CICC: No Basis for Consecutive Sharp Fed Rate Hikes; AI Returns Face Validation in H1 2025
nashnova research
CICC argues the Fed lacks grounds for consecutive sharp rate hikes, marking H1 next year as the critical window to verify AI investment returns and the rate path — traditional demand is already buckling under high rates, inflation should fall on base effects, and AI's own financing stress is shaking the market narrative.
What did the Fed actually say this time?
The October FOMC held rates steady at 3.75%–4%, with limited new information.
Fed Chair Waller reiterated he dislikes forward guidance — no pre-commitment to the next move.
Post-meeting, the 10-year Treasury yield edged above 5%; equities stayed choppy. CME rate futures still price an 85% probability of a December hike.
Why does CICC think consecutive sharp hikes won't happen?
First, traditional demand can't take it. The 30-year mortgage rate has hit 7.5%, well above the 6% rental yield (rental yield = annual rent ÷ property price — a gauge of whether buying makes financial sense) → existing-home sales have weakened again since May; small-business short-term loan rates have also climbed to 8.2%.
In plain terms = rates are already too high for ordinary homebuyers and small businesses; more aggressive hikes would crush the traditional economy.
Second, inflation should drop fast on a high base. CICC estimates that if oil stays near $95/barrel, US CPI will fall to 2% by May 2027, far below the current 3.7%.
Third, AI investment returns are wobbling. AI has yet to open a convincing new application "ceiling." CoreWeave's credit risk keeps rising — This means → market confidence in AI's "burn-cash-for-growth" path is loosening, echoing late last year's private-equity credit scare.
How is AI turning from the economy's "friend" into its "enemy"?
AI's positive spillovers (positive externality — AI growth that lifts the broader economy along the way) previously propped up US GDP and gave the Fed room to hike. Q2 data: US real GDP grew 2.2% annualized, with AI-related demand contributing 28% of that growth.
But as AI needs more external funding, crowding-out (crowding-out = AI financing drains capital from other sectors) is kicking in.
This means → AI has gone from "donating blood" to the traditional economy to "competing for the same blood supply." CICC estimates the 10-year yield around 6% is the tipping point — AI can tolerate 7% or higher, but traditional demand tops out at roughly 4%–4.5%. Beyond that line, AI's positives can no longer offset the negatives, and the Fed may even need to cut rates.
What needs to be proven in H1 next year?
CICC flags H1 next year as the key checkpoint, drawing an analogy to Q1 this year when Anthropic's results blew past expectations and proved the coding-business loop works.
In plain terms = the market needs to see AI actually making money in the first half; otherwise the "burn cash, sell the story" narrative collapses and the funding chain tightens.
This reflects that the biggest uncertainty in AI investment today is not the technology itself but whether commercial returns can keep pace with the cash burn.
What does this mean for major asset classes?
Treasuries: Assuming one more hike per the dot plot, the fed funds rate reaches 4%–4.25%. Add 70–90 bps of term premium (term premium — the extra yield investors demand for holding long-dated bonds), and the 10-year yield tops out around 5.3%.
US equities: CICC raised its mid-year S&P 500 target to 7,800–8,000. The index is stuck oscillating in that range as rising AI financing risk and slowing revenue recognition weigh on the narrative; the market is "narrowing" into lower-valued cloud names. CICC remains not bearish — earnings are still solid.
Gold: CICC frames it as a "long call option" — downside is manageable; the biggest risk is precisely consecutive sharp hikes. Upside needs a narrative catalyst — a resonance of concerns over AI returns, policy stability, and Fed credibility.
Where do rate-hike expectations diverge across assets?
Implied hikes over the next year vary wildly: rate futures 3× > gold 2.4× > copper 1.9× > Dow 1.7× > Treasuries 1.7× > S&P 500 1.1× > dot plot 1× > Nasdaq 0.4×.
This means → rate futures are pricing the most hawkish outcome; the Nasdaq is pricing the most dovish — tech investors are betting the hikes won't materialize.
CICC's bottom line: traditional demand under pressure + inflation falling on a high base both argue against consecutive sharp hikes. If AI demand slows more than expected, rate cuts could arrive ahead of schedule.
市场有风险，内容仅供研究参考，不构成投资建议。