The Long End Stopped Waiting for the Fed

1. What had already happened

Everyone was watching Wednesday. The target range entered the meeting at 3.75% to 4.00%, after a quarter-point increase on September 16 that passed twelve to nothing, and the market priced a hold at about 91% with a further increase by December at roughly three in four.^[1][3][4] This is a meeting without projections, so the next set of dots arrives December 9, which left Wednesday as a meeting about language.^[14]

The more interesting move came earlier, in bonds. After the September increase the ten-year kept climbing, from 5.01% on the day of the decision to a 5.31% cycle high on October 5. Then it turned. By last Friday it had closed at 5.00%, giving back the entire rise since the meeting, while the three-month bill, which simply tracks the policy rate, sat at 4.22%.^[2] The front end still carries the September increase. The long end has stopped waiting for the next one.

A falling ten-year is usually good news for the stocks most sensitive to it, which is what made the equity tape hard to explain. Between the September 16 close and October 6 the Nasdaq Composite rose 6.2% while the Dow was flat. From October 6 to last Friday the Nasdaq gave back 6.0% and the Dow added 1.3%, with long yields falling the whole way.^[2][5] The familiar story, that higher rates were weighing on expensive growth, does not fit a stretch in which rates fell and growth fell with them.

So we kept coming back to the financing. The October 19 outlook argued that the question had moved on from how much the largest technology companies would spend on AI capacity to what it costs the companies building that capacity to borrow, and whether lenders were growing less willing to fund them.^[12] CoreWeave’s five-year default swaps are the clearest print of that: a record 1,194.5 basis points intraday in mid-October, a level ICE Data Services read as a better-than-even chance of default within five years, and about 1,038 at the last settle reported before the meeting.^[6][13] Prices alone cannot prove that credit explains the Nasdaq’s October. They rule out rates as the explanation, and the credit prints moved in the right direction at the right time.

The curve puts the meeting in scale. Against June 17, the Chair’s first meeting, every tenor was 35 to 58 basis points higher at Wednesday’s close. Against September 16, no tenor had moved more than seven. Whatever the committee did Wednesday, it did it into a market that had already repriced the whole year and then undone the last month of it.

2. The decision

In the October 19 outlook we wrote: “A hold changes nothing in this framework. Any increase is a genuine surprise that tightens the same financing channel this note is about, and we would expect the AI complex, not the Dow, to take it.”^[12]

A hold resolves that sentence as written. What the meeting adds is language and a vote, so both are set against the two previous statements.

Statements quoted from the Federal Reserve’s own releases.^[1][8] The September statement ran to 114 words; the October statement runs to 123.

In July the committee held nine to three, with Beth Hammack, Neel Kashkari and Lorie Logan preferring an increase; in September all twelve voted for one.^[1][8] The previews we read expected a pause now and an increase in December,^[9][10] and Continuum Economics expected Hammack and Logan to dissent against the pause, with Kashkari a closer call.^[9] The dissents came from somewhere else. Kashkari and Logan voted with the centre; Hammack dissented as expected, and Lisa Cook, a governor, joined her. So the wing is two votes, it sits partly on the Board, and the centre chose to wait with the inflation sentence rewritten to say why: the committee wants “the accumulation of more information” before it sets the stance. September’s sentence said the increase itself would do the work; October’s says the data will decide. The activity sentence moved the other way, from “solid” to “strong,” with capital investment still “robust.”^[1]

The Chair has said he does not prejudge meetings, and he has avoided forward guidance since June.^[11] He gave no December line at 2:30. Asked whether policy was working, he said there was “monetary medicine starting to work some” and “no firm conclusions,” adding that if the committee had one “we’d announce it to much fanfare today.” On financial conditions he was “unwilling and unable to call them tight.”^[15] The December contract did not move on any of it. The one answer that bears on this note came late, on whether higher rates reach AI investment at all. He said he was skeptical of the view that they do not: “Interest rates get in the cracks. Interest rates affect directly or indirectly almost everything,” and, on the investment surge, “do interest rates matter? I believe they do.”^[15] That is the Chair saying policy reaches the capital-spending cycle. The market’s version is that it reaches it through the long end, which rose on the day he said it.

3. What it does to the financing question

The Fed reaches the financing channel only through the long end and through risk appetite, and it sets neither directly. The table carries the measures the October 19 note set, before and after Wednesday.

Two things outside the Fed will tell us more this week. Microsoft and Alphabet report after Wednesday’s close and Amazon on Thursday, with the first 2027 capital-spending guidance inside them, and the October 19 note already set the test for those.^[12] Then there is the borrower itself. Nvidia’s chief executive said last week that CoreWeave’s financing costs are falling and that Nvidia does not need to backstop it, while CoreWeave’s own filings show Nvidia already committed to buy its unsold capacity through April 2032.^[7][13] In the same week the swap market priced the company at better than even odds of default. The gap between those two readings is the question we are carrying into the capex prints. Wednesday’s decision could not close it.

4. What changes in the book

Nothing today. No published level in the book keys off the policy rate. The levels that can trade this week are the ones the capital-spending guides reach, and in Bloom Energy a target review is due Friday, so that name stays inside a window through November 4. Vistra is inside the window ahead of its November 6 report. This note rates no security and opens no window of its own.

5. What would make us reconsider

1. December pricing, by Friday’s close. The December contract closed Wednesday at 79% on the prediction markets, with fed funds futures nearer 85%. At or above 72% at Friday’s close, the lowest reading of the pre-meeting week less a point, the committee is still on the path its September projections implied and the hold cost it nothing. Below 50%, the market has stopped believing the median, and the long end’s October rally was the start of a different regime. Between the two, Friday’s capital-spending reads carry the argument.
2. The ten-year, standing. A Friday close at or above 5.31%, October’s cycle high, keeps the pressure on long-duration equity. Sustained below it, one source of pressure is easing even if the credit question is not; sitting under a recent high proves little on its own.
3. CoreWeave credit, standing. A five-year default-swap spread back below 800 basis points says the credit scare is healing. At or above 1,140, the October high, says it is deepening.
4. Hyperscaler 2027 capital spending, reported this week and graded in the October 19 note’s own terms: combined guidance pointing higher than 2026 keeps the demand case; flat or lower across more than one major buyer and we treat the cycle as turning.

Housekeeping

Scope. This note reflects the statement and implementation note released at 2:00pm ET and the press conference that followed. Statement language is quoted from the Federal Reserve’s releases, never from coverage of them. Treasury yields are official par-curve closes; index levels are daily closes; anything intraday is labelled. Probabilities are from named markets at a stated time and are approximate.

Trading window. Our policy restricts trading in a covered name from ten days before a publication to five days after it. This note rates no security. Bloom Energy is inside a window through November 4, with a target review due Friday; Vistra is inside the window ahead of its November 6 report note. Protective stops and kills at previously approved levels remain available at any time and are logged when used; none was triggered today.

Pre-registration. The dispersion test in section 2 was logged with its one-point threshold on October 26, before the meeting, in our internal ledger. The December-pricing threshold in section 5 could only be set after the decision, and was logged at the close before Friday’s read.

Sources

1. Federal Reserve, FOMC statement, September 16, 2026: increase to 3¾–4 percent, 12–0. federalreserve.gov. FOMC statement, October 28, 2026: hold at 3¾–4 percent, 10–2, Cook and Hammack preferring an increase. federalreserve.gov
2. U.S. Department of the Treasury, daily par yield curve rates, 2026. home.treasury.gov
3. CME Group FedWatch tool, read October 26 (hold 90.6%). cmegroup.com
4. Kalshi and Polymarket Fed-decision markets, read via an aggregator: October 26 (hold about 93%; December increase about 76%), October 27 morning (73%) and October 28 at 7:04pm ET (December increase 79%, hold 18%). A secondary source for a number that moves by the hour. defirate.com
5. Yahoo Finance daily closes: Nasdaq Composite, Dow Jones Industrial Average, CBOE ten-year yield index. ^IXIC, ^DJI, ^TNX
6. CNBC, October 21, 2026, on CoreWeave’s bond yields and default-swap settle. cnbc.com
7. CNBC, October 24, 2026, on Nvidia’s role as financier to the neoclouds, including the capacity commitment disclosed in CoreWeave’s filings. cnbc.com
8. Federal Reserve, FOMC statement, July 29, 2026: hold at 3½–3¾ percent, 9–3, with the dissenting names. federalreserve.gov
9. Continuum Economics, “FOMC Preview for October 28,” October 20, 2026. continuumeconomics.com
10. ING, “Federal Reserve preview: A pause before a December hike.” think.ing.com
11. CNBC, October 16, 2026, on the Chair’s remarks ahead of the meeting. cnbc.com
12. Our October 19 outlook: “Insurance on a Company You Don’t Own”, section 4, where the four tests were set.
13. TNW, on Jensen Huang’s Bloomberg interview in Stockholm and the record default-swap print. thenextweb.com
14. Federal Reserve, FOMC meeting calendar, 2026: meetings with projections are marked. federalreserve.gov
15. Chair’s press conference, October 28, 2026, 2:30pm ET: federalreserve.gov (video; the transcript posts later). Quotations and the fed funds futures reading (Bloomberg data, about 85% for December) from Yahoo Finance’s live coverage of the press conference: finance.yahoo.com. To be checked against the transcript when it posts.
16. Company dates, confirmed on the issuers’ own pages: Microsoft FY27 Q1 call October 28 (microsoft.com); Alphabet Q3 release dated October 28 (sec.gov); Amazon Q3 call October 29, 5:00pm ET (ir.aboutamazon.com).