The week the question changed from inflation to the consumer

The week was built around one question: would inflation force another hike. By Friday the answer was effectively no, and a different question had taken its place.

1. The consumer, not the CPI, was the week's real news

July retail and food services sales fell 0.6% month on month to $763.6B, the largest decline since May 2025, against a +0.1% consensus; excluding autos, sales fell 0.3% versus +0.2% expected (Census). The University of Michigan's preliminary August sentiment index came in at 51.0 against 54.5 expected and 55.2 in July, ending two months of improvement, with weakness across income, age, education and political lines (InvestingLive).

The detail most reports skipped: year-ahead inflation expectations rose to 4.3% from 4.2% (Yahoo Finance). Households said they are spending less and expecting prices to rise faster, in the same survey.

Why it matters. Three soft prints in three days, CPI on Wednesday, PPI on Thursday, retail sales on Friday, took September hike odds down to roughly 32% to 33% from nearly 50% a week earlier (Kitco). The hike debate is largely over. But it ended because demand softened, not only because prices did, and those are two very different reasons for a central bank to stop.

So what (ABC). The regime stays risk-on, and Beta requires no action. What changes is what to watch: the next consumer print now carries more weight than the next inflation print.

2. The tape sold off and it still was not a risk-off day

The S&P 500 fell 0.17% to 7,785.76, the Nasdaq 0.28% to 26,729.16 and the Dow 0.20% to 53,732.41. Energy and basic materials led, technology and healthcare lagged. But the Russell 2000 rose 0.51% to 3,068.42 and the VIX fell 2.60% to 14.25 (Motley Fool via Yahoo Finance). The index still closed a third consecutive weekly gain, and Thursday marked its 27th record close of the year.

Meanwhile the long end went the other way: the Treasury sold 30-year bonds at 5.216%, described as the highest in 25 years, attributed to inflation, energy prices and the deficit. The 10Y sat in the 4.6% to 4.7% area with sources disagreeing on the day's direction, so we will not claim one.

Why it matters. Small caps up, volatility down, large-cap tech down is a rotation signature, not a fear signature. And a front end pricing a hold while the long end auctions at a 25-year high is a curve that disagrees with itself about the same economy.

So what (ABC). Read the shape of the curve, not just its level, from here. A pause caused by disinflation supports both Beta and Alpha. A pause caused by a tiring consumer supports the discount rate and threatens the earnings.

3. Broadcom fell 6% on a financing structure, not a business

Bank of America estimated that Broadcom's chip-financing vehicle could reach $370B of senior debt by mid-2029 at 20-gigawatt scale, including roughly $150B of issuance in 2027 alone; the stock fell 5.94% to $392.99 on a day with no company news (24/7 Wall St.). The $370B is not Broadcom's debt. Its disclosed obligation is a five-year backstop of lease payments with maximum exposure up to $29B on the initial transaction. Fundamentals were untouched: fiscal Q2 revenue $22.19B, +47.9% y/y, AI chip revenue +143%, $16B of AI semiconductor revenue guided for the current quarter.

NVIDIA, with a structurally similar roughly $500B financing framework alongside Goldman Sachs, Blackstone and Apollo, did not sell off, because no specific figure was attached. AMD rose 5.13% to $507.77 on Baird's Street-high $1,250 target. The semiconductor ETF moved only 0.7%, so this was single-name repricing, not a sector de-rating. Applied Materials extended Thursday's post-earnings decline, down 5.59% to $504.67.

Why it matters. Every layer of the AI stack proved demand this week. What is now being priced is how the buildout is financed, and much of that leverage sits in vehicles and lease commitments that credit spreads and CDS cannot see. That makes the repricing a step function that arrives whenever someone publishes a number.

So what (ABC). Screen AI exposure in the Alpha sleeve on three axes now: funding source, cash conversion, and financing disclosure. The number to underwrite at Broadcom is the backstop, not the revenue.

4. Gold rose, and it explained itself

Spot gold rallied to a session high of $4,387/oz on the retail sales miss, last quoted $4,380.88 (+0.68%) intraday, and was marked +0.24% at $4,373.48 at the equity close; silver traded $64.740, +0.66% (Kitco).

This is the same equation that made gold fall 1.31% on Thursday, running in reverse. Thursday: a flat PPI cut expected inflation faster than nominal yields fell, so real yields firmed and gold dropped even as the 10Y fell. Friday: the retail miss pushed nominal yields down while UMich inflation expectations rose to 4.3%, so real yields fell and gold rose. Two consecutive sessions, opposite directions, one mechanism.

The move was capped by oil, which firmed on Hormuz risk after two UAE tankers were attacked, with WTI near $82.80 and Brent near $88.50, and by the fact that gold has still not reclaimed the $4,448 resistance Kitco flags. Kitco's weekly survey has Wall Street turning fully bullish (Kitco), which is a crowding note, not an endorsement.

So what (ABC). Gold stays a Cash-sleeve hedge sized as optionality against a real-yield or geopolitical shock. Read it against nominal yields minus expected inflation, never against the nominal 10Y alone. A consensus that has gone fully bullish is an argument against adding into strength.

Housekeeping

Reddit rose 12.63% to $178.09 on news it joins the S&P 500 next week, a reminder that index inclusion is a flow event, not a fundamental one, and that passive Beta buys whatever the committee adds.

Takeaway of the week: the inflation question closed and the consumer question opened. We spent eleven days watching the 10Y for a breakout that never came, and the variable that finally moved the tape was household spending. Track the next retail sales and sentiment prints the way we tracked CPI, and watch the front end and long end separately, because right now they are describing two different economies.

Analytics & education, not advice. DYOR.