1September's hike was priced out in a week, and a weak consumer did it
The deepest market on the Federal Reserve's September decision now prices no change at 73.5%, up 11 points in a week. The same market prices a 25 basis-point hike at 24.5%, down 11 points. That is $34.1m of volume on the event, with roughly $630,000 of liquidity sitting on the no-change leg alone — deep enough to take seriously.
The trigger has a date on it. Advance retail sales for July printed on 14 August and contracted. Reuters read it the same way the market did, running the headline "Oil prices rally, US data dents chances of Fed rate hike" the evening it landed. The mechanism is direct: a shrinking consumer means demand-side inflation pressure is fading, which removes the case for tightening into it.
The labour data underneath is doing something stranger. July unemployment printed 4.1%, improving from 4.2% in June and 4.3% in May. Payrolls went the other way — 158,858,000 in July against 158,881,000 in June, a fall of 23,000. The rate is improving while the level of employment shrinks, which usually means people are leaving the workforce rather than finding work. That is a soft labour market wearing a strong number.
The hike did not die, though. It moved. A separate contract on whether the Fed hikes at any point in 2026 still sits at 46.5% — a coin flip — even after falling 8 points on the week. Liquidity there is much thinner, about $202,000 against the September event's $3.5m, so treat the level as indicative rather than firm. The argument has shifted from when to whether.
2The truce is holding and the deal is dying — those are not the same thing
Two Iran markets moved in opposite directions this week, and conflating them would be a mistake.
The Israel–Iran ceasefire holding through 31 August prices at 95.5%, up 10 points on the week, on $184,000 of 24-hour volume. Meanwhile a final US–Iran nuclear deal by 30 September sits at 3.05%, down 4.5 points, and the year-end contract fell 10 points to 17.5%. The shooting pause is firming and the diplomatic settlement is being priced out, simultaneously.
The news flow explains why. The UAE says Iran attacked an ADNOC vessel in the Strait of Hormuz and has urged the waterway be reopened. Reuters reports Hormuz traffic slowing further after the US threatened more economic pressure, and US officials saying the naval blockade can be maintained "indefinitely". President Trump told Americans to accept high gasoline prices. Oil climbed more than a dollar on the tanker attacks, and European shares snapped a four-week rally as higher crude tempered otherwise strong earnings.
The distinction matters because the two contracts pay off through different channels. A ceasefire stops the escalation premium building. Only a deal returns sanctioned barrels to the market. Prices moved this week toward "no shooting, no barrels" — the combination that keeps crude bid on transit risk without any new war headline to point at.
That runs directly against the first claim. Claim one's dovish repricing rests on demand weakness pulling inflation down; this one is a supply-side impulse pushing energy costs up. Near-term the demand read looks likely to dominate, because the Fed market repriced 11 points on the retail print while oil was already rallying — traders saw both and chose the consumer. But that is a judgement about which force the market is currently weighting, not a resolution. A sustained crude move through the September meeting would test it directly.
3Record order books are trading at 40-to-57 percent discounts
Brian, on Business with Brian, spent a 24-minute video on a single screen: companies whose revenue is already contracted rather than forecast, and which the market has marked down anyway. He calls it buying receipts instead of guesses. Four of his five names fit that shape and the numbers behind them are unusually concrete.
carries $638bn of signed contracted revenue, up from about $138bn a year earlier. Roughly half of that leans on one customer, OpenAI, under a five-year agreement worth about $300bn — around $60bn a year, close to the size of Oracle's entire current business. Building the capacity to deliver it pushed quarterly construction spending from about $6bn to roughly $16bn, and swung free cash flow to about negative $24bn. The stock trades 57% below its high. His own framing of the risk is blunt: a contract is only as good as the customer, and this customer loses money and is not public.
has a backlog above $5.5bn, up by half in a year, with revenue nearly doubling last quarter, no net debt and return on equity around 40%. It trades about 46% below its high after operating margin slipped roughly four points — the result of bolting on a lower-margin electrical contractor, which management attributes to mix rather than deterioration. holds a record backlog above $21bn, up 30%, and beat and raised guidance, yet sits about 40% off its high after its smaller telecom division hit an air pocket.
is the odd one out: not a builder waiting to be paid but a cash generator, converting close to 80 cents of every revenue dollar into operating profit, above Visa's mid-60s. It fell more than 50% from its high, and dropped about 20% on the day it disclosed that the SEC had closed its investigation with no action — because the same report carried its first guidance miss since going public. Revenue still grew about 53%.
Every one of these carries a hedge, and the hedges are part of the call — detailed in the recommendations below. One caveat belongs with the claim itself: this is a single voice on a single day, and the video contains a disclosed paid segment for a gold developer, which is advertising rather than analysis and is excluded from the calls recorded here.
The testable part is narrow. The thesis says these drawdowns are about financing, mix and customer concentration, not about demand — and the order books are the evidence for that. Cancellations, not share prices, are what would falsify it.
What the sources recommended
Five single-name buys came out of one video, all from the same person, and every one arrives with a caveat attached by the person making it. There were no sell or avoid calls.
is the anchor, bought "slowly on weakness in small pieces" with the explicit rider that the chart is still broken and he is not calling a bottom. he holds "with a little more conviction than most on this list", while noting it is not statistically cheap at roughly 40 times earnings and that growth is set to cool from about 60% this year toward the low 20s next. he already owns in size and intends to add to, but only once it convinces him it has found a floor.
is the weakest of the five as a call — he says outright this is "not a backup the truck moment", that he is waiting rather than rushing, and that he will start small and add on weakness, citing real debt and a backlog weighted toward next year. he flags as the highest-risk name on the list and sizes accordingly, at roughly 46 times earnings, with short interest near 14% and climbing, and insider selling that included the founder chief executive taking out about $24m. Its appeal rests on being a neutral supplier of training data after Meta bought roughly half of rival Scale AI for about $14bn, which pushed other labs to move work elsewhere.
The consumer contraction gets its corporate cross-check this week
The retail sales print that moved September's odds was one number from one agency. This week the retailers themselves report, and they either corroborate it or complicate it.
reports Tuesday before the open. , , and all follow Wednesday, with and Thursday and Friday. That is most of the US consumer's discretionary and home-improvement spend inside four sessions, against a July print that said spending shrank. also reports Thursday, days after news that its AI models passed 3 billion downloads.
If the retailers' guidance echoes the July contraction, the case for a September hold hardens and the 73.5% gets more expensive. If they beat and guide confidently, the weak print starts to look like a single soft month, and that 46.5% chance of a hike somewhere in 2026 becomes the live question again.
