Observe the numbers. August core CPI, year-over-year, came in at 2.4%. That is not the 2.0% the market had priced in. That is not the "disinflation trend" the bulls have been selling for six months. That is a fault line.
I have spent the better part of three decades auditing mechanisms—smart contracts, token economies, stabilization protocols. The first rule of any audit is simple: when the output deviates from the expected output, you do not adjust the expectation. You trace the input. You find the variable that moved.
The variable here is energy. And services. And something else—something the CICC report flagged that most retail commentary will miss entirely: AI-related price pressure. Do not glaze over that. It is a new input in the system. And new inputs change the output.
Context: The System State Before the Shock
Let me establish the baseline state of the machine before we dissect the failure mode.
The Federal Reserve has been running a tightening cycle for over two years. The target range sits at 3.50% to 3.75% as of the last meeting. The market narrative for the past three months has been that the cycle is done. The "last hike" was the phrase du jour. Bond traders have been positioning for a cut cycle starting in the first half of next year. The equity market, particularly the tech-heavy indices, has been rallying on this assumption.
Here is what the August data actually shows:
- Headline CPI, month-over-month: +0.4%
- Core CPI, month-over-month: +0.3%
- Core CPI, year-over-year: +2.4%
That 2.4% figure is the headline the market will obsess over. But in my line of work, we look at the month-over-month prints. That is where the momentum lives. The annual figure is just a lagging output of twelve prior months of data. The 0.3% core MoM print is the accelerometer. And it is not decelerating.
CICC's call, based on this data: the Fed hikes 25 basis points at the September 16 meeting. Target range moves to 3.75% to 4.00%.
I have read their research note. The logic is sound. Let me walk you through the mechanism, because the why matters more than the what.
Core: The Mechanism Autopsy
Component One: Energy as the External Shock
Energy prices have moved from being a passive drag on headline inflation to an active contributor. Crude oil has rebounded into the $80-$85 per barrel range. OPEC+ supply cuts are holding. Geopolitical risk premium is creeping back in.
Here is the transmission chain: energy feeds into headline CPI directly through gasoline prices and indirectly through airfares, shipping costs, and industrial inputs. The August headline print of +0.4% MoM is primarily an energy story. But energy is a lagging indicator in the CPI basket—it takes two to four weeks for oil price movements to show up in the published data. That means the September and October prints will still be carrying this energy load.
Silence in the data is the loudest warning sign. The energy component is not silent. It is screaming. And the market is choosing not to hear it.
Component Two: Service Price Stickiness
Now we get to the structural part—the component that will outlast any short-term energy fluctuation.
Services inflation remains stubbornly persistent. The two big sub-components:
Shelter costs: Owner's Equivalent Rent (OER) is still running at +0.3% to +0.4% month-over-month. This is the heaviest weight in the CPI basket. And here is what most retail investors do not understand: shelter inflation lags housing prices by 12 to 18 months. The moderation in home prices that began in 2023 will eventually feed through to CPI. But "eventually" means late 2024 to early 2025, not this quarter.
Super-core services: This is the Fed's preferred gauge of underlying inflation momentum—core services excluding housing. It has been running at an annualized pace of 4% to 5%. That is not consistent with a 2% inflation target. That is consistent with an economy that still has excess demand.
Complexity is often a veil for incompetence. The Fed's reliance on "transitory" narratives in 2021 was exactly this—a complex story that obscured a simple truth: too much money chasing too few goods. The current "soft landing" narrative risks the same failure mode.
Component Three: The AI Price Signal
This is the component that most market commentary will miss. CICC flagged it, and they deserve credit for that.
AI-related price pressure is emerging. Think through the mechanism:
- Data center construction is surging. That means concrete, steel, and specialized cooling infrastructure—all of which feed into producer prices.
- Electricity demand from AI data centers is straining regional grids. Utilities are filing for rate increases to fund grid upgrades.
- The cost of GPU compute is being passed through to cloud service prices. And software vendors are adding AI features at premium price points.
This is a new variable in the inflation model. It was not present in the 2022 inflation surge. And it is structurally sticky—AI infrastructure buildout is not going to pause because the Fed raises rates by 25 basis points.
From my experience auditing token economies: when a new variable enters the system, you do not assume it is exogenous noise. You model it as an endogenous factor until proven otherwise. The AI capex supercycle is a multi-year demand driver. It will keep upward pressure on certain CPI components regardless of the Fed's policy stance.
Component Four: The Labor Market Connection
The Fed also has to reconcile its dual mandate. And the labor market is sending mixed signals.
- Unemployment rate: approximately 3.8% to 4.0%—historically low, still at full employment.
- Wage growth: 3.5% to 4.0% year-over-year—cooling from the 5%+ peaks of 2022, but still above the level consistent with 2% inflation.
CICC's prediction is that the Fed will lower its unemployment forecast in the upcoming SEP (Summary of Economic Projections) while raising its inflation forecast. That is a policy-relevant shift. It signals that the Fed sees the economy as still running hot enough to require restrictive policy for longer.
The Contrarian Angle: What the Bulls Got Right
I have spent this entire piece dissecting the hawkish case. Intellectual honesty requires the flip side.
The bulls were not entirely wrong. Let me stress-test my own bearish bias.
First, the base effect. Last year's September and October prints were high. That means even if month-over-month inflation continues at 0.2% to 0.3%, the year-over-year figures will mechanically decline in Q4 2025. This is not a forecast—it is arithmetic. The annual comparison gets easier precisely because the prior-year base is elevated.
Second, the labor market is indeed cooling. The unemployment rate has drifted up from cycle lows. Job openings are declining. The Beveridge Curve is shifting inward—which historically signals that the economy can achieve lower inflation without a sharp rise in unemployment. The "soft landing" scenario is not impossible. It has historical precedent (1994-1995).
Third, the policy transmission lag. The Fed's tightening operates with a lag of 12 to 18 months. The full effect of the rate hikes from 2023-2024 has not yet fully propagated through the economy. There is an argument that the Fed has already done enough, and the data just needs time to catch up. CICC's call for one more hike may be the last one in the cycle—a "confirmation hike" rather than the start of a new tightening phase.
Fourth, the market was correct that the peak policy rate is near. The question is not whether the Fed stops hiking. The question is how long it stays at the peak. The bulls may be wrong about the timing of cuts, but they are likely right that the terminal rate is close to being reached.
The Fed's Decision Tree: What to Watch
Let me lay out the September 16 FOMC meeting as a decision tree, because that is how the market will trade it:
Scenario A (CICC's Base Case, ~50% probability): Hike 25bp to 3.75%-4.00%. Updated dot plot shows higher median rate for 2026 and 2027. SEP shows lower unemployment forecast, higher inflation forecast. This is a hawkish hike—the market will initially sell off on the "higher for longer" message.
Scenario B (~25% probability): Hold rates. But the dot plot shifts dramatically higher, signaling multiple hikes in Q4 2025/Q1 2026. This is arguably more hawkish than a hike, because it shifts the entire rate path higher without the "relief" of a hike being "done."
Scenario C (~20% probability): Hike 25bp, but signal that this is the terminal rate. Dovish guidance. The market rallies on "end of cycle" narrative.
Scenario D (~5% probability): Hike 50bp or unconventional action. Extremely hawkish. Only triggered by a catastrophic inflation surprise.
My base case aligns with CICC: Scenario A. The Fed will hike, and the hike will be accompanied by hawkish projections. But the market will eventually digest this and realize that the marginal policy shift is smaller than the cumulative shift.
The Market Impact: Who Bleeds, Who Benefits
Let me trace the consequences through the major asset classes, because narrative matters less than allocation.
US Treasuries: The 2-year yield will move higher on a hawkish hike. The 10-year is more ambiguous—it is driven by growth expectations, not just policy expectations. But a higher rate path extension will push the 10-year toward 4.5% to 5.0%. The curve remains inverted, but the inversion will flatten as short rates rise and long rates stay anchored.
Equities: The initial reaction will be negative. Higher discount rates compress multiples. But the composition matters. High-duration, high-multiple tech names are the most vulnerable. Value stocks with real cash flows and dividends will be more resilient. The "AI trade" is particularly exposed—if the AI narrative is partially an "ease of financial conditions" narrative, then a hawkish Fed removes the fuel from that fire.
Dollar: The dollar index (DXY) is likely to strengthen on a hawkish surprise. The yield differential between the US and other developed markets (Europe, Japan) is already wide. A further widening pushes the dollar higher.
Gold: The real yield channel dominates—higher real rates are negative for gold. But there is a counter-force: if the market interprets the Fed's hawkishness as a precursor to a policy error (tightening too much), then the risk premium on fiat assets rises, which is positive for gold. Net: range-bound, with support near $1,850 and resistance near $1,950.
Emerging Markets: This is the most exposed asset class. A stronger dollar + higher US rates = tighter global financial conditions. Countries with dollar-denominated debt and current account deficits get squeezed. The ex-China emerging market complex (Indonesia, Philippines, Mexico, Brazil) will face capital outflow pressures.
Crypto (since this is a Blockchain news article): The digital asset market has become increasingly correlated with US liquidity expectations. A hawkish Fed pushes this timeline longer. The "risk-on" bid for crypto that we have seen during the past few months of dovish expectations is likely to be deferred. But note: higher duration assets like Bitcoin are going to be more sensitive to the marginal direction of liquidity, not the level. The Fed is still tightening, but the pace of tightening is slowing. That is the binding constraint.
The Hidden Variable: Fiscal Dominance
I cannot write a thorough macro analysis without flagging the elephant in the room.
The US federal deficit is running at approximately 6% of GDP. That is not a sustainable level in a non-recessionary economy. And fiscal policy—Congress, not the Fed—has been running expansionary while the Fed tries to run tighter. This is the "fiscal dominance" problem.
The mechanism: the Treasury needs to issue more debt to finance the deficit. More debt issuance = higher term premium = higher long-term yields. This does the Fed's work for it on the long end, but it also creates a paradox: the Fed's tightening is being partially offset by fiscal stimulus.
This is why the Fed's forecast changes are so important. If the Fed lowers its unemployment forecast, it is implicitly acknowledging that fiscal policy is providing more support than previously estimated. That means the Fed needs to keep rates higher for longer to offset fiscal stimulus.
Trust is a variable, verification is a constant. The Fed's credibility is on the line. If it signals a pause now and inflation reaccelerates in Q4, it has to reverse course—which would be the third policy error in five years (2021 transitory, 2023 pause, 2025 premature cut). The Fed has institutional incentives to err on the side of hawkishness. That is why I lean toward CICC's call.
The Takeaway: The Question That Matters
The September CPI print will not tell us anything. The October print will not tell us anything. The November print will start to matter. The January 2026 print will matter a lot.
Here is the question that matters more than the immediate rate decision: What is the path of the federal funds rate for the next 24 months?
CICC is signaling that the path is higher than the market expects. The market is pricing a peak rate of 3.75%-4.00% and then cutting back toward 3.00%-3.50% by late 2026. CICC's analysis suggests the peak might be 4.00%-4.25% and the "resting rate" in 2027-2028 might be 3.50% or higher.
That is the output that matters. The September meeting is just the input.
The market will trade the headline. The professionals will trade the dots. And the dots are pointing higher.
Disclosure: This analysis is based on publicly available data and the CICC research note as described. I hold no direct positions in the assets discussed and have no conflict of interest to declare. This is not investment advice. Verify everything.
