New Highs, an Unmoved 10-Year, and a Household That Is Not Celebrating | Week Ending 9 October 2026

The Week at a Glance

Wall Street spent the week proving it can make a high with the bond market unconvinced. The index did the celebrating. The 10-year did not blink. The household survey declined the invitation — on the headline, not on every line of it.

On Tuesday the S&P 500 and the Nasdaq printed all-time highs. Friday they finished just shy of those marks, and still up on the week. The final scoreboard, Friday 2 October close to Friday 9 October close: the S&P 500 finished at 7,811.54, up 88.82 points, or 1.2%. The Dow closed at 51,654.95, up 0.9%. The Nasdaq Composite ended at 27,366.17, up 0.6%. The Russell 2000 went the other way, down 0.9% to 2,806.98. Large beat small. That is not a broad risk-on week. It is a cap-weighted one that survived a bond yield still parked above 5.2%.

The 10-year closed Friday at 5.24%. That close is the digest. It is not the week. On Wednesday, the day the minutes landed, the yield traded as high as about 5.36% — a fresh high since 2002, through last Thursday’s 5.34% — before a solid 10-year auction pulled it back. Friday’s 5.24% is where it settled, not where it was asked to settle. A week in which equities make records and the long bond gives back the spike, but not the level, is a week in which the discount rate is still the constraint.

The number that did not bounce was the household headline. Preliminary October sentiment fell to a five-month low, and current conditions printed the lowest reading on record. Expectations ticked the other way. The index and the survey are describing different economies. One is capitalised earnings. The other is a petrol price and a borrowing rate.

1. The Index Made the High. The Bond Market Did Not Blink.

The week is not Friday’s 0.6% on the S&P 500, 0.8% on the Dow and 0.6% on the Nasdaq. Those closes were tidy. The week is that they sat on top of a 10-year that made a new high since 2002 on Wednesday and still closed Friday at 5.24%, the same neighbourhood as last Friday’s rout.

Investment-grade US corporate bonds, on the ICE BofA 7–10 year index, yielded 6.14% to 6.24% across the week. Expensive money is no longer a forecast. It is in the discount rate an equity holder is not marking, and in the borrowing rate sitting behind that record-low current-conditions print.

The Russell’s 0.9% weekly decline is the other tell. When the cap-weighted indices make highs and small caps do not, the rally is not a vote for the economy. It is a vote for the earnings concentrated at the top of the index — the same earnings the New York Fed’s Desk, in the minutes below, says have been carrying the rally while multiples compress. That is a healthier rally than a multiple expansion at a 5.3% 10-year. It is also a rally that dies if the earnings do.

2. The Minutes Explained the Yield

Wednesday’s minutes from the 15–16 September meeting are why the yield was willing to print 5.36% before it was willing to close at 5.24%.

The vote. The Committee had voted 12–0 to lift the funds rate to 3.75%–4.00%, the first hike since 2023. Most participants assessed that another increase would likely be appropriate by year-end. Several described the current rate as not restrictive, or only mildly restrictive.

The split that matters. It is not whether they hiked. It is why. Some wanted insurance against energy and other supply shocks sticking. A more hawkish core wanted to lean against demand. Several noted that core-goods inflation was being pushed by the AI build-out even as tariff effects faded. That last point is the one an Australian holder of the picks-and-shovels complex should not skim. A central bank is now treating the capex boom as an inflation input, not only as an earnings input.

The Desk’s read. Staff and the Desk flagged competition for capital: heavy private debt issuance to fund AI infrastructure, and wide spreads on hyperscaler paper. On the equity side, the rally this year has been earnings, not multiples. Multiples have been compressing under the yield. That is one reading of a healthier tape. It is also a description of a market that has less multiple left to give if the 10-year stays here.

The funds rate sits at 3.75%–4.00%. Friday’s 10-year, at 5.24%, is still about 125 basis points above the top of that range — the same gap that defined last week’s rout. Wednesday’s 5.36% widened it to about 135. The market is not only pricing the next quarter-point. It is pricing oil, debt supply, and a neutral rate that may simply be higher than the Committee’s range assumes.

3. Oil: The Premium Shortened. It Did Not Leave.

Brent settled Friday at US$104.72, up 0.4% on the day, after a session that ran from about US$102.50 to US$105. WTI sat near US$91.65. The gap between the two is the story, not the settlement.

Hormuz risk prices the global barrel. Inland US crude does not wear the same premium. Thursday’s spike faded after President Trump ruled out strikes on Iran before the November midterms. The war premium did not leave. It shortened its horizon. A reader who treats US$90 WTI as “the oil price” is looking at the wrong contract for Australian petrol, and for the inflation print the Fed is actually watching.

Oil was the other hand on the yield. A 10-year that can touch 5.36% and still close at 5.24% with Brent above US$100 is not waiting on a growth scare to tell it what to do. Last week’s jobs miss already tested that. This week’s new highs tested it again.

4. A Leak That Moved Hundreds of Billions

Thursday’s other scare was an accounting argument dressed up as a demand shock.

The Financial Times reported that OpenAI had told investors its annualised revenue run-rate was nearing US$50 billion at the end of September. A figure nearer US$68–70 billion had been in circulation, built by comparing OpenAI with Anthropic on definitions that do not match. OpenAI, unlike Anthropic, does not count the gross value of sales through cloud partners. Nvidia fell 2.9% to US$230.48 and shed roughly US$170 billion of market value. Broadcom, AMD, Oracle and CoreWeave fell harder.

Friday the other half of the argument arrived, via Bloomberg: OpenAI has told investors it expects the run-rate to reach or exceed US$70 billion by year-end, with enterprise growth doing most of the work. Accounts of the FT report also carried third-quarter run-rate growth of 77% and enterprise run-rate growth of 107%. OpenAI declined to comment on the figures. Chips bounced.

The useful residue is not the percentage point. It is that a private company’s revenue definition moved public suppliers by hundreds of billions in a session. Anyone holding the picks-and-shovels complex is holding a mark that can gap on a leak. That is a position-size fact, not a view on the technology. The minutes had already said the equity rally is earnings, not multiples. A leak that reprices the earnings — even for a day, even on a definition — is the risk that framing implies.

5. The Household Declined the Invitation

The University of Michigan’s preliminary October sentiment index fell to 46.3, a five-month low, against a 47.6 consensus. Current economic conditions sank to 44.7, the lowest reading on record, from 50.9. The expectations index rose, to 47.3 from 46.3, the first increase since July. Year-ahead inflation expectations ticked up to 4.7% from 4.6%. Long-run expectations rose to 3.5% from 3.4%. The drop in the headline was steepest among lower-income households and among people with smaller share portfolios. Just over half of respondents in a gasoline special report said they would cut spending on goods, cars, dining and travel. About 31% expected to spend as usual.

That is the same split as the Russell. The S&P holder and the respondent with a smaller share portfolio did not have the same week. One marked a high. The other marked petrol, borrowing costs, and a current-conditions print with no precedent in the series. The forward-looking half of the survey did not confirm the collapse. The present-conditions half did. A preliminary reading can be revised. The survey window ran from 22 September to 5 October, so it closed before Friday’s bounce and before the chips retraced Thursday’s leak. A record-low current-conditions number is still the print that did not bounce when the index did.

For an Australian portfolio the exposure runs through both economies. The first is the capitalised earnings in the US holding. The second is the currency on that holding, and the Brent contract rather than WTI in the petrol price. The survey is not a trading signal. It is the check on whether the high was broad. It was not.

The Scorecard

US weekly performance is Friday 2 October close to Friday 9 October close. Oil levels vary by contract and snapshot; treat the intraday range as approximate. The OpenAI run-rate figures are as reported, then clarified by a separate account of what the company told investors. They are not an audited print.

The Read

Last week the problem was a bond rout that a jobs miss could not stop. This week the problem is the same price of money, with a new equity high painted over it. The Fed has hiked once and, on the minutes, most officials want to hike again by year-end. Wednesday asked for 5.36%. Friday closed at 5.24%, where last Friday’s rout closed. When the index makes a record and the discount rate does not ease, the discount rate is the number to watch.

For an Australian holder of US shares, the other number is the currency — this time with the sign flipped. Over the week the S&P 500 rose 1.2% in US dollars. The Australian dollar rose from 0.6933 to 0.6979, about 0.7%. An unhedged holder’s Australian-dollar return was roughly 0.5% (WSDU calculation, spot-to-spot, ignoring dividends and costs) while a hedged holder kept something close to the 1.2%. Last month the currency did the work: from 11 September to 2 October, a falling Aussie turned a roughly flat US tape into an AUD gain of about 4%. This week it gave a slice back. Same mechanism, opposite sign. It is a tendency, not a contract, and not a hedge policy. The arithmetic is the one set out in Currency Watch. The position-size point in section 4 is the one the picks-and-shovels notes keep running into.

What this week did not settle is the path from here. Three questions, in order. Did the discount rate actually ease? No — the close did, the high did not. Did the earnings story survive a private-company accounting leak? For a day, no, then yes — which means the mark is fragile. Did the household confirm the index? The expectations line flickered. The current-conditions line did not. An Australian holder who marks the portfolio off the S&P close and ignores the 10-year, the Brent settlement and the AUD fix is marking the wrong three numbers.

Data Sources:

• Associated Press, 9 October 2026 (Friday index closes, weekly moves, 10-year at 5.24%).
• NBC News and CNBC, 7 October 2026 (10-year intraday high near 5.36%, highest since 2002).
• ICE BofA 7–10 Year US Corporate Index, via FRED (yields above 6% through 8 October).
• Federal Reserve, minutes of the 15–16 September 2026 meeting, released 7 October 2026.
• Financial Times and Reuters, 8 October 2026 (OpenAI September run-rate near US$50 billion); Bloomberg, 9 October 2026 (year-end run-rate pointed at US$70 billion or more). CNN, 8 October 2026 (Nvidia −2.9%).
• University of Michigan, Surveys of Consumers, preliminary October 2026. Bloomberg, 9 October 2026 (consensus).
• Reserve Bank of Australia, 4pm exchange rates, 9 October and 2 October 2026.
• LSEG I/B/E/S, 9 October 2026 (Q3 earnings growth 30.6%; 26.5% ex-energy).

Oil levels vary by contract and snapshot. OpenAI figures are as reported, not an audited print. The unhedged AUD return is a WSDU spot-to-spot calculation and is not a forecast.

Wall St. Down Under | Australia

Subscribe | wallstdownunder.com.au

Disclaimer: Wall St. Down Under is an independent financial newsletter for informational and educational purposes only. Nothing published here constitutes financial advice, a recommendation to buy or sell any security, or a solicitation of any investment decision. Always do your own research and consult a licensed financial adviser before making investment decisions. Australian investors should consider their own financial situation, objectives, and risk tolerance. Past performance is not indicative of future results.