October hike odds top 72% as Cook names AI capex a near-term inflation driver; the long end grinds out another multi-decade high

0. Weekly Arc

The cumulative direction over the past week is unchanged hawkish accumulation, now hardening from pricing into a structural story. It runs from the Sept 16 unanimous hike to 3.75%–4.00% with a hawkish dot plot and no forward guidance, through the Sept 23–25 cluster in which Barkin, Collins, Paulson, Hammack and Williams converted expectation into a priced cycle, then the Sept 26–28 stretch in which the 30-year passed 5.5%, the 2s10s compressed toward inversion and October odds consolidated near two-thirds. Today’s marginal move is a broadening of the inflation justification onto AI capital expenditure while the long end sets fresh highs.

1. Policy Narrative & Expectations

No new policy action in the past ~24h — the target range stands at 3.75%–4.00% after the unanimous 12-0 September hike ^[1][2][3] — so the marginal information sits in three places. First, the sole substantive official remarks of the session came from Governor Lisa Cook, who located near-term inflation pressure in AI-driven demand and energy/labour competition and explicitly warned that lowering rates could fuel inflation ^[4][5][6]. Second, the price of the cycle rose again rather than consolidating: October hike odds now read above 72% on CME ^[1], 72.5% on a separate tracker ^[7], 70.3% ^[8] and roughly 70% ^[9][10], and LSEG-recorded money-market pricing now implies a cumulative 100bp of hikes over the next 12 months ^[11], against the roughly 88bp priced in yesterday’s batch. Third, the analytical distribution around the number of hikes remains wide and has partly inverted versus the market: Deutsche Bank keeps a base case of about 50bp of further hikes while conceding 100bp is fully possible and identifies only two developments — renewed labour-market tightening and a reversal of rent disinflation — that would make the Fed materially more hawkish in 2027 ^[12]; Morgan Stanley expects two further hikes and holds the upper bound at 4.125% in Q4 2026 and 4.375% across 2027 ^[13]; JPMorgan conditions an October move on a strong labour report plus a 0.2% monthly core PCE print ^[14]; Citi argues the reaction function has become more sensitive to spot inflation and energy prices, leaving the October meeting hinging on a single September core CPI print ^[15][16]; and Goldman Sachs argues the actual number of hikes may be two to three rather than the seven to eight implied by far-forward pricing ^[17].

1.1 FOMC Officials’ Remarks

• [NEW] Hawkish — Governor Lisa Cook: per Gelonghui, she said bluntly that “trillions of dollars of AI infrastructure spending has intensified competition for energy and labor, and in the short term has instead become an inflation driver, which could support the Fed in maintaining its tightening path” ^[4]. Per one-line relays of the same speech she said she sees “some economy-wide pressure coming from AI-fueled demand” ^[5], that she is highly attentive to AI potentially causing a temporary rise in unemployment, that the Fed has limited tools in that case, and that “lowering rates could fuel inflation” ^[6][18] — single source / unverified social relays. JPMorgan’s read is that she pointed to an AI-driven rise in the neutral rate while maintaining hawkish comments ^[19], and Deutsche Bank groups her among officials citing resilient consumer spending and a surge in AI capex as demand-side inflation drivers ^[9]. On the labour market she said it looks able to withstand further rate hikes, with unemployment falling and conditions “roughly balanced and gradually improving” ^[20] (per Yicai). On forward policy she said the number and magnitude of any future adjustments will be informed by inflation and labour-market data and by how the economy has reacted so far, and that she will consider the policy rate needed to keep guiding inflation back to target ^[21][22] — i.e. explicitly data-dependent rather than path-committed.
• [ONGOING] Chair Kevin Warsh (listed separately): the operative framework is unchanged — no forward guidance, the hike framed as “removing some of the accommodation”, a denial that the market coerced the committee, growth seen as strengthening, employment risks broadly balanced and inflation risks skewed to the upside, with the “underlying inflation clearly and fast enough falling back to target” standard judged unmet ^[23].

1.2 Policy Signals & Institutional Communication

• [ESCALATED] October hike pricing broke above 70% across venues: CME FedWatch shows more than 72% for October ^[1], a separate reading 72.5% ^[7], another 70.3% ^[8], with JPMorgan, Deutsche Bank and Jin10 all recording roughly 70% ^[14][9][10]. LSEG-recorded money-market pricing now implies a cumulative 100bp of hikes over the next 12 months ^[11], up from the roughly 88bp in yesterday’s batch, and Deutsche Bank notes the market’s implied peak fed funds rate is above 4.80%, equivalent to nearly 100bp of additional tightening ^[12].
• [NEW] Deutsche Bank’s hawkish-2027 test: its base case remains about 50bp of further hikes (December and March) with 100bp “fully possible”, achievable by a slightly higher neutral rate than the latest SEP assumes without a fuller reassessment; it names two developments that would force a hawkish reassessment — renewed labour-market tightening and a reversal of rent disinflation ^[12][9]. Its published forecasts put core PCE at 3.3% in Q4 2026 falling to 2.4% in Q4 2027 and the funds rate at 4.13% in Q4 2026 and 4.38% in Q4 2027 ^[9].
• [ONGOING] Morgan Stanley keeps two further hikes and holds the funds upper bound at 4.125% in Q4 2026 and 4.375% across 2027, with the 10-year forecast at 4.80% at end-2026 and 4.70% at end-2027 ^[13][24].
• [NEW] JPMorgan’s explicit October trigger: a strong labour-market report combined with firmer September inflation pointing to another 0.2% monthly core PCE print would be enough to trigger an October hike, with the market pricing close to a two-thirds probability ^[14].
• [ESCALATED] Citi reduces the October decision to one variable: the reaction function has become more sensitive to spot inflation and energy prices, so the October FOMC outcome will depend mainly on the September core CPI reading; it expects payrolls of about 85k with unemployment edging up to 4.2%, and says only negative payrolls or an unemployment rate of 4.3% would materially move pricing ^[15][16].
• [NEW] Goldman Sachs argues the market is over-pricing the cycle: it sees limited core inflation pressure and expects two to three actual hikes rather than the seven to eight implied, creating an opportunity to hedge at the far end of the curve ^[17].
• [NEW] Barclays expects the Fed to restart reserve-management purchases: it flags potential Q4 funding pressure from TGA rebuilding, heavy T-bill supply and seasonal GSIB balance-sheet contraction, expecting RMPs to resume at about $10bn in October and $20bn in November and to hold that pace thereafter ^[25][26].
• [NEW] Flow-of-funds shift toward duration: fixed-income fund inflows moderated to $8.3bn in the five days to Sept 24, below the four-week average, but short-term government fund demand cooled sharply to the 26th percentile while intermediate and long-term government fund demand rose to the 78th and 69th percentiles — investors “moving out the curve” ^[26][25]. Bank securities holdings fell about $13bn in the week through Sept 16, but Barclays attributes this mainly to valuation losses rather than selling, with year-to-date bank securities portfolios still up about $174bn including about $110bn of Treasuries, and foreign official Fed custody holdings down about $11.4bn in Treasuries in the week of Sept 23 ^[26][25].
• [NEW] Institutional governance: the WSJ reports a Federal Reserve employee transferred potentially sensitive files outside the central bank just before retiring, and that the Fed told its internal watchdog it will improve security after a finding that a researcher extracted potentially confidential information in 2024 — an institutional-risk item with no rate-path content ^[27].
• [NEW] Speaker slate: the WSJ week-ahead lists Fed Governors Michael Barr, Christopher Waller and Michelle Bowman and New York Fed President John Williams as speakers ^[28][29], with Goolsbee, Musalem and Williams also scheduled on the following session ^[30].
• [ONGOING] The delivered decision, statement wording and projections stand: a unanimous 12-0 hike to 3.75%–4.00%, the first since July 2023 and less than seven weeks before the November midterms, with the statement adding that the move supports a “timelier return” to 2%, describing domestic spending as resilient, softening capex language from “strong” to “robust”, and dropping the reference to energy-sector supply shocks; the SEP raised 2026 growth to 2.3%, PCE to 3.7% and core PCE to 3.4%; the dot plot implies one more hike this year and no change next year; and the Chair again did not submit a dot ^{[23][3][1][31]}.

2. Key Data & Market Read

• [NEW] Dallas Fed manufacturing beat: the index came in above expectations, which JPMorgan reads as confirming US activity remains solid and as supportive of both the dollar and Treasury yields ^[19].
• [NEW] September composite PMI read through: the September composite PMI rose from August’s level to show activity still fairly strong, described as a direct near-term stimulus pushing yields higher and as evidence the economy is not as moderate as previously assumed ^[32]; JPMorgan notes the developed-market flash PMI rose to a four-year high, adding to the risk of further central-bank tightening ^[14].
• [NEW] Texas manufacturing described as “hot and expensive”, with output and orders strengthening and price pressures rising in tandem ^[33].
• [NEW] Claims and capex: weekly unemployment insurance claims remain near 57-year lows, supporting the view the labour market is in good shape and may even be accelerating ^[34][35], and fresh capital goods orders data suggest the AI infrastructure boom remains very much alive ^[35].
• [NEW] Forecasters converge on a soft-but-not-weak payroll print and a firmer PCE print: JPMorgan expects September payrolls to rise and unemployment to hold with core PCE at 3.3% year-on-year ^[14]; Citi expects payrolls of about 85k with unemployment edging to 4.2% ^[15]; Barclays expects a smaller payroll gain with unemployment unchanged ^[36]; the Bloomberg-surveyed consensus expects a solid headline with unemployment unchanged and both core and headline PCE ticking up on the month ^[34]. Citi expects the core PCE released Wednesday to be revised down substantially — dovish at the margin, but already priced ^[15].
• [NEW] Activity tracking stayed hot: Goldman Sachs keeps Q3 US GDP tracking at 3.4%, its MAP economic surprise index rose to +0.72 and its current activity indicator edged up ^[37][38], with 2026 full-year growth tracking near its estimated potential and consumer spending running slightly above the FRB/US model’s post-oil-shock estimate ^[38].
• [NEW] The AI composition fact: UBS calculates that over the past four quarters more than 100% of US growth came from AI, with AI-related investment up about 30% while other investment components showed no growth and the non-AI “other” sector contracting 0.7% year-on-year ^[39].
• [NEW] August core CPI remains the anchor: the 0.3% month-on-month print against 0.2% expected triggered the repricing and the Fed’s 25bp hike, and most officials are assumed to have preset a total of 50bp of hikes this year ^[15].
• [NEW] Narrative impact: the data flow continues to confirm the hike rather than settle the path, and the burden of proof has migrated onto one print — September core CPI — for the October decision ^[16][15]. The hawkish leg now rests on AI-related capex demand, resilient consumption and a labour market that is not a constraint ^[9][14]; the dovish leg rests on the expected large downward revision to core PCE, a soft payroll distribution and Goldman’s view that actual tightening will be smaller than priced ^[15][17]. That asymmetry is why the market trades the number of hikes, not their direction.

3. Financial-Conditions Signals

• [ESCALATED] Rates: the 10-year traded down about 1bp at 5.2278% after Monday’s fresh high, the 30-year down 1bp at 5.466% and the 2-year flat at 4.9243% ^[1]; Monday’s close put the 10-year at 5.24% and the 30-year at 5.56% ^[19], with the 10-year having spiked to a 19-year high above 5.27% overnight and the 2-year up more than 57bp this month to the 5% threshold ^[40]. The 10-year’s four-day rise of 27.5bp is the largest since April 2025, and the WSJ notes it has come without fresh inflation news ^[41].
• [ESCALATED] The driver decomposition is now explicit and is real rates: the five-year real yield rose 0.66 percentage points over the month against a 0.72 percentage point rise in the five-year note, with inflation expectations accounting for just 0.06 percentage points, and real yields have been the key driver of Treasury yields all year ^[35]; Deutsche Bank independently finds the move mainly reflects repricing of the expected real policy rate path, with breakevens, swap spreads and nominal term premium broadly unchanged, and puts 10-year fair value at 5–6% with the 1y1y nominal SOFR at about 4.95% already at or above policy-rule levels ^[42]. HSBC adds that cash Treasuries have outperformed swaps and dealer inventories are range-bound, indicating fiscal concerns are not the main driver of the recent rise ^[43], while Deutsche Bank places the 10-year term premium low once the 2010s anomaly is excluded ^[42].
• [NEW] Technicals caution on sustainability: Reuters’ technical read says the rise above 5% may be set to slow, with the 10-year above its upper yearly Bollinger Band for only the second time since 1984 and annual RSI at its highest in more than 40 years, though no outright easing is signalled ^[44].
• [ESCALATED] Credit spread widening broadened: rate-volatility has suppressed bond demand and pushed the liquid US IG index to 96bp, only 3bp off the year-to-date widest of 99bp ^[45]; BB-rated spreads widened 28bp (up 18%) from Sept 18 to Sept 28, more than single-B (up 13%) and BBB (up 9%), moving the BBB/BB ratio back to more typical levels ^[45]; high-yield spreads widened to levels not seen since April, with CCC spreads at 1,128bp from 800 a year earlier and BB at 176bp from 153 ^[46]. Euro investment-grade outperformed US, and cyclical or supply-hit sectors — media, autos, leisure, banks — underperformed while defensives outperformed ^[45]. UBS frames the regime as widening dispersion between stronger and weaker borrowers, favouring issuers with durable cash flows and cautioning on technology, communications and CCC credit ^[46].
• [NEW] Liquidity plumbing is still benign on the surface: reserve balances sit near $2.92tn with Fed total assets roughly stable at $6.8tn including $4.55tn of Treasuries and modest continued Treasury purchases; standing repo facility usage has averaged under $100mn and discount window usage has stayed at historically very low levels; the EFFR-IOER spread has held at -2bp; SOFR-EFFR and SOFR-IOER turned stably negative again from mid-September; Libor-OIS has kept falling and SOFR-OIS has deepened negative, indicating ample dollar liquidity ^[23]. Against that, the US commercial paper-EFFR spread has risen sharply to a 16.7bp average from 10.3bp, German government bond bid-ask spreads have broken above 0.5bp and Japanese government bond bid-ask spreads remain elevated ^[23].
• [ONGOING] Dollar: the dollar index rose 0.7% last week to 101 ^[24], and Deutsche Bank estimates it was about 2% undervalued over the past month on rates and oil, sees downside limited and judges a break above the top of its one-year range unlikely, while noting the market prices a Fed terminal rate above 90bp versus its own 50bp and the dot plot’s 25bp ^[47]. Morgan Stanley forecasts the index at 104 by mid-2027 with EUR/USD at 1.10 on wider rate differentials and strong US growth ^[24][13], while Goldman sees near-term support without broad-based dollar strength and expects gradual renminbi appreciation ^[17]. Spot FX: the yen hovered at 157.31 per dollar and the euro held at $1.1367 ^[40].
• [ESCALATED] Fiscal term-premium channel: US federal debt has exceeded $40tn with interest expense rising markedly, and the CBO projects the deficit staying high over the next decade — with a rising term premium effectively lifting the entire floor under Treasury yields ^[32]; net interest expense as a share of GDP has exceeded its early-1990s peak and is expected to keep rising, while advanced-economy net public debt has gone from 52% of GDP in 2000 to 80% in 2025 and is projected higher ^[48]; HSBC notes more than half of the current ~6% deficit is driven by net interest expense at about 3.3% of GDP and that this share rises further as debt is refinanced above 4%, with the CBO projecting publicly held debt from about 100% of GDP in 2026 toward about 120% over ten years and a $41.1tn debt ceiling hit in 2027 ^[43].
• [NEW] Household channel: DM household debt-to-income ratios are near a 25-year low and debt-service ratios sit below pre-pandemic levels after the 2024–2025 cutting cycle, but JPMorgan’s model projects debt-service ratios facing upward pressure as policy rates rise broadly through mid-2027, especially in high-debt, short-duration countries such as Sweden, Canada and Australia ^[14].

4. Global Central-Bank Linkages

• [NEW] RBA: the Reserve Bank of Australia said monetary policy is well positioned to address developments with the board focused on its price-stability and full-employment mandate ^[49], while a separate relay states that inflation remains elevated and the board decided further financial tightening is needed to ensure inflation returns to target ^[50] — single source / unverified social relays. Governor Bullock said bond markets are reacting in an orderly way but the RBA is watching closely ^[51] — single source / unverified. Reuters reports an RBA hike is fully priced along with another by February ^[40], JPMorgan expects the RBA to resume hiking with 100bp of cumulative tightening this year ^[14], and Westpac questions whether the governor can be hawkish enough to validate current market pricing, especially if the decision is not unanimous ^[40].
• [NEW] ECB: President Lagarde pushed back on expectations of an October hike, saying the central bank is already at the upper end of the neutral rate range and will respond more cautiously ^[19][14]. JPMorgan expects euro-area inflation to stay sticky with the policy rate at only the upper end of neutral, skewing risk toward further tightening with hikes in December and March lifting the rate to 3% ^[14]; Morgan Stanley expects a December hike to 2.75% followed by a single December 2027 cut to 2.5% ^[13]; Barclays expects a third 25bp hike in December ^[36]; and the ECB has warned of an energy-price crisis ^[52].
• [NEW] BOJ: JPMorgan sees improving Tankan corporate sentiment supporting further hikes but judges the probability of an October hike lower than for the Fed or the ECB ^[14]; Goldman’s FX strategists raised their yen forecast on the BOJ’s more hawkish stance and a potential shift in portfolio flows ^[53]; Barclays expects hikes in January and July next year to a terminal 1.75% ^[36]; and a former BOJ official expects hikes every quarter ^[52].
• [NEW] BOE: Morgan Stanley adjusted its expectation to hikes in November 2026 and February 2027, driven by commodity prices and the UK regulatory and fiscal outlook, and says the BOE could hold if oil and refined-product supply improve ^[13]; Barclays expects 25bp in November and February if Middle East tensions persist ^[36].
• [ONGOING] PBoC: the Q3 monetary policy committee statement’s framework stands; in this batch the PBoC-affiliated view is that liquidity and structural tools will front-run fiscal support with a 10bp cut and/or 25–50bp RRR cut still possible within the year; the renminbi is seen stable-to-slightly-strong with the year-end USD/CNY forecast trimmed to around 6.65, though Fed hikes, the China-US rate differential and potential domestic easing act as constraints ^[54][55]. The committee’s exchange-rate language — a decisive market role, greater FX resilience and prevention of herd effects and self-reinforcing irrational expectations — is unchanged ^[56].
• [NEW] SNB: the Swiss National Bank kept its policy rate unchanged at zero ^[52].
• [NEW] Others: a Bloomberg poll of analysts and traders found the Federal Reserve and the outlook for further US rate hikes to be the dominant driver of Chile’s bond market ^[57].

5. Asset Implications

Quadrant	Current probability tilt	Key asset implication	Anchoring narrative
Growth↑ + Inflation↑	Rising further	Energy and commodities plus TIPS remain the cleanest expressions, with Brent above $100 and up sharply year-to-date against nominal long bonds capped by a 10-year above 5.2% and a 30-year near 5.5%; front-end carry, curve structures and short-duration credit still dominate outright duration	§2 / §3
Growth↑ + Inflation↓	Falling further	The disinflation leg lost its clearest remaining support as an official named AI capex as a near-term inflation driver and the DM flash PMI hit a four-year high; what survives is the expected large downward core PCE revision and the view that actual tightening will be smaller than priced, favouring quality cash-flow equity beta over duration	§1.1 / §2
Growth↓ + Inflation↑	Rising (tail)	The supply leg is policy-insensitive and now explicitly named from the demand side as well — AI competition for energy and labour — while the household channel transmits first through debt-service ratios modelled to rise into mid-2027, so commodities and gold, not duration, are the hedge	§1.1 / §3
Growth↓ + Inflation↓	Falling	The cycle tail stays expensive against pricing of roughly 100bp of further hikes over twelve months, a market-implied peak above 4.80% versus a dot plot implying one more move, and a sell-side distribution whose lowest published paths still hold policy restrictive through 2027	§1.2 / §3

Stock-bond correlation call: the regime remains inflation/policy-driven positive correlation — the configuration hardest for risk parity — and today’s evidence sharpens the mechanism in a specific and unfavourable way. The decisive new fact is the decomposition: essentially the entire rise in the five-year note over the past month came from the real-yield component, with inflation expectations contributing a trivial slice, and the same conclusion is reached independently by two sell-side desks that find breakevens, swap spreads and nominal term premium broadly unchanged. Duration is therefore repricing precisely the variable equities reprice — the expected real policy path — through the same discount-rate channel, and it is doing so while officials describe the tightening as risk management aimed at a supply-and-AI-demand mix that a policy rate reaches only with a lag. Layered on top is a correlation-structure fact that argues against treating commodities as a clean hedge on a portfolio basis even though they remain the cleanest expression of the quadrant: Barclays finds the average pairwise correlation between the S&P 500, Brent and Treasury yields at its highest in thirty years since the Iran conflict began, while the returns of those three assets diverged sharply — equities up, oil volatile, bonds sold off. High correlation with divergent returns is what makes cross-asset hybrids attractive and simple risk-parity diversification unattractive. Two counter-prints deserve to be named precisely. First, the technicals: the 10-year has breached its upper yearly Bollinger Band for only the second time since 1984 with annual RSI at a four-decade high, which does not signal easing but does argue the advance is becoming harder to sustain. Second, the positioning: US equity futures positioning shows only about a tenth of contracts long with the premium-weighted put/call Z-score beyond two standard deviations, and JPMorgan’s own note records that despite those oversold signals no mean reversion occurred and long strategies were stopped out — which is the signature of a discount-rate market, not a positioning accident. Directional call: positive correlation persists, and the tail risk is that it worsens before it improves.

Risk-budget implication: keep commodity and energy risk at or above benchmark, but recognise that its diversification value has fallen even as its directional case has strengthened — with thirty-year-high cross-asset correlation, energy is simultaneously the stated inflation driver, the lowest-cost stagflation hedge and a less effective portfolio diversifier than it has been. In rates, do not rebuild long-end duration into a move that is provably real-rate-driven and still being validated by data: the sell-side distribution spans two further hikes with the upper bound below 4.4% against market pricing of nearly 100bp of additional tightening and a peak above 4.80%, front-end pricing may already have over-reflected tightening, and the robust expressions remain belly and curve structures — one desk’s stated trade is a long 2s10s spread position funded out of an outright duration short. Note the two structural supports that argue against treating duration as uninvestable: fund flows have begun moving out the curve into intermediate and long government funds, and the Fed is expected to restart reserve-management purchases of $10–20bn a month in the fourth quarter, both of which support the belly and the long end relative to the front end rather than the front end itself. In credit, keep duration short and prefer spread over pure duration, but respect that the widening has become broad rather than idiosyncratic: the US investment-grade index is only 3bp from its year-to-date wides, BB underperformed single-B and BBB, and high-yield spreads are at their widest since April with CCC spreads up by hundreds of basis points over the year — the operative structure is dispersion, with defensives outperforming cyclical and supply-hit sectors, and the concrete technical to monitor is the Q4 funding pressure that would force the Fed back into bill purchases. In equities, hold moderate beta expressed through cash-flow durability rather than rate-sensitive cyclicals: utilities, real estate and small caps are the sectors taking the damage while large-cap tech and momentum lead, the AI complex is both the equity beta and a major issuer of the debt being repriced, and market breadth has deteriorated to a configuration historically seen only in 1999–2000. Hold gold as the structural fiscal and dollar-credit hedge but size it for continued real-rate pressure: the metal has fallen to a near-seven-week low and at least six major banks have cut 2026 forecasts since June, yet central-bank purchases running multiples of their pre-2022 pace and a structural debasement bid are the explicit reason the long-horizon targets remain far above spot — the trade is now a low-correlation diversifier with a real-rate headwind, not a duration substitute.

6. Contrarian & Tail Risks

• Consensus fragility: the baseline is now “October hike above 70%, roughly 100bp more over twelve months, and the long end has repriced for it.” Four falsifiable pillars hold it up. First, the pricing is a market construct built against a Fed that has deliberately withdrawn guidance: the dot plot implies one more move while the market prices several, and the readings themselves are a band across instruments and snapshots — above 72%, 72.5%, 70.3% and roughly 70% — so venue and vintage dominate any single print. Second, the sell-side cannot all be right: one desk expects a peak with the upper bound below 4.4%, another says 100bp of hikes is entirely possible, another argues actual hikes will be two to three rather than seven to eight, and Citi reduces the entire October decision to a single September core CPI print with only negative payrolls or a 4.3% unemployment rate able to move pricing materially. Third, the justification is circular at the margin: officials are explicitly watching whether market expectations become self-validating, and the historical base rate is that markets underestimate the total extent of hikes. Fourth, and most fragile, the analysis rests on a framework in which the rule set itself is the uncertainty — no forward guidance, an operational de-emphasis of r-star, and data-dependence affirmed in every appearance.
• Second-order transmission: four chains are not in market pricing. The AI-financing-and-demand chain has now become two-sided rather than one-directional: an official has stated publicly that AI infrastructure spending is competing for energy and labour and is a near-term inflation driver, while an investment bank calculates that more than all of US growth over the past year came from AI with the rest of the economy contracting — meaning tightening bites the non-AI sector harder and exacerbates the imbalance, and pushing the unemployment rate high enough to hit 2% inflation would risk recession. The funding-plumbing chain is dated and mechanical: TGA rebuilding plus heavy short-term T-bill supply plus seasonal balance-sheet contraction is expected to force the Fed back into bill purchases at $10–20bn a month, and a commercial paper-EFFR spread that has already jumped by roughly two-thirds is the first visible crack. The fiscal-reflexivity chain is quantified on both sides: debt above $40tn, net interest above its early-1990s peak as a share of GDP and more than half of the current deficit, a term premium whose rise lifts the entire floor under yields, and a debt ceiling to be tested in 2027 with the historical precedent of short-bill yields spiking into the X-date. The household chain is slower but now modelled: debt-service ratios sit below pre-pandemic levels but are projected to face upward pressure through mid-2027 as policy rates rise, concentrated in high-debt, short-duration economies. Layered on top, a reserve-status chain: a central bank’s independence is tested less by any single hike than by personnel turnover across years, and the operational channel through which that risk prices is the long end, where more political pressure historically demands a higher premium rather than a lower one.
• Source quality control: a large share of the most quotable items today are single-source social relays and should be treated as indicative only — Governor Cook’s remarks arrive through one-line posts covering AI-fuelled demand pressure, limited Fed tools, the risk that cuts fuel inflation and data-dependence, and the “further inflationary pressures ahead” headline is a relay of a wire report condensed to a social post; the RBA’s policy positioning and inflation language, its governor’s “orderly” comment, and the European Central Bank’s and Swiss National Bank’s lines are all relayed or secondary. Direct conflicts deserve side-by-side reading: the size of the tightening cycle is disputed in the same batch by at least three incompatible paths (two hikes with a sub-4.4% upper bound, roughly 50bp with 100bp possible, and two to three hikes rather than seven to eight) and neither side should be traded alone; the driver of the long end is attributed both to real-rate repricing with term premium unchanged and, at the same time, to a term premium that has risen enough to lift the floor under all yields; the dollar is simultaneously called 2% undervalued with downside limited and forecast to a multi-year high by mid-2027; and the severity of credit stress depends entirely on whether the widening is valuation-thin dispersion or the start of a refinancing cycle — the two readings come from the same week’s data. Two data-quality flags: the level and timing of the 10-year’s high differ across venues, so level-based triggers must specify source and timestamp; and fund-flow series carry their own coverage caveats. Finally, several items are model- or survey-derived rather than observed — the AI share-of-growth decomposition, the debt-service ratio projections, the policy-rule tables, the Phillips-curve unemployment requirement and the technical Bollinger/RSI readings — and are useful for framing a distribution rather than as price levels.

---

Appendix: Additional Sources

• ^[5] Financial Juice — Cook: sees some economy-wide pressure from AI-fueled demand
• ^[6] Financial Juice — Cook: highly attentive to AI potentially causing a temporary rise in unemployment; limited Fed tools; cuts could fuel inflation
• ^[21] Financial Juice — Cook: number and magnitude of future rate adjustments informed by inflation and labour-market data
• ^[22] Gelonghui — Cook: will consider the policy rate needed to keep guiding inflation back to target
• ^[20] Yicai — Cook: labour market looks able to withstand further rate hikes
• ^[58] Christophe Barraud — Cook sees further inflationary pressures ahead (Reuters relay)
• ^[10] Jin10 — hike pricing at 70% with the yield curve rising across the board
• ^[9] Deutsche Bank — Fed Watcher: officials reinforce hawkish expectations; baseline 25bp in December and March
• ^[42] Deutsche Bank — yield rise reflects real policy-rate repricing; long 2s10s spread recommendation
• ^[25] Barclays — US rates flow of funds: out-the-curve demand and expected resumption of RMPs

---

This report is a macro-mechanism analysis, not investment advice.