U.S. Treasury yields surged in September as markets repriced the return required to hold long-dated government debt, with the 10-year yield rising 54 basis points to 5.29%, according to QuantStreet Capital. The move was driven overwhelmingly by higher real yields rather than a deterioration in inflation expectations, creating a difficult backdrop for rate-sensitive stocks while leaving semiconductor companies tied to artificial intelligence infrastructure among the market’s strongest performers.
QuantStreet Capital said the 10-year Treasury yield rose from 4.75% during the month, contributing to losses ranging from 2.3% to 5% across parts of the U.S. fixed-income market. The yield climbed further in early October, briefly reaching 5.35%, a level last seen roughly 24 years ago.
The scale and composition of the selloff matter for financial markets because real yields raise the effective cost of capital across the economy. Companies, households and governments face higher borrowing costs when inflation-adjusted Treasury returns rise, even if markets are not substantially lifting their inflation forecasts.
Real yields accounted for most of September’s Treasury move
The 10-year real Treasury yield, measured through Treasury inflation-protected securities, increased to 2.93% from 2.44% in September, a 49-basis-point rise, QuantStreet Capital said. Over the same period, the 10-year breakeven inflation rate — the market-based measure derived from the gap between nominal and inflation-protected Treasury yields — edged up only 5 basis points to 2.36%.
That means more than nine-tenths of the increase in the nominal 10-year yield reflected the real-yield component rather than inflation compensation.
The distinction places the September move closer to a repricing of growth, funding demand and risk compensation than to a simple inflation scare. Rising real yields can reflect expectations for stronger long-run economic growth, but they can also include a higher premium demanded by bondholders for duration risk, fiscal uncertainty and volatile Treasury supply.
Reuters reported on October 7 that the 10-year Treasury term premium, the additional return sought for holding long-term bonds rather than rolling short-term debt, had climbed to around a 12-year high. That raises the possibility that part of the increase in measured real yields reflected investors demanding more compensation for the uncertainty of holding long bonds, rather than a pure improvement in the economic outlook.
The United States is also confronting substantial borrowing needs. Higher Treasury issuance can pressure long-duration securities when the supply of bonds rises faster than demand from buyers willing to hold them at prevailing yields. A projected federal deficit of $1.9 trillion and elevated energy prices have added to market concerns about the fiscal and inflation outlook, although September’s breakeven data indicates that inflation pricing itself moved relatively little during the month.
AI spending offers one explanation, but the evidence remains incomplete
QuantStreet Capital’s monthly discussion pointed to a second force: rising capital demand connected to AI infrastructure. Data centers, graphics processing units, networking equipment and power systems require large upfront investment, often financed over long periods. If businesses seek more long-term funding while the supply of long-duration savings does not increase at a comparable pace, borrowing costs can rise.
ING research has similarly argued that AI could affect yields through several channels, including corporate capital expenditure, productivity prospects and long-run growth expectations. A stronger growth outlook would typically raise real yields by increasing the expected return available from productive investment.
The U.S. dollar strengthened in September, according to the discussion cited by QuantStreet, an outcome that does not neatly fit a broad retreat from dollar-denominated assets. That pattern leaves room for an interpretation centered on domestic growth and capital demand, alongside concerns about Treasury supply and term premium.
Neither explanation excludes the other. TIPS yields themselves can contain a real term premium, so the 49-basis-point increase in the 10-year real yield cannot be treated as a precise measure of improved growth expectations. Markets may be pricing a combination of stronger demand for capital, heavy government borrowing and greater uncertainty around long-term interest rates.
Semiconductor shares separated from the broader equity market
The rise in yields divided U.S. equities sharply. The VanEck Semiconductor ETF, which trades under the ticker SMH, gained about 9.4% in September, while the equal-weight S&P 500 index declined about 4.8%, according to QuantStreet Capital. The Nasdaq remained positive for the month, aided by its heavy exposure to large technology companies.
Small-cap shares, real estate investment trusts, utilities and financial stocks weakened as yields rose. These sectors tend to be more exposed to refinancing costs, interest-rate-sensitive valuations or changes in the shape of the yield curve.
The divergence also showed how narrow the market’s earnings optimism had become. QuantStreet identified companies such as AMD, Micron, Intel, Cisco and Applied Materials as part of the momentum trade around AI infrastructure. Chipmakers and equipment providers can benefit early when companies commit to new computing capacity, because their sales are directly connected to the investment cycle.
The firm described the remainder of the corporate sector outside these hardware-linked winners as “ROCS.” Its argument was that many downstream companies funding AI projects had yet to demonstrate a broad improvement in profits that would match the scale of spending on chips, data centers and supporting equipment.
That gap creates pressure on the economics of the AI buildout. Higher real yields increase financing costs precisely as many companies are committing capital to longer-payback projects. If productivity gains and revenue growth emerge slowly, returns on those investments could be squeezed even while upstream suppliers continue reporting strong demand.
U.S. Bureau of Labor Statistics data has shown labor-productivity growth improving from the average pace recorded after 2010. The data does not isolate AI’s contribution, and it does not establish that productivity gains have already become large enough to offset the cost of the current investment cycle.
Bond volatility complicates the case for long-duration risk
QuantStreet said it began modestly extending duration in lower-risk portfolios after the 10-year yield approached 5.25%, while maintaining an overall duration position below its benchmark. The move suggests the firm sees yields as becoming more attractive, but not enough to abandon caution while term-premium and fiscal pressures remain elevated.
For cryptocurrency markets, the Treasury move offers a less direct signal than it does for bonds or highly rate-sensitive equities. The material provided cited a 90-day rolling correlation of negative 0.17 between Bitcoin and the 10-year Treasury yield in September, indicating little stable short-term relationship between the two measures.
Bitcoin’s decline to roughly $81,162 during the recent market repricing nonetheless showed that sharp moves in global funding conditions can coincide with stress across risk assets. Its reported correlation with gold also rose to 0.59, the highest level in six years, though correlations can change rapidly and do not establish a durable shift in trading behavior.
The immediate test for markets is whether higher yields begin to slow spending beyond the most profitable AI suppliers. A sustained rise in real rates would place greater scrutiny on debt-heavy businesses, property markets and companies making large technology investments whose financial returns remain years away.
Explore how rising yields intersect with crypto markets and tokenized Treasuries in our guide to TradFi and its mechanics today.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
