The Zhitong Finance App learned that CICC released a research report saying that there is no need or basis for the Federal Reserve to raise interest rates continuously and drastically. There are the following reasons: First, traditional demand does not support drastic interest rate hikes. Second, when inflation is driven by a high base, it will fall quickly first. CICC roughly estimates that the US CPI will return to 2% in May 2027, which is significantly lower than the current 3.7%. Third, the return on investment and financing pressure on AI itself has also wavered. The first half of next year is a key verification point. Currently, AI has been slow to lack new narratives and catalysts to open the “ceiling” of application scenarios, which is the key to the continuous rise in weak links such as CoreWeave's credit risk. This is the same lack of progress in AI at the end of last year, and the market is just as worried about private equity credit risks.
At the October FOMC, which ended early this morning Beijing time, the Federal Reserve stood still as scheduled, keeping the benchmark interest rate at 3.75 to 4%. The incremental information released at this meeting was very limited. Federal Reserve Chairman Walsh still “spared words like gold” and once again expressed his dislike of forward-looking guidance. In addition, not raising interest rates at this meeting was entirely within market expectations, because market changes were also very limited. After the meeting, interest rates on US bonds remained basically unchanged. The 10-year interest rate rose slightly to more than 5%, US stocks also remained volatile, and the implied probability of a December rate hike in CME interest rates remained around 85%.
Although expectations of interest rate hikes have been repeated several times since the last rate hike in September, and the CPI and PPI exceeded expectations during this period, the overall policy direction was also in line with the CCC's previous judgment, that is, the September rate hike was necessary because it was necessary to maintain its credibility (this was basically achieved, and the term premium went back to 60 bps before concerns escalated in July), but there is no need for continuous and drastic interest rate hikes (“Is it enough for the Fed to raise interest rates once?”).
After the interest rate hike in September, the market environment has quietly changed: rising financing costs are already squeezing traditional demand such as real estate. Weak links in AI, such as NeoCloud's CDS and junk bonds representing CoreWeave, are also rising rapidly, all reflecting the tightening of financial conditions. As a result, the demand for AI and the prospects for sustainability in capital development have cast a shadow over it. Interest rates on US bonds seem to be peaking, and oil prices are still high. So is the Fed's interest rate hike in December certain? How many more additions will the Federal Reserve make in the future?
Chart: Interest rate hikes in October are expected to rise above 70% by the end of September, but have dropped to less than 10% before the meeting
Source: Bloomberg, CICC Research Division
October FOMC core information: Suspend interest rate hikes and see data later; strong economy, but tightening financial conditions; subsequent reduction in the number of interest rate negotiations
The incremental information provided at this conference was limited and mainly conveyed the following signals:
1) Maintaining the benchmark interest rate at 3.75% to 4%, in line with market expectations; 2) the economy is still strong (strong); 3) financial conditions have begun to tighten, which is reflected in some credit spreads, especially high-yield bonds; 4) Future decisions depend on data. There were 2 non-agricultural and 1 inflation before the FOMC in December. According to the minutes of the September meeting, participants generally emphasized that inflation is still high, and labor market risks have abated, so most officials stated that interest rates may still need to be raised once before the end of the year; 5) Walsh once again emphasized that they do not like forward-looking guidelines, and that the number of interest rate discussions will be gradually reduced in 2027 and 2028.
Future policy path: There is no need for continuous and drastic interest rate hikes. The first half of next year is a key verification point; market expectations are basically unreliable
The Federal Reserve's suspension of interest rate hikes this time was in line with market expectations, and is also in line with CICC's previous judgment. Even before the September rate hike was implemented, CICC suggested that the Federal Reserve needed to raise interest rates to restore the market's trust in it, but it also emphasized that US fundamentals did not support continuous and drastic interest rate hikes “What would happen if interest rates were raised?” Now that the purpose of restoring trust has been basically achieved for the time being (the 10-year US bond maturity premium has been reduced to the level expected by the market before Walsh's speech at the July interest rate meeting), non-agricultural agriculture is weakening, AI concerns have increased, financial conditions have been partially tightened, and supply constraints have not yet been greatly strained, then why should the Federal Reserve take a look, let alone that the midterm elections are just around the corner.
Looking ahead, CICC still believes that there is no need or basis for the Federal Reserve to raise interest rates continuously and drastically. There are the following reasons:
First, traditional demand does not support drastic interest rate hikes. The US economy is still in K-type differentiation with “strong technology but weak traditional demand”, and current interest rates have begun to clearly squeeze traditional demand. Taking real estate as an example, the 30-year mortgage interest rate has risen to a multi-year high of 7.5%, which is clearly higher than the rental income of 6%. Existing home sales have weakened again since May; interest rates on short-term loans for small businesses have also risen to 8.2%.
Chart: The US economy is still in K-type differentiation with “strong technology but weak traditional demand”
Source: Haver, CICC Research Division
Chart: Existing home sales continue to weaken under pressure from higher US mortgage interest rates
Source: Haver, CICC Research Division
Chart: Mortgage interest rates and short-term loan interest rates for small businesses in the US have risen rapidly in 30 years
Source: Bloomberg, CICC Research Division
However, the strength of AI has hedged and overshadowed the pressure of traditional demand in terms of total volume. However, when AI begins to need to draw resources from the whole society, it will become more obvious that traditional demand is crowding out physical and financial factor resources. Previously, positive externalities brought about by AI demand supported US economic growth and provided “motivation” for the Federal Reserve to raise interest rates, but as AI began to require more external capital to support investment, the crowding out effect was also unavoidable. It may gradually change from a “friend” of traditional demand to an “enemy” and “AI from a “friend” to a “enemy” of traditional needs?” Typical examples are AI issuing large amounts of debt and the need for the Federal Reserve to raise interest rates. Once AI's return on investment and positive externalities weaken, the combined effects cannot be supported, making it difficult for the Federal Reserve to continue to raise interest rates.
Chart: In the second quarter, US real GDP grew 2.2% month-on-month, and AI-related demand contributed 28% to growth
Source: Haver, CICC Research Division
Second, when inflation is driven by a high base, it will fall quickly first. The high base from the first half of this year until May due to the situation in Iran will cause the overall inflation in the US to fall rapidly. Based on an assumption of maintaining the current oil price (around 95 US dollars/barrel), China Gold roughly estimates that the US CPI will return to 2% in May 2027, which is significantly lower than the current 3.7%.
Chart: CICC estimates that a high base will drive CPI down to around 2% year over year in the first half of 2027
Source: Haver, CICC Research Division
It is true that geography and oil prices are short-term disturbances and affect the path; however, if you look at it longer, they do not affect the trend. If the short-term geological situation escalates again and pushes up oil prices, it will indeed increase the pressure for CPI to rise above expectations and raise interest rates. This is also the biggest variable in the short-term path. However, looking at it over time, the geographical conflict will not necessarily cause crude oil supply to continue to shrink. The reason is that even if the Strait of Hormuz is not open, various alternative transportation routes (such as pipelines, ship-to-ship, etc.) have basically returned the Gulf's exports other than Iran to pre-conflict levels. The same was true during the Russian-Ukrainian conflict in 2022. Although the geographical situation has not completely eased, the supply scale of Russian crude oil has gradually recovered to pre-conflict levels, and oil prices have gradually declined. Therefore, there are short-term path disturbances, and the medium term may not affect the trend.
Chart: Crude oil exports from Gulf countries other than Iran have also returned to pre-conflict levels, according to Kpler's data
Source: Kpler, CICC Research Division
Chart: The gradual release of crude oil supply in the medium to long term will depress oil prices
Source: Haver, CICC Research Division
Third, the return on investment and financing pressure on AI itself has also wavered. The first half of next year is a key verification point. Currently, AI has been slow to lack new narratives and catalysts to open the “ceiling” of application scenarios, which is the key to the continuous rise in weak links such as CoreWeave's credit risk. This is the same lack of progress in AI at the end of last year, and the market is just as worried about private equity credit risks.
The first half of next year is a critical juncture, just as Anthropic's performance in the first quarter of this year surpassed expectations to prove to the market that it has gone through a closed loop in the programming business. Otherwise, if there is a delay and the market runs out of patience, the outlook for capital expenditure and returns in 2028 will be further bleak. As the main driving force for growth and profits, if we don't have strong confidence in AI, then how can we expect the Federal Reserve to continue to raise interest rates?
In summary, the beginning of pressure on traditional demand and the rapid decline in inflation in the first half of the year did not seem to support the Federal Reserve's continuous and sharp interest rate hikes. What's more, if demand for AI slows beyond expectations, the Federal Reserve may even need to cut interest rates. The first half of next year is a key verification point for AI and policy paths. The upward risk of CICC's view is that oil prices continue to exceed expectations and AI progress has greatly exceeded expectations. As for market expectations, judging from historical experience, almost all of them are unreliable.
Chart: Judging from historical experience, there is a deviation between the implied interest rate path of interest rate futures and the final actual path
Source: Bloomberg, CICC Research Division
How high interest rates can the US afford? 6% could be a watershed
On the one hand, AI is strong, and on the other hand, traditional demand is gradually being pressured and squeezed out. So when might the positive externalities of AI not be able to offset the weakness of traditional demand?
CICC in “From the “Friend” to the “Enemy” of Traditional Needs? According to estimates, this watershed may be about 6% of 10-year US debt (AI can withstand 7% or even very high; traditional demand tolerance is around 4-4.5%). This also shows why the recent rise in interest rates on US bonds has not completely hurt overall growth, yet the erosion of some weak points has begun to occur.
If it exceeds this inflection point, the positive externalities of AI cannot continue to hedge against negative externalities, causing overall growth to begin to be pressured, then the Federal Reserve may even need to start cutting interest rates.
How will the market trade? Short-term interest rate hikes are still in line with the “land and cash out” rule
The recent peak and fall in US bond interest rates is basically in line with the rule that interest rate hikes are implemented and expectations are gradually realized. In September, CICC “Is it enough for the Federal Reserve to raise interest rates once?” As a reminder, the implementation of interest rate hikes is not bad for the market; on the contrary, it may mean “running out of profit.” A typical example is the “preventative interest rate hike” in 1997. Interest rates on US bonds peaked and fell soon after the policy was implemented; in turn, the “preventive interest rate cuts” in 2019, 2024, and 2025 also staged a similar “script”. Interest rates on US bonds gradually rebounded the moment or soon after interest rate cuts were implemented (“What Happens If Interest Rates Are Raised?”). Compared to this time, although the October inflation data caused expectations of interest rate hikes to “rebound” and interest rates on US bonds rose again, if a brief rate hike is the basic assumption, it still conforms to this rule. According to the bitmap, it is estimated that the benchmark interest rate of 4 to 4.25% plus the term premium of 70 to 90 bp also corresponds to about 5.3% of 10-year US bonds. Therefore, if CICC's judgment on the Fed's future interest rate hike path is accurate, then the moment when financial conditions are tightest is gradually approaching.
Chart: Interest rate expectations of 4 to 4.25% plus a term premium of 70 to 90 bp, the top corresponding interest rate on US bonds is around 5.3%
Source: Bloomberg, CICC Research Division
US stocks, on the other hand, reflect the AI narrative and the emotional turmoil of the midterm elections. In the June mid-year outlook, CICC further raised the S&P 500 point level to 7,800 to 7,800. The market became entangled and volatile after reaching this point. In addition to liquidity disturbances brought about by rising interest rates, the recent rise in AI financing risks and the slowdown in revenue redemption have all suppressed the overall AI narrative, so the market is also “shrinking” to cloud vendors with low valuations and expectations that don't rely entirely on future high growth assumptions. Since US stock profits are still steady, and valuations have shrunk, CICC is still not pessimistic about the outlook for US stocks.
Chart: The AI pressure index built by CICC has also continued to rise since the end of September
Source: Bloomberg, FactSet, CICC Research Division
The US dollar strengthened as interest rates on US bonds fell and expectations of interest rate hikes eased, partly reflecting Europe's weaker “backing” and a further decline in capital. Although interest rates on US bonds have begun to fall, the US dollar has not weakened at the same time. The main reason is that Europe's political situation and finance are facing greater variables, so some funds have fallen back again, but the difference between choosing US bonds or US stocks.
Chart: The US dollar benchmark situation fluctuates, and the Fed's interest rate hike has revived trust and still supports the US dollar in the short term
Source: Bloomberg, CICC Research Division
Gold is currently more similar to “looking at long options”. It has three levels of meaning: first, you need to pay a certain “option fee,” or time cost; second, downward pressure is manageable, and the biggest risk is continuous and drastic interest rate hikes; third, options implementation requires “exercise conditions” or narratives. For example, at the end of last year, market concerns about AI investment returns, US policy stability, and the reputation of the Federal Reserve have jointly catalyzed a sharp rise in gold.
Chart: More space for gold comes from grand narratives
Source: Polymarket, CICC Research Division
Looking at interest rate hikes for the next year, which includes various types of assets, interest rate futures (3 times) > gold (2.4 times) > copper (1.9 times) > Dow Jones (1.7 times) = US bonds (1.7 times) > S&P 500 (1.1 times) > bitmap (1 time) > NASDAQ (0.4 times).
Chart: Judging from the interest rate hike expectations for the next year, including various types of assets, interest rate futures > gold > copper > Dow Jones = US bonds > S&P 500 > bitmap > NASDAQ
Source: Bloomberg, CICC Research Division