The 10-year Treasury yield breaking above 5% has not crushed the unprecedented AI investment wave built on massive AI infrastructure expansion and expectations of widespread AI adoption, but the market is now examining a deeper risk.
As pricing probabilities for at least three more Federal Reserve rate hikes this cycle rise significantly, the 2-year Treasury yield has begun climbing faster than the 10-year Treasury yield, and the U.S. bond market appears to be signaling that inflation expectations and fiscal deficit-related term premiums are pushing up risk-free yields on 10-year and longer maturities, along with a more complex signal.
Inflation and fiscal concerns, combined with record-breaking AI-related debt issuance, are driving yields higher, but continued Fed rate hikes could also gradually erode the soft-landing expectations the Fed has long sought and the future growth prospects of the U.S. economy.
Bond market pricing shows the U.S. Treasury market is approaching a critical signal: a series of Fed rate hikes will begin to shift the market narrative toward U.S. economic stall and even toward the dual risk of an earnings recession plus an economic recession in the stock market.
Last week, the extra yield investors demanded to hold 10-year U.S. Treasuries over 2-year notes briefly narrowed to just 17 basis points, the tightest gap since early 2025.
This so-called yield curve flattening raises the possibility that the 10-year Treasury yield, known as the anchor of global asset pricing, will soon fall below shorter-term Treasury yields.
This closely watched phenomenon, in which the 2-year Treasury yield exceeds the 10-year Treasury yield, is called a yield curve inversion, and it is one of the important leading indicators of an approaching recession.
An even more important market pricing trajectory is that, as a 10-year Treasury yield above 5% has failed to crush this unprecedented AI investment boom centered on massive AI infrastructure expansion and expectations of large-scale AI application penetration, the market seems to be starting to worry that the real killing line for this AI investment wave may be the so-called U.S. Treasury yield curve inversion.
A 10-year Treasury yield above 5% mainly reflects higher nominal financing costs, and tech giants such as Microsoft, Meta, and Google, with abundant free cash flow and extremely high expected returns on investment, can easily absorb interest costs.
By contrast, once the yield curve inverts again deeply or inverts because recession expectations escalate, it often means a collapse in financial system liquidity, a collapse in corporate earnings expectations, and a contraction in aggregate demand across society, at which point even powerful tech giants would be forced to cut capital spending on AI data centers and computing hardware, truly touching the killing line of the AI boom.
If the curve does invert further in the future, accompanied by tighter credit conditions, corporate financing difficulties, and weakening order expectations, market worries about the end of the AI investment boom could escalate across the board.
Previously, the rise in the discount rate on the denominator side of the DCF valuation model mainly compressed valuations; from a theoretical perspective, the 10-year Treasury yield is equivalent to the risk-free rate r on the denominator side of the DCF valuation model, an important stock market valuation model.
Then, after a yield curve inversion market pattern emerges, there could also be a sharp downward revision in earnings forecasts on the numerator side.
And the previously seen long-end Treasury yields above 5%, meaning that even if 10-year and longer yields fall back due to growth concerns, earnings downgrades and a widening risk premium could offset the valuation support brought by a lower risk-free rate.
However, inversion itself is an early warning indicator and cannot be mechanically equated with recession or a peak in AI tech benchmark stock indices; the Cleveland Fed has also clearly pointed out that the yield curve inversion from late 2022 to late 2024 sent an unrealized recession signal.
Muse and Astra aggressively expand application boundaries: AI growth expectations fully confront global funding cost pressure
The earnings imagination brought by the expansion of AI agents and the valuation pressure caused by rising 2-year and 10-year-plus long-term Treasury yields are becoming two of the most striking opposing forces on Wall Street, yet at least so far the answer from U.S. equity market capital is that the unprecedented AI computing expansion and AI application boom brought by AI agents have outweighed a series of important bearish factors, including Treasury yields rising above 5%.
The 10-year Treasury yield breaking above 5% has not yet crushed the AI rally, but the market is examining a deeper risk: whether the Fed, if it truly begins a sustained monetary tightening cycle, will move from compressing stock market valuations to further transmitting into slower earnings growth or even earnings declines against the backdrop of a yield curve inversion, and even into a new recession for the entire economy.
In the week ended September 25, the Nasdaq 100 still rose 3.3%, the largest weekly gain since early August, and had earlier closed at a record high through Tuesday, showing that AI commercialization and corporate earnings expectations are still supporting risk appetite.
However, this resilience only significantly proves that growth expectations have temporarily offset part of the rate pressure; it does not mean tech stocks have escaped the grip of high interest rates.
The support that AI agents provide to the stock market comes from the expansion of application scope and potential revenue sources.
In the Connect summary published on September 24, Meta announced plans to bring Muse to AI glasses in the coming months and to expand retail, payment, and office connectors, including GitHub, Notion, and Box work scenarios; OpenAI's GPT-6 Astra strengthened computer operation, software engineering, and multi-step professional task capabilities.
Moving from answering questions to executing tasks means AI can enter more workflows originally completed by humans and create new commercial space for subscriptions, usage-based pricing, and enterprise services.
This change provides an important basis for long-term earnings growth in AI-driven stocks and explains why the market is still willing to preserve growth expectations for the AI supply chain in a high-rate environment.
But for stock pricing, increased workload must ultimately translate into revenue, profit, and free cash flow to continuously counter valuation compression caused by rising discount rates.
Wall Street financial giant Jefferies recently said that, driven by the twin engines of the AI investment boom and better-than-expected earnings from AI-related companies, the S&P 500 is expected to surge to 8,000 by the end of 2026 and further reach 9,000 in 2027.
Jefferies' core logic is clear and powerful: in a cycle where AI-driven earnings growth exceeds the historical average by more than twice, fighting the earnings trend is dangerous.
Jefferies' 2026 baseline forecast of 8,000 for the S&P 500 is based on earnings per share of $373, up 35% year over year and far above the market consensus of 29%, and a price-to-earnings multiple of 21.5 times.
From the perspective of inference system architecture, a single user instruction may trigger multiple steps such as planning, retrieval, file reading, code execution, tool calling, result checking, and regeneration.
GPUs and other AI accelerators handle model computation, while CPUs handle virtual machines, browsers, task scheduling, and tool execution; long context and concurrent sessions expand model state and KV cache management needs, while HBM, server DRAM, and enterprise SSDs handle different storage tiers based on performance, capacity, and cost.
NVIDIA's engineering materials explicitly discuss the impact of CPU execution efficiency on agent throughput and the need for tiered cache management across GPU memory, CPU memory, and storage.
From this it can be inferred that rising agent penetration is expected to push AI applications into industries worldwide with even stronger penetration trends, especially as incremental AI computing demand spreads from AI accelerators to high-performance CPUs, storage, and networking systems; the scale of demand ultimately depends on the combined effect of user numbers, task frequency, concurrency level, and efficiency improvements.
However, this market confrontation will ultimately be decided by cash flow.
AI applications expand revenue opportunities and improve production efficiency, which is expected to revise future corporate cash flows upward; rising long-term Treasury yields raise discount rates and project financing costs, reducing the present value corresponding to the same cash flows.
For AI data centers, a higher cost of capital also raises the utilization, pricing, and collection requirements needed for projects.
Therefore, strong demand linked to AI computing and AI applications and pressure on tech stock valuations can occur at the same time: the former determines the room for business growth, while the latter tests the price investors pay for that growth.
If the curve flattens further and is accompanied by tighter credit or even a yield inversion, the market's focus may also expand from valuation multiples to customer budgets, order delivery, cash collection, and a broader macro data flow measuring whether AI can drive sustained economic growth under an efficiency expansion trend, forming a stricter stress test for the AI super bull market.
From inflation trade to growth worries: as rate hike expectations rise, recession alarms approach, and the bond market seems increasingly close to sounding an economic warning
Last week, the extra yield investors demanded to hold 10-year U.S. Treasuries over 2-year notes briefly narrowed to just 17 basis points, the tightest gap since early 2025.
This so-called yield curve flattening raises the possibility that the 10-year Treasury yield will soon fall below shorter-term Treasury yields; this closely watched phenomenon is called a yield curve inversion, which is why the market has begun discussing curve inversion and its recession warning significance again.
Here it is necessary to distinguish between the absolute level of yields and the gap between long and short maturities: even if the 10-year yield remains at its highest since 2007, as long as the 2-year yield rises faster, the curve will still flatten.
Therefore, 17 basis points was the narrowest spread seen at one point last week, and the core change at this stage is that the market has broadly increased vigilance over excessive future tightening; it cannot yet be described as the curve having already inverted.
The rise in short-end yields first reflects repricing of the monetary policy path.
The Fed raised rates by 25 basis points on September 16, lifting the target range for the policy rate to 3.75%-4.00%, while emphasizing that economic growth is solid and inflation remains elevated; interest rate futures market pricing shows traders are preparing for at least three more 25-basis-point hikes over the next year.
Two-year Treasuries are especially sensitive to policy changes over the next few quarters, while 10-year and longer yields comprehensively reflect longer-term short-rate expectations and the term premium investors demand for bearing long-term interest rate risk.
Therefore, growth resilience, persistent inflation, and tightening policy expectations can together keep pushing yields higher, but the magnitude of increases across maturities may not be the same; moreover, market pricing does not mean the Fed has committed to the corresponding number of hikes.
Uncertainty over energy supply is still reinforcing this interest rate transmission chain.
On September 27, after rejecting Iran's ceasefire proposal, Trump said he expected the two sides to resume negotiations in the following week; Iranian Foreign Minister Araghchi insisted that reopening the Strait of Hormuz must be premised on his conditions being met, with related disputes involving arrangements such as lifting the blockade, sanctions, and frozen assets.
There is still room for negotiation, but normalization of shipping and energy supply has not yet become a certain outcome.
From a macro transmission mechanism perspective, persistently high energy costs may both push inflation higher and delay monetary easing, while also eroding household real purchasing power and corporate profits, leaving the bond market facing both higher near-term rates and pressure on future demand.
Historically, an inverted yield curve has provided a powerful signal: since the 1960s, all eight past U.S. recessions were preceded by a curve inversion, although this indicator failed in its prediction earlier this decade.
This is essentially bond investors expressing a judgment: the Fed is raising rates high enough to hinder economic development in order to contain inflation.
Such an outcome would have broad effects on financial markets, especially for stocks trading near record highs.
As shown in the chart above, the U.S. yield curve appears to be starting to flatten, possibly signaling a potential yield curve inversion; such changes have historically been uncommon and may signal an economic slowdown or even a new recession.
This month, the Fed delivered its first rate hike in three years and hinted at possible further increases, after which more and more investors began preparing for the above scenario.
This also highlights that, after a bond selloff reflecting rising price pressure amid strong growth, a hawkish Fed is changing the balance among various risks.
Zach Griffiths, head of investment-grade bonds and macro strategy at research firm CreditSights, said: "Seeing the 2-year and 10-year yield curve invert, or flatten significantly, would make people question the view that the economy is very strong, and that is part of what is priced into the bond market."
An inversion would reverse the process of global yield curve normalization since 2024.
Bond investors usually demand higher returns, meaning higher yields, to compensate for the greater uncertainty of locking up money for longer.
This means the yield curve usually slopes upward.
Just last month, the curve was moving in that direction, when long-term yields surged partly because of market concerns that the Fed's anti-inflation credibility under Chair Kevin Warsh was weakening.
But after the Fed's September rate hike, shorter-term bonds began to lead yields higher.
Traders are betting that the Fed's rate hikes over the next year will amount to at least three 25-basis-point moves.
Some do not expect the curve to invert soon, because the market has already priced in a considerable amount of tightening, making it harder for short-term rates to rise further relative to long-term yields.
Gennadiy Goldberg, head of U.S. rates strategy at TD Securities, said: "The market has already priced in significant Fed rate hikes, which has driven the yield curve to flatten noticeably in recent weeks. This leads us to believe that the 2-year and 10-year yield curve may steepen in the coming weeks."
At present, it is also hard to imagine a significant weakening of the economy.
According to the latest monthly survey by Bloomberg Intelligence, economists have just raised their forecasts for U.S. third-quarter economic growth due to stronger demand.
But others believe the curve-flattening trend still has room to continue.
Ed Al-Hussainy, a portfolio manager at Columbia Threadneedle Investments, said that as the Fed tightens policy to cool the economy and inflation, he is positioning for inversions in the 2-year versus 10-year and 5-year versus 30-year yield curves over the next six months.
He said: "The best sign that monetary policy is tightening is that the yield curve is flattening and ultimately inverting."
Entering this week, 2-year and 10-year Treasury yields were about 4.9% and 5.2%, respectively.
The 10-year Treasury yield, an important benchmark for the global bond market, is currently near its highest level since 2007.
A yield curve inversion often reflects concerns about the growth outlook, because the purpose of rate hikes is to address inflation by suppressing loan demand.
Slower economic growth may ultimately create conditions for the Fed to cut rates, pushing long-term yields down relative to short-term yields.
Statistics compiled by institutions show that since 1978, the 2-year and 10-year yield curve has inverted on average about 15 months before the start of a recession, with that interval ranging from six months to two years.
As shown in the chart above, multiple rounds of historical data show that a yield curve inversion often signals an approaching recession.
However, in recent years, the predictive power of the yield curve has come under increasing scrutiny.
In 2022, several U.S. yield curves inverted, and most economists expected the economy to fall into recession within 12 months.
But that did not happen, because the U.S. economy proved broadly able to withstand the Fed's tightening from 2022 to 2023, the regional banking crisis, the global trade war, and this year's surge in energy prices.
Although the 2-year versus 10-year yield curve is the indicator most often cited by bond investors, policymakers seeking recession signals also study other curves related to 3-month borrowing rates.
The yield curve between 3-month and 10-year U.S. Treasuries remains relatively steep.
The recent flattening of the 2-year versus 10-year Treasury curve has caused losses for bond investors who positioned for steepening early in the year.
This change is also spreading to the U.S. stock market, especially bank stocks.
Because banks typically fund themselves at short maturities and make longer-term loans, a narrowing spread between the two erodes net interest margins.
The KBW benchmark bank stock index, which tracks the performance of large bank stocks, entered technical correction territory last week, meaning a sharp decline of 10% from a recent high.
The move toward inversion reflects a possible complete shift in the economic outlook since the war between the United States and Iran broke out in February.
Before that, traders were betting on a series of rate cuts that would push short-term yields lower, rather than the rate hikes they are now preparing for.
Jamie Patton, co-head of global rates at TCW Group, said an inversion "would be a sign that the Fed is making a policy mistake."
She said: "It is raising rates too much and will have to cut them sharply in the future. Therefore, for us, a yield curve inversion is not a healthy signal for the macroeconomy."