved past continuity: growth, broadening.
June delivered a clean break in the inflation story. After four months of energy-driven acceleration, headline CPI fell to 3.5% from May’s 4.2% peak on the largest monthly energy decline since April 2020. Core CPI eased to 2.6% and core PCE to 3.3%. May now looks like the high-water mark, and the disinflation is arriving from altitude rather than from target.
The consumer did the opposite of what the sentiment surveys implied. Retail sales posted their strongest reading since early 2022, real spending accelerated, and the consumer composite reached its best level in four years. Output confirms it: both ISM business surveys sit solidly in expansion territory, with new orders leading, though industrial production and permits stay soft. Growth is still running through order books and services rather than physical output.
Labor is the dissent. June payrolls rose just 57,000 against expectations north of 100,000, prior months were revised down by a combined 74,000, and the unemployment rate’s dip to 4.2% flattered the report only because participation hit its lowest level since early 2021. The composite holds positive on well-behaved claims and a firmer workweek, but the flow measures, payrolls and the employment-to-population ratio, keep deteriorating. Jobs are not being lost; they are simply not being created.
Financial conditions remain a tailwind: an upward-sloping yield curve on both measures, corporate borrowing costs far below where they sat a year ago, equities up 18% year over year. July complicated that at the margin. The semiconductor correction pushed hyperscaler borrowing costs higher and knocked the market’s leadership into a bear market, and the Fed’s hold arrived with a hawkish edge: a 9-3 vote with three members dissenting in favor of a rate hike, the most since 2016, and long-term Treasury bonds selling off to their highest yields since 2007. Read against re-accelerating growth, those dissents look less like caution than arithmetic.
| Indicator | Jul-26 | Jun-26 | May-26 | Apr-26 | Mar-26 | Feb-26 | Jan-26 | Dec-25 | Nov-25 | Oct-25 | Sep-25 | Aug-25 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Labor | ||||||||||||
| Non-Farm Payrolls (YoY%) | -0.5 | -0.4 | -0.5 | -0.5 | -0.5 | -0.3 | -0.0 | 0.2 | 0.4 | 0.3 | 0.5 | 0.5 |
| Initial Unemployment Claims (Inverse YoY%) | 9.6 | 8.9 | 10.4 | 14.9 | 6.2 | 12.4 | -1.0 | 9.1 | -1.4 | -0.9 | 0.0 | 0.9 |
| Employment-to-Population Ratio (YoY%) | -1.2 | -0.8 | -1.5 | -1.2 | -1.0 | -0.5 | -0.5 | -0.3 | -0.5 | -1.0 | -0.7 | -0.7 |
| Average Weekly Hours Worked (YoY%) | 0.6 | 0.3 | 0.0 | 0.0 | 0.6 | 0.9 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| RQA Labor Composite (YoY%) | 2.1 | 2.0 | 2.1 | 3.3 | 1.3 | 3.1 | -0.4 | 2.2 | -0.4 | -0.4 | -0.0 | 0.2 |
| Commercial Output | ||||||||||||
| ISM Manufacturing PMI (% over Base) | 6.6 | 8.0 | 5.4 | 5.4 | 4.8 | 5.2 | -4.2 | -4.0 | -2.4 | -2.2 | -2.2 | -3.2 |
| ISM Services PMI (% over Base) | 8.0 | 9.0 | 7.2 | 8.0 | 12.2 | 7.6 | 8.8 | 5.2 | 4.8 | 0.0 | 4.0 | 0.2 |
| Industrial Production Index (YoY%) | -1.3 | -0.9 | -1.3 | -2.0 | -1.6 | -1.1 | -0.8 | -0.2 | 1.6 | 1.2 | 0.8 | 1.1 |
| Residential Real Estate Permits (YoY%) | -2.1 | 1.4 | 2.1 | -7.4 | -5.5 | -2.4 | -4.8 | -12.8 | -7.3 | -8.1 | -11.1 | -3.0 |
| Income & Consumption | ||||||||||||
| Real Personal Incomes (ex. Transfer Receipts) (YoY%) | 0.7 | 0.2 | -0.3 | 1.1 | 2.0 | 1.1 | 1.3 | 1.9 | 2.1 | 1.8 | 1.9 | 3.8 |
| Retail Sales (YoY%) | 3.2 | 2.5 | 0.7 | -1.0 | -0.6 | -0.6 | -1.4 | -1.3 | -0.8 | -0.3 | 0.0 | -0.4 |
| Real Personal Consumption Expenditures (YoY%) | 2.1 | 1.8 | 1.2 | 1.2 | 1.7 | 1.7 | 1.1 | 1.2 | 1.5 | 1.6 | 2.0 | 1.8 |
| RQA Consumer Spending Composite (YoY%) | 2.6 | 2.1 | 1.0 | 0.1 | 0.5 | 0.5 | -0.1 | -0.0 | 0.4 | 0.7 | 1.0 | 0.7 |
| Financials & Sentiment | ||||||||||||
| Treasury Yield Curve Spread - 10-Yr Less 3-Month | 0.9 | 0.6 | 0.8 | 0.7 | 0.6 | 0.3 | 0.6 | 0.5 | 0.1 | 0.2 | 0.1 | 0.0 |
| Treasury Yield Curve Spread - 10-Yr Less 2-Yr | 0.5 | 0.3 | 0.5 | 0.5 | 0.5 | 0.6 | 0.7 | 0.7 | 0.6 | 0.5 | 0.6 | 0.6 |
| Corporate Bond Spreads (Inverse YoY%) | 30.4 | 28.0 | 16.4 | 12.5 | 15.0 | 12.9 | 1.7 | 6.2 | 10.3 | 16.2 | 11.0 | 9.7 |
| U.S. Monetary Base (YoY%) | -4.5 | -1.9 | -4.6 | -5.5 | -4.0 | -3.8 | -4.1 | -5.6 | -3.7 | -2.0 | 0.3 | 1.5 |
| S&P 500 Return (YoY%) | 18.1 | 20.9 | 28.2 | 29.4 | 16.3 | 15.5 | 14.9 | 16.4 | 13.5 | 19.9 | 16.1 | 14.4 |
| Consumer Sentiment (YoY%) | -10.5 | -18.5 | -14.2 | -4.6 | -6.5 | -12.5 | -21.3 | -28.5 | -29.0 | -24.0 | -21.4 | -14.3 |
| Inflation & Money Supply | ||||||||||||
| CPI (YoY%) | 3.4 | 4.2 | 3.8 | 3.3 | 2.4 | 2.4 | 2.6 | 2.7 | 2.8 | 3.1 | 2.9 | 2.7 |
| Core PCE (YoY%) | 3.4 | 3.6 | 3.5 | 3.4 | 2.9 | 2.9 | 2.8 | 2.6 | 2.5 | 2.8 | 3.1 | 3.1 |
| M2 Money Supply (YoY%) | 5.2 | 5.1 | 4.3 | 4.2 | 4.6 | 4.1 | 4.1 | 4.1 | 4.6 | 4.7 | 4.8 | 5.0 |
By the levels framework (growth against zero, inflation against 2%), the economy sits where it has for nearly the entire post-COVID era: Quadrant II, the inflationary boom, with the growth model at 0.29 and CPI at 3.8% on the three-month basis the map uses. What changed is the vector. Growth is rising and inflation is falling, which points at Quadrant I without qualification. That combination has held once in this sample: the seven months from October 2020 through April 2021.
That is a long way to travel. Inflation has roughly 1.8 percentage points to cover before it crosses the line, and almost all of that has to come from energy. If the Gulf premium keeps bleeding out and June’s decline extends, the disinflationary expansion is available before year-end, and the framework favors equities, corporate credit, and precious metals. If oil reprices instead, inflation re-accelerates against a Fed with three members already voting to hike, and the setup inverts quickly.
The July semiconductor correction sharpens the sector implication. Regime transitions reward durable earnings over beta, a stock’s tendency to amplify the market’s moves, and the spotlight’s conclusion applies directly: the market’s largest sector is being re-underwritten on the sustainability of its profit engine in real time. Until that resolves, the regime favors quality balance sheets, diversification away from the most crowded leadership, and patience on adding cyclical risk. The path runs through three gates: whether energy stays contained, whether the labor slowdown stabilizes or steepens, and whether September becomes the hike the dissenters wanted. Futures now put those odds better than even.
Growth is broadening, inflation has crested, and the index’s largest sector is being repriced on the durability of its earnings. Quadrant II with a Goldilocks vector — constructive, but the route runs through the energy market and a Fed that is no longer unanimous.
Disclosures. These materials have been prepared solely for informational and educational purposes and do not constitute investment advice or a recommendation to make or dispose of any investment or to engage in any particular investment strategy. Information and data shown were obtained from sources believed to be reliable, but accuracy is not guaranteed. All investments involve risk, including the potential loss of principal. Past performance is not indicative of future results. References to specific securities and issuers are for illustrative purposes only and are not intended as recommendations. Third-party research and market commentary are cited for context and do not represent the views of RQA. Richmond Quantitative Advisors, LLC is an SEC-registered investment adviser; registration does not imply a certain level of skill or training.