Hammack Just Tipped the Fed Toward a December Hike — The 5.36% 10-Year Is Already Pricing Both the Bank NII Trade and the AI Capex Funding Test

Cleveland Fed President Beth Hammack used her Oct. 12 speech — titled \"Trimming the Sails: Communicating My Reaction Function\" — to do more than repeat the September dissent she filed at the July FOMC. She laid out the conditional logic that makes another move in December the path of least regret for a voter who already thinks current policy is insufficiently tight.

The core of the message: \"a restrictive monetary policy stance is appropriate to ensure a timely return to 2 percent inflation,\" and if it \"remains elevated relative to my forecast or the settling in of an inflationary mindset could require pointing the bow of monetary policy rates toward an even higher latitude.\" She frames reaction functions as conditional — not a promise — but the directional bias is unambiguous: more restrictive, not less.

Even before Hammack's speech, the September Summary of Economic Projections pointed toward a fourth-quarter move: the median participant's end-2026 federal funds rate is 4.1%, exactly 25 bp above the post-meeting 3.75–4.00% target. Sixteen of eighteen participants expect at least one more hike in 2026; four of those expect two. No participant's2026 dot sits below the post-meeting target.

Hammack's speech placed her squarely in the four-participant \"two more hikes\" camp. Her inflation risk language — \"skewed\" to the upside and the threat of an \"inflationary mindset\" — is consistent with the 4.375% high-end dots. With the October 28 meeting still 16 days away and the September CPI and PPI prints not yet digested, the December9–10 meeting is the calendar slot where the median dot catches up to her reaction function.

The arithmetic is stark. To keep the median at 4.1% by year-end from the current 3.75–4.00% target requires exactly one more 25 bp hike. If the Fed delivers zero hikes, the median would fall to roughly 3.875% — implying four participants (Hammack included) vote against any cut and the dot plot reverts higher at the December meeting.

The 10-Year Treasury touched5.36% on Oct. 7 — its highest level since April 2002 — before a solid $39 billion 10-year note auction pulled it back toward 5.28% by Oct. 9. Hammack herself, in a late-September interview, attributed the move to \"real rates that have moved up more than inflation expectations,\" driven by stronger growth, a repricing of monetary policy, and heavy technology investment. Breakeven10-year inflation expectations have stayed anchored near 2.4%, not at the levels the move would imply if this were a tariff-inflation scare.

The supply side is doing real work. Hyperscaler bond issuance has reached roughly $220 billion through Aug. 10, per BNP Paribas data — already more than double 2025's $93 billion run rate. Alphabet, Amazon, Meta, Microsoft and Oracle are absorbing long-end supply that the Fed is no longer offsetting. Front-end yields are pricing the policy path: the 2-year is at 4.80%, up 133 bp year-to-date, reflecting market conviction that the Fed's terminal rate has stepped higher. The 30-year at 5.60% is the curve's other anchor.

Q2 2026 net interest income at the four largest US commercial banks was up sharply year-over-year, and the Q3 consensus now expects the curve to keep repricing in the banks' favor. Q2 prints from Q2 2026 10-Q filings show year-over-year NII growth of 9.9% at JPMorgan Chase, 9.0% at Bank of America, 5.2% at Wells Fargo and 12.9% at Citigroup — and that was with the 10-Year only at 4.6–5.2% during the period.

JPMorgan Chase raised its FY2026 NII outlook after Q2, guiding to roughly $103 billion in total NII and $96.5 billion excluding Markets. The Q3 2026 consensus now models managed NII of $26.84 billion, up from $24.07 billion a year earlier. Q3 earnings land Oct. 14 — three trading days after Hammack's speech and one day before this article publishes.

The compression risk is real but receding. The St. Louis Fed's banking analytics showed industry NIM fell from 3.30% in Q4 2025 to 3.22% in Q1 2026 — the first sequential NII decline since mid-2024 — as loan yields dropped from 6.75% to 6.51%. With the 10-Year now well above 5%, asset yields are climbing again while deposit costs lag, reopening the rate-cycle NIM tailwind that powered 2022–2024 earnings beats.

The flip side of the 5.36% 10-Year is the cost of funding the biggest capital expenditure program in corporate history. The five largest US hyperscalers — Alphabet, Amazon, Meta Platforms, Microsoft and Oracle — are on track to spend roughly $725 billion on capex in 2026, up from about $430 billion in 2025. Goldman Sachs Research pegs the global AI investment figure even higher, at $1 trillion globally and $581 billion in the US, with US AI investment forecast to reach 1.8% of GDP this year and 2.8% by 2028.

The funding gap is closing, but at a price. Oracle's fiscal Q1 202710-Q shows $1.12 billion of negative net interest income on $5.1 billion of interest expense — the AI capex build is now showing up as a100+ bp drag on the income statement, and the long-end reset is the proximate cause. Meta Platforms's and Amazon's debt service costs are scaling similarly; every 100 bp move on $100 billion of long-dated debt adds roughly $1 billion of pre-tax expense per year before any operational offset.

The credit market is rewarding scale. Investment-grade hyperscaler paper continues to clear at tight spreads — the AI-credit premium is barely 95 bp over Treasuries, despite the volume — but the absolute coupon is the variable that matters. A 30-year issuance at 5.60% locks in a cost of capital 110 bp higher than the same deal would have cleared at in January. That math is the test Hammack's December hike sets up for 2027 capex plans: every basis point of Fed tightening pulls forward a non-trivial wedge between AI infrastructure ROI and the cost to fund it.

Short-term (days to quarters), the trade is unambiguous. The Q3 bank earnings cycle that opens Oct. 14 with JPMorgan Chase, Bank of America, Wells Fargo and Citigroup will print NII numbers that lap the lower rates of Q3 2025. A December hike fully priced into the SEP by the December 9–10 meeting would push the 10-Year toward 5.50% and add roughly another 50 bp to bank asset yields before FY2027 guidance is set.

Long-term (1–3 years), the question flips. If Hammack's reaction function forces the Fed to a 4.375% terminal and holds it there through 2027, the bank NII trade continues to compound, but the AI capex ROI math compresses. Goldman Sachs Research models US AI investment reaching 2.8% of GDP by 2028 — a path that requires the long end to stay below 5.50% for issuers to clear with positive NPV. The 5.36% print on Oct. 7 is the boundary. Beyond it, hyperscaler capex plans get renegotiated, not abandoned; debt-funded M&A in the AI value chain slows; and bank lending standards quietly tighten as the curve works against the issuers the banks underwrite.

Hammack's Oct. 12 speech is not new information in isolation — she has been the FOMC's most vocal hawk since the July dissent. What makes it operative now is the convergence: the September SEP median (one more), the September CPI print (still sticky at 3.0–3.4%), the October 10-year auction at multi-year highs, and a hyperscaler debt-supply pipeline that is pulling real yields higher without Fed action. The Fed is being out-hawked by the bond market, and Hammack is simply the most on-the-record voice pointing it out.

The investable conclusion is two-sided. JPMorgan Chase, Wells Fargo, Bank of America and Citigroup are long a curve that has done three years of repricing in eight months. Goldman Sachs and Morgan Stanley are long the underwriting pipeline that comes with the volume. The short side — if there is one — is the hyperscaler capex math: Oracle's negative net interest income is the cleanest signal that the funding cost has caught up with operating leverage, and Microsoft, Amazon and Meta Platforms are not far behind. The trade works until the long end breaks5.50%; Hammack's reaction function suggests it will.