Market Week · 13 November 2026
Market Week, 13 November 2026
Covers trading sessions from 4 to 12 November.
The week in one question
On Thursday 5 November the Nasdaq fell 5.52% and Treasury yields fell with it. A week later yields were at their highs for the month and the Nasdaq had recovered about three-quarters of that day's loss. Stocks went down when money got cheaper and came most of the way back while it got dearer. Whatever moved prices this week, it was not the discount rate alone.
The better explanation sits in four earnings releases and one bond offering. Demand for AI computing is not in doubt. Who gets paid for it is.
A note from Bob Katz. I have watched this shape before. In the dot-com years telecom firms laid fiber far ahead of the traffic, and a great deal of it sat dark, unneeded and unused. The resemblance to today's hyperscale build is eerie. I also remember the memory business, where in a shortage customers would place outsized orders hoping to raise their allocation. When the market turned, they found themselves holding the excess, and the orders on the books turned out to be partly a measure of fear, not of use. Memory and storage are short again now, and prices are rising. I can't tell from here how much of a $132.5 billion backlog is need and how much is allocation. Nobody could tell then either.
What happened
Stocks. The S&P 500 closed at 7,688.62 on 4 November, lost 4.36% the next day, and stood at 7,592.40 on 12 November, 1.3% below where the stretch began. The Nasdaq Composite closed 5 November at 24,315.00, which the Associated Press put at 11.9% below its October record. Small companies never joined the recovery: the Russell 2000 fell another 1.66% on 12 November.
Rates. This is the part that changed most.
| 5 Nov | 12 Nov | Change | |
|---|---|---|---|
| 2-year Treasury | 4.46% | 4.86% | +40 bp |
| 10-year Treasury | 4.84% | 5.10% | +26 bp |
| 10-year real yield (TIPS) | 2.60% | 2.86% | +26 bp |
| High-yield spread over Treasuries | 3.48 pts | 3.20 pts | −28 bp |
| High-yield index yield | 8.07% | 8.16% | +9 bp |
| Investment-grade spread | 0.93 pts | 0.89 pts | −4 bp |
The 10-year nominal yield and the 10-year real yield rose by the same 26 basis points. Investors did not ask for more inflation compensation. They asked for a higher real return.
Credit did not flinch. The extra yield investors demand on high-yield bonds peaked at 3.48 points on the day of the selloff and narrowed to 3.20 by 12 November. The all-in cost of high-yield borrowing still rose, to 8.16%, because the Treasury yield underneath it rose by more than the spread fell.
The data behind it. October payrolls rose 165,000 with unemployment at 4.1%, and the prior two months were revised up by 84,000. October consumer prices rose 0.4%, core 0.3%, the third straight month at that core pace. The Federal Reserve's target range is 3.75% to 4.00%, and two members already dissented in favor of a rise at the October meeting. Against that, third-quarter productivity grew at a 3.6% annual rate and unit labor costs fell 0.7%.
Auctions. The Treasury sold $42 billion of 10-year notes at 5.020% and $25 billion of 30-year bonds at 5.441%. Bids covered the 30-year sale 2.36 times.
Sellers and buyers
Four companies reported within three days of each other. Two sell the equipment. Two buy it and rent out the capacity.
| Role | Quarterly revenue | Growth | What the quarter produced | What it took | |
|---|---|---|---|---|---|
| Cisco | Seller | $18.5B | +24% | $4.4B cash from operations | — |
| Applied Materials | Seller | $10.41B | +53% | $3.30B cash from operations | — |
| CoreWeave | Buyer | $3.66B | +168% | net loss of $709M | about $46.9B of debt outstanding |
| Nebius | Buyer | $0.98B | +571% | $273M cash used in operations | $7.3B of capital spending |
Nebius spent roughly seven and a half dollars on equipment for every dollar of revenue it booked in the quarter. It now expects about $27 billion of capital spending this year against revenue at the low end of $3.0 to $3.4 billion, and says the annual run-rate it had guided for December will arrive in February instead. The contracts are still in place; the capacity is late.
CoreWeave reported a backlog of about $132.5 billion and raised its 2026 revenue outlook to $13.0 to $13.3 billion. Its balance sheet carries about $46.9 billion of debt, by our sum of the four debt lines in its release.
Then the price of the money. On 2 November Oracle priced $10 billion of notes with coupons from 7.000% to 8.650%, alongside about $10 billion of new shares. Treasuries of similar maturities yielded between about 4.9% and 5.4% on 12 November. A company of that standing is paying roughly two to three points over the government to fund this build. Its longest notes, at 8.650%, pay more than the high-yield index's 8.16%. The average investment-grade borrower paid 0.89 points over Treasuries on 12 November.
One more data point belongs here without interpretation. On 12 November Nvidia announced that its chief executive had resigned, effective immediately, with the chief financial officer named interim chief executive. The company gave no reason. Its quarterly results remain scheduled for 17 November.
Evidence vs speculation
- Evidence: equipment sellers are converting AI demand into cash now. Cisco and Applied Materials together generated $7.7 billion of operating cash in one quarter.
- Evidence: capacity operators are growing revenue several-fold while consuming cash and adding debt. These are company-reported figures, not estimates.
- Evidence: the real cost of long-term money rose 26 basis points in a week, and all of the rise in the 10-year yield was real.
- Evidence: the broad credit market did not treat the week as a credit event. High-yield spreads ended it 28 basis points tighter than on 5 November. The strain shows up in what individual AI builders pay, not in the index.
- Speculation: that the operators will not earn their cost of capital. A backlog of $132.5 billion may convert on good terms. Nothing reported this week settles it either way.
- Speculation: that one cause explains 5 November. Positioning, policy expectations after the election and doubts about returns were all cited, and price action cannot separate them.
- Speculation: that one quarter of 3.6% productivity growth is a trend, or that AI produced it.
What would change our mind
The skeptical reading is that operators are borrowing at 7% to 8.65% to buy hardware that depreciates in a few years, on contracts whose cash arrives later than promised. It is a reasonable reading, and it could be wrong. These would count against it:
- an operator reporting positive cash from operations after interest for two consecutive quarters;
- backlog converting to recognized revenue on schedule, with no further slips of the kind Nebius described;
- new financing priced meaningfully closer to Treasury yields than Oracle's;
- real yields falling back while credit spreads stay near 3.2 points.
One of these is already half true. Spreads tightened this week even as real yields rose. If the skeptical reading were plainly right, lenders to weaker companies should have been the first to notice, and the index says they have not. Both things can hold: the market for ordinary credit is calm, and the market for funding this particular build is expensive. Which one gives way is the open question.
Next week
- Tuesday 17 November: October import and export prices; Nvidia quarterly results.
- 8–9 December: Federal Reserve meeting.
Sources
Index closes: Associated Press daily tables, cross-checked against a second price source. Treasury yields, real yields and auction results: U.S. Treasury. Credit spreads and high-yield index yield: ICE BofA option-adjusted spread and effective yield indexes, as published by the Federal Reserve Bank of St. Louis (no observation for 10 November). Employment, prices and productivity: Bureau of Labor Statistics. Policy rate and dissents: Federal Reserve statement of 28 October. Company figures: each company's own earnings release or SEC filing. Oracle terms: its pricing announcement of 2 November.
This note is drafted with AI research assistance and reviewed and signed by Bob Katz. It is market commentary, not investment advice, and not a recommendation to buy or sell any security named. Company figures are as reported by the companies and have not been audited by us.