THE DEAN’S LIST: The AI Trade Gets The Bill.

Open your brokerage app and one corner of the market looks ugly. The S&P 500 fell 1.95%, the Nasdaq Composite fell 4.48%, and the Dow rose 0.67%. Chips did the damage. Hot inflation data tested markets’ resolve, oil jumped again, and bond yields fell anyway. The Single-Digit Millionaire portfolio lost 1.65% and beat the S&P 500 fund it holds by about 0.3 points. The scorecard is at the bottom of this letter.

Dean’s note:
If you own an S&P 500 fund, you gave back the previous week’s gains and then some. The index closed Friday at its lowest level of the week. The Nasdaq Composite did the damage, and one weekly review had the chip index down about 11.4%, with Friday its eighth straight loss. Money moved elsewhere. The portfolio dashboard’s sector table shows energy up 4.86%, consumer staples up 5.02%, and health care up 3.45% for the week, against technology down 7.70%. Equal weight, which gives every company the same weight, also beat the regular index by a wide margin. That is money changing seats in a way we’ve been expecting. This is why we hold equal weight in the portfolio.

Financing and valuation worries drove most of it. Investors again asked who pays for the AI buildout and whether some borrowers can carry the debt. Gartner put numbers on that worry Wednesday. CoreWeave, an AI cloud provider with a heavy debt load, took the worst of it. TSMC’s reported results and outlook stayed strong the next day. A great company and a great stock are not the same thing. Great technology and a great balance sheet are not the same thing either. Dive one has the detail.

The bond market did something odd. Inflation data ran hot, and yields still fell across the curve. The long bond is still the tell, and this week it said two things at once. The inflation scare is real, but global money wants a safe place to hide. Our read is that the second force won the week. Dive two sorts out the competing causes.

September CPI came in hotter than forecasts, and producer prices came in far hotter. Gasoline drove more than half of the monthly CPI increase, but producer prices showed pressure beyond fuel. We expect inflation to cool in the second half, and this week leans against that call. Core inflation, which strips out food and energy, was a smaller miss. Our call now rests on oil.

The Fed is in its quiet period before the October 27 to 28 meeting. Chair Kevin Warsh spoke Friday and declined to preview the decision. A hike got less likely. That is not a cut. Dive three has the numbers and the dates.

Oil jumped again. Over the weekend, Iran’s Revolutionary Guards said another tanker hit a mine, and Foreign Minister Abbas Araghchi repeated Iran’s seven-day offer to reopen the strait. Washington wants nuclear terms on the table from the start. Our Hormuz view is unchanged. Dive four says what would change it.

A quiet start, a hot Wednesday, and a Friday that sent chips to eight straight losses.

Monday (October 12): The bond market was closed for Columbus Day. Stocks drifted. The S&P 500 slipped 0.06% to 7,807.24 and the Dow rose 0.38%. Broadcom CEO Hock Tan compared many AI investors to casino gamblers at a conference, and the PHLX chip index fell 2.67%.

Tuesday (October 13): Bank earnings began. JPMorgan missed on profit and revenue. Goldman Sachs beat estimates by a wide margin. Johnson & Johnson fell 5.73%. The S&P 500 finished flat at 7,805.45.

Wednesday (October 14): September CPI came in at 3.7%, a tenth above forecasts. Gartner issued its AI infrastructure warning. The S&P 500 fell 1.11% to 7,718.94, the Nasdaq Composite fell 1.88%, and the Russell 2000 fell 1.93%. Oracle lost 8.83%.

Thursday (October 15): Producer prices jumped 1.3% for September. Retail sales rose 0.8%, and weekly jobless claims were 199,000. Stocks bounced anyway. The Dow gained 1.08% to 52,177.69 and the S&P 500 rose 0.33% to 7,744.06. Nvidia CEO Jensen Huang defended CoreWeave, and CoreWeave shares rose 5.94%.

Friday (October 16): The S&P 500 fell 1.09% to 7,659.44 and the Nasdaq Composite fell 2.14%. 68 of the 75 tech stocks in the S&P 500 declined, while energy led on higher crude. Warsh spoke in Bangkok.

Monday (October 19, yesterday): U.S. stock futures are higher early Monday and oil is lower. Over the weekend, Iran reported more tanker mine strikes, unverified. China reported third-quarter growth of 4.5% from a year earlier, up from 4.3% in the second quarter. Trump’s Friday order waiving specified new Graham Act measures on Russia is in dive four. Meta, Tesla, IBM, Texas Instruments, and Lam Research report Wednesday after the close. 

The week had a theme, and it was the cost of the buildout. Monday, Hock Tan, the Broadcom CEO, said many AI investors act like casino gamblers and questioned whether loss-making AI labs can keep funding themselves. That is a paraphrase from press reports. Wednesday, Gartner said more than $3.5 trillion of AI infrastructure spending could be exposed to write-downs by the end of 2030 if providers keep spending at the current pace. It also said over 80% of neocloud providers, the newer companies that rent out AI computing capacity, could stop operating independently by 2030. That can mean being bought or merged, not only failing. Those are forecasts built on Gartner’s assumptions, including a 12% return hurdle.

CoreWeave fell 15.46% Wednesday and another 11.91% Friday, with investors focused on its debt load. Reuters reported Thursday that some investors in Anthropic’s planned November share sale were discussing a valuation near $1.5 trillion, below the roughly $2 trillion floated earlier, after third-quarter revenue grew 45% from the prior quarter. The PHLX chip index finished the week down about 11.4%. TSMC’s reported results and outlook, released Thursday, remained strong.

Dean’s note:
Keep three questions apart. Is AI demand real? TSMC’s results and guidance say demand for its chips is. Can every company building the capacity earn its cost of capital? Gartner says many will not, and that is a forecast, not a fact. And what does the stock price already assume? That one changes every day. This week the selling hit the borrowers and the chip stocks together, while TSMC’s results and guidance stayed strong. Our view: the buildout is real, the financing will sort winners from losers, and we may see more weeks like this one. That is partly why we own the S&P 500 fund, equal weight, and foreign stocks together. We have watched capex booms before. The tech usually wins. The lenders and the latecomers do not always.

September CPI rose 0.6% on the month and 3.7% from a year earlier, against 3.6% expected and 3.4% in August. Gasoline rose 9.1% and made up more than half of the monthly gain. Core CPI, which leaves out food and energy, rose 0.3% on the month and 2.6% on the year, against 2.5% expected. Producer prices rose 1.3% on the month, about double forecasts of 0.5% to 0.6%. Excluding food, energy, and trade services, they rose 0.5% on the month and 5.0% on the year, so the pressure was not only fuel. Retail sales rose 0.8% against 0.3% expected. Gas station sales rose 6.4% and sales excluding gas rose 0.2%. The control group, which feeds GDP estimates, rose 0.1%. Spending rose sharply at gas stations and only modestly elsewhere. These are nominal dollars, not inflation-adjusted volumes.

Then the odd part. Treasury’s daily par yields fell for the week. The 2-year ended at 4.74% from 4.80%. The 10-year ended at 5.09% from 5.24%. The 30-year ended at 5.43% from 5.60%. The 10-year sat 35 basis points above the 2-year. Plausible causes include a flight from tech stocks into Treasuries, lower hike odds after Warsh spoke, growth worries, and a lower term premium, which is the extra pay investors demand for holding longer debt. A single week can never separate them.

Dean’s note:
Our read: we think a risk-off bid for Treasuries outran the inflation scare, and that bid can reverse. Watch the long bond for the tell. Now the inflation call. We expect headline CPI to run below August’s 3.4% by the report published in December, which covers November. This print leans against us, and the weak point is plain. The call leans on oil, and the core PPI reading is the harder fact for us, because it shows pipeline pressure beyond fuel. If crude stays near $100, we will probably miss the mark. If a Hormuz deal is on the table, gasoline can unwind fast. Stagflation means stubborn inflation alongside weak or stagnant growth, often with rising unemployment. Low claims and a 0.7% gain in September industrial production, above the 0.4% forecast, say we are not there. CPI is last month’s newspaper. The market trades on what it expects next.

The Fed meets October 27 and 28, and the decision lands Wednesday, October 28. It went quiet Saturday, October 17, so no Fed speakers this week. The next CPI report is November 10, after the meeting. CME FedWatch odds of an October hike jumped from about 21% before Wednesday’s CPI to 54% after it. They fell to roughly 28% by Friday morning and to about 15% after Warsh spoke Friday. Polymarket, a prediction market and not CME, had the October hike near 14% and a December hike near 79%. All of these move daily.

At an IMF event in Bangkok Friday, Warsh said, “I don't prejudge meetings.” He said developments had broadly matched the Fed’s expectations since the September hike. He called AI “probably the most consequential change” to the economy in his lifetime. He said the transition could be “very bumpy,” with employment effects over the next 12 to 18 months uncertain. The Fed targets PCE inflation, and its September median for 2026 was 3.7% for PCE and 3.4% for core PCE.

Dean’s note:
Precision pays here. A hike got less likely Friday. A hike did not become a cut. Our read is that the Fed waits in October and that December is the live question. That depends on oil and on November 10’s CPI. Warsh gave us half of our AI view. He sees big productivity gains and AI spending that lifts demand before the benefits arrive. He did not sign on to the jobs half. He called the employment effects uncertain, and he is right. Our view is that AI can lift profits at some firms without adding jobs there. 

Sunday, Araghchi said Iran’s offer stands. The strait would reopen on the sixth day if Washington ends its blockade of Iranian ports, lifts sanctions on Iranian oil, and agrees to a ceasefire on all fronts, with talks starting on day seven. The U.S. wants its nuclear demands on the agenda from the start, and Vice President JD Vance has called for a meaningful cut in Iran's enrichment capacity, per Fox News. Trump warned the U.S. could “soon finish the job” unless Tehran changes course. These are statements, not an agreement. Separately, Trump signed an order Friday waiving specified new sanctions, restrictions, and tariffs on Russia under the Graham Act. It keeps earlier Russia sanctions in place and sets no end date. Rubio described a review of up to 60 days. An October 9 Treasury license covers specified Russian diesel transactions. Both are oil-price stories more than Hormuz stories.

Saturday, the Revolutionary Guards said a supertanker hit a mine and caught fire. Sunday, Iranian media said another tanker exploded after striking a mine. Neither claim has been independently verified. The UK Maritime Trade Operations center counted 10 reported ship attacks in the week ending October 16. CENTCOM says the strait is open to all ships except Iran’s and that about 97% of ships attempting passage transit safely. That is the share of attempts not hit. It is not a count of traffic against normal. If we get a three-way meeting between Putin, Xi and Trump in the meantime to discuss Ukraine, bet your bottom dollar Iran will be on the menu as well. 

Dean’s note:
Our forecast, labeled a forecast: a brokered arrangement that reopens the strait before the November 3 midterms is possible, and we still rate it less likely than not, meaning below 50%. By reopen, I mean commercial transits measurably returning toward normal. This week gave the forecast a headline and little else.

A seven-day offer is a list of conditions, and the sides disagree on what comes first. That is a sequencing gap, and reported mine strikes make it wider. Evidence that would raise the odds: a public agreement on sequencing, and weekly transit counts rising. Evidence that would lower them: more ship attacks, or Saudi retaliation that widens the war. Watch weekly transit and cargo volumes, not the speeches. Meetings to resolve Ukraine without Zelensky don’t seem useful, except to speed up the resolution of Iran. But we don’t know how many legitimate closed-door offers for peace he turned down to date. Maybe some, maybe none. What we do know is his relevance is fading on the global stage.

JPMorgan missed on both lines Tuesday. Profit was $5.57 a share against $5.90 expected, and revenue was $50.17 billion against $51.22 billion, per Fortune. Excluding $658 million in securities losses, profit was $5.75. Expenses rose 12% and fixed-income trading fell 4%. Goldman Sachs earned $17.01 a share against $13.02 expected, per estimates cited by Fortune, on revenue of $18.68 billion, with equities revenue up 72%. Wells Fargo missed on revenue and beat on profit; Citigroup beat, and Bank of America missed, and its quarter included a $1.2 billion FinCEN penalty. Separate that one-time charge from operating results before judging the miss. Wells Fargo cited low loan delinquencies.

TSMC reported third-quarter revenue of $46.72 billion with a 66.4% gross margin and guided fourth-quarter revenue to $52.3 billion to $54.3 billion. ASML raised its full-year sales outlook. The chip index still fell hard on the week.

Dean’s note:
 Read the line item, not the headline. Goldman’s quarter was a trading quarter, driven by equities. That is a profit story, not a read on the economy. JPMorgan’s misses look more tied to costs, trading, and the securities loss than to broad credit deterioration. It still booked $2.7 billion in credit costs, including $2.2 billion of net charge-offs and a $539 million reserve build, so credit stays on the watch list.

Chips are the same lesson in reverse. Strong results and a falling stock are not a contradiction. Investors may have been reassessing valuation, financing costs, or future returns, and a falling price does not tell you which. A profit miss and a demand miss are not the same animal. Wednesday, keep Tesla’s margins and cash flow apart from its deliveries.

A two-percent dip is not a verdict. Know which evidence you are waiting for.

Here is what I am watching this week.

•  The long bond. It rallied in a week of hot data, and we think a risk-off bid did it. If yields rise again while stocks fall, the adult in the room is saying something different. Wednesday’s 20-year bond reopening tests the long end.

•   Wednesday’s earnings after the close. Meta’s revenue against its spending, Lam Research for chip-equipment demand, and Tesla’s margins. Earnings are evidence.

•  Hormuz transit and cargo volumes, and how the Graham Act waiver and the diesel license play out in fuel prices. They say more than the speeches.

•   Friday’s flash PMIs, including their price gauges, and the University of Michigan’s final October reading at 10 am ET. The last reading had year-ahead inflation expectations at 4.7%. That is a survey of what consumers expect, not what prices did.

•  Where your plan, your cash flow, and your emergency reserves support it, a choppy week is not, on its own, a reason to stop contributing to your 401(k). The 2026 elective deferral limit is $24,500. If you are 50 or older, a standard catch-up of $8,000 is generally allowed. A super catch-up of $11,250 replaces it, taking the total to $35,750, but only if you turn 60, 61, 62, or 63 during 2026 and only if your plan permits catch-up contributions.

If your 2025 Social Security wages from that employer topped $150,000, your catch-up generally has to go in as Roth. If your plan does not offer a Roth option, ask whether you can make catch-up contributions at all. Confirm your eligibility with your plan administrator.

The VIX closed Friday at 19.99, up from 14.84. It is the market’s expected 30-day volatility for the S&P 500, expressed as an annualized percentage and implied by option prices. It estimates the size of moves, not their direction. It does not directly measure oil-market or credit risk.

Mix matters. Through Friday this year, emerging markets (the EEM benchmark fund) are up 14.88%, and the Nasdaq 100 (QQQ) is up 15.54%, ahead of the S&P 500 fund’s 11.88%. Equal weight (11.76%), small caps (9.86%), and developed foreign stocks (the EFA benchmark, 7.83%) trail it. Those are the dashboard’s benchmark funds, not all of our holdings. This week, equal weight gained 0.49% while the Nasdaq 100 lost 5.52%. Our emerging markets holding, the best sleeve this year, lost 5.66%. No sleeve wins every week. That is why we hold several.

- Dean

P.S. The number I keep coming back to is $8.5 trillion. Gartner says AI data center owners would need about that much cumulative revenue through 2030 to cover announced spending, financing costs, and a 12% return. Gartner’s own forecast of AI revenue attributable to them is no more than $3.6 trillion, and it expects less than half of AI revenue to go to providers that own significant infrastructure. It is a forecast with assumptions, and forecasts like it have missed in both directions. The gap is the question. It is not the answer.

And one more thought. Size the dip. The S&P 500 sits about 2% below its record close and is still up 11.9% this year, per the Associated Press. The Nasdaq Composite is about 5.3% below its October 6 record close of 27,599.79. A 10% drop is the conventional line for a correction, and the index remains short of it, though it is more than halfway there. Chips are another story. SOXX, an ETF that holds chip stocks, closed Friday at $493.39, down about 16.2% from its October 2 close of $588.90. That group has crossed the line. Do not confuse a correction in one group for a collapse in the market. Money is changing seats, and it has not left the building. Stay invested. Stay selective.

👉 Using the portfolio dashboard's data through October 16, with VOO's Friday close corrected to the market’s $701.61, the portfolio fell 1.65% for the week, against 1.95% for VOO, the S&P 500 fund that makes up half of it. Equal weight (up 0.49%), developed foreign stocks (up 0.01%), and cash (up 0.08%) cushioned the week. Emerging markets (down 5.66%), small caps (down 3.05%), and Ethereum (down 2.64%) held it back. Those are asset returns, not weighted contributions. For the year, the portfolio is up 8.54% against 11.88% for VOO, a gap of 3.34 percentage points. Among our holdings, only emerging markets lead VOO this year. A mix of assets can help or hurt over any stretch, and diversification does not eliminate losses in a broad selloff. Gold can fall hard. Crypto can halve. Cash costs you when stocks run. Small caps and foreign stocks can lag for years. See every holding, every weight, and the reasoning behind them at My Single-Digit Millionaire portfolio.


This newsletter is for informational purposes only and does not constitute investment advice. Past performance is not indicative of future results. Consult with a qualified financial advisor before making any investment decisions.