1A rate hike is 78% priced into an economy adding 50,000 jobs a month
Two venues agree, and neither is thin. Polymarket puts a 25 basis-point increase at the 15–16 September meeting at 78.5%, on $12.1m of standing liquidity and $7.4m traded in the last day. Kalshi's contract on the same decision last traded at 78 cents, with 6.9m contracts of open interest behind it. A week ago Polymarket had the same outcome near 48%, and the "no change" bracket has fallen 31 points over those seven days.
The repricing has a date on it. August consumer prices rose 0.40% on the month, which annualises to 4.9%, and 3.35% against August 2025. That print landed on 11 September. It followed a June reading that had actually fallen month-on-month, so this is a reacceleration rather than a continuation.
The same week's labour data is harder to square with a hike. Payrolls stand at 159.08m against 158.47m a year earlier — 603,000 jobs in twelve months, about 50,000 a month, a rate of 0.38%. Unemployment is 4.1%, unchanged from July.
Jim Paulsen, on Excess Returns, builds his case on that gap. He combines the two series into what he calls a job-market misery index — the unemployment rate minus annual job growth — and puts it at 4.1, higher than roughly 88% of post-war readings. The only higher readings come inside recessions or the first months of a recovery. His description of job growth as "about zero" overstates the print: 0.38% is very low against a typical 1.5% to 2%, but it is not zero, and August alone added 162,000. He also argues the causation runs opposite to the common read. Payrolls have led jobless claims for the past decade, so claims near 200,000 should rise to meet flat payrolls rather than payrolls rising to meet benign claims.
The bond market has already done much of the work the committee is being asked to do. The ten-year printed 4.95% on 10 September, up from 4.78% six days earlier and 4.55% in mid-July. The ten-to-two-year spread narrowed to 0.33 points by 11 September from 0.43 a week before. With the effective funds rate at 3.63%, the two-year sits roughly 90 basis points above it — most of a tightening cycle is priced before anyone has voted on it.
Paulsen's own forecast runs against that tape, and he states it plainly. He expects the ten-year to trade with a four handle again before it breaks five, because the Bloomberg economic surprise index built on hard data alone has led the bond market through this cycle and has rolled over. The market is at 4.95% and pricing a hike. One of the two is wrong, and the next payroll and claims prints are what settle it.
The consequence: a tightening decision is being taken against inflation data that is a month old and labour data that is not. Whichever series the committee weights on Wednesday reveals which reading of this economy it has adopted.
2The oil market has stopped pricing a quick reopening of Hormuz
The physical evidence arrived first. The IEA reported Saudi oil supply at a more-than-three-decade low after Houthi attacks. Shipping traffic through the Strait of Hormuz fell to single digits. Tanker rates hit record highs. The US average diesel price passed $6 a gallon for the first time, according to GasBuddy, and Saudi Arabia shut its East-West pipeline after drones launched from Iraq struck it. Brent and US crude both passed $100 on 11 September, and Reuters reported crude up more than 3% again on 13 September after fresh strikes on Saudi targets and on Hormuz.
Then the diplomatic route closed. Oman's foreign minister confirmed on 13 September that the meeting between Gulf states and Iran had been postponed. Polymarket's ladder on an Iran–Oman Hormuz management agreement fell across every rung within the day: by 21 September down 27 points to 12%, by 30 September down 23 points to 19%, by 31 October down 18 points to 39.5%. The nearest rung, by 14 September, is now 1%. At $250k traded it was the most active Middle East contract on the venue over the day.
The market's view of the strait itself is the number that matters most, and it is low and falling. Hormuz traffic returning to normal by 31 December prices at 17.5%, down 8 points over the week on $536k of standing liquidity. By 31 October it is 6.5%. Neither rung has a bullish tilt anywhere in it.
Crude's own distribution repriced to match. On Polymarket's September WTI ladder, a touch of $105 moved from roughly 18% to 77.7% over seven days, gaining 21.8 points in the last day alone. A $110 touch is 46%, up 36.5 points on the week; $120 is 16%. In the other direction a $90 touch fell 44 points to 46.3%, $85 fell 34.5 points to 27%, and $80 fell to 13%. A Saudi East-West pipeline restart by 30 September sits at 52.5%, down 33.5 points in a day, but that contract is thin at $62k of total volume and the level should be treated as weak.
One contract was excluded rather than reported. Polymarket's "US x Iran effective ceasefire begins by" ladder still quotes rungs dated 4, 7 and 11 September at 91.5% to 92.5%, dates that have passed without resolving, alongside 95.25% by 30 September. A 95% chance of a ceasefire inside two weeks cannot be reconciled with active strikes on 13 September, nor with the same venue's own oil ladder. The stale rungs are the defect this feed flagged on 11 September, and the contract is not evidence in either direction.
The consequence: the escalation premium in crude is not a headline that decays in a week. Through the fuel bill it feeds the inflation print pushing the committee toward a hike, which makes the first two claims here one chain rather than two stories.
3The AI build-out's funding has moved from cash flow to debt
Two people reached the same mechanism from different directions this weekend. They are describing one thing, not independently confirming it, and that distinction matters for how much weight it carries.
Paulsen tracks it through credit. CCC spreads and credit default swap spreads have widened this year while the S&P 500 has not reacted — a pattern that on his charts preceded a pullback the last three times it appeared in this bull market. His underlying point is about who pays. Technology capital spending used to be financed almost entirely out of cash flow, and the ratio of corporate cash flow to new-era investment spending has fallen sharply over the last two quarters.
Henry Jennings, on Equity Mates, arrives at the same place from the funding market. The large technology companies spent their cash piles, then raised equity, and are now issuing into the bond market, where they compete for the same buyers as a US government funding roughly $40 trillion of debt. He notes bond yields at highs not seen since 2007 and draws the consequence: more competition for capital means higher rates, and higher rates compress the valuations of exactly the companies doing the borrowing.
Jennings does not take that to a bearish conclusion. He separates this cycle from the dot-com comparison — those were "Mickey Mouse companies" reliant on clicks, where this is "real now" — and cites Anthropic's revenue figures as mind-boggling. Paulsen makes the opposite historical analogy. The last time a technology sector changed its stripes this way, he says, was the Bell operating companies laying the same fibre down the railroad beds across the country — "no one made any money on it."
Paulsen is the one with a number attached. He is explicit that he is not forecasting a bear market for the index. He expects a decline of 20% or more in technology and telecommunications, with the other seven sectors falling less, adding to a 10% to 15% overall correction. His earnings decomposition drives it. Forward earnings for technology and telecoms, just under half of index market capitalisation, have soared. Commodity-linked earnings, about 5% of capitalisation, have soared on oil. The remaining seven sectors are up about 6% annualised year to date, which after inflation of 3.5% to 4% is roughly 2.5% real. He expects the commodity leg to roll over regardless of the oil price, because profit growth needs prices to keep rising and crude has been flat since March even at $100.
The near-term catalyst came from the laboratories themselves. Bloomberg reported Sam Altman telling OpenAI staff he is open to slowing cutting-edge development, and Anthropic published an account of Claude being used for weapons, spying and cyber operations. A Bloomberg piece carried on Yahoo Finance reads those calls as a near-term weight on chipmakers and their supply chain with limited long-term effect. Nasdaq futures were down 1.39% before the US open, against 0.66% for the S&P — the read-through is concentrated in the AI complex rather than broad.
The consequence: the feature that separated this capital-expenditure cycle from 1999 was that it funded itself. That is the feature that has changed.
4Cash flow rose at four companies whose shares fell 19% to 42%
Brian, on Business with Brian, screened four names on three tests — free cash flow, sales growth and operating margin. All four generated more cash over the last twelve months than the year before while their shares fell.
GE Vernova passed all three: free cash flow of $12.4bn against $2.7bn a year earlier, sales up 13%, operating margin from 1.9% to 4.4%, with the shares 19% below their high. He puts the bear case himself rather than answering around it. The chief financial officer said $6.4bn of that cash is customer deposits reserving turbine slots, and guided the second half lower as a result. Against that he sets an order book of 116 gigawatts of gas turbine orders and reservations, 20 of them signed in a single quarter, targeting at least 125 by year end. Analysts are split: of 37 covering it, 18 revised in the last month, 8 up and 10 down, leaving the average almost unchanged.
Uber passed all three: free cash flow of $10.1bn, up 19%, sales up 17%, operating margin from 9.5% to 12.5%, shares 28% below their high. The overhang is Waymo, which ended the Phoenix arrangement inside the Uber app in June 2026 and is reported to be leaving Austin after 2028. The day it goes, a partner paying Uber becomes a competitor taking its rides. Uber's answer is roughly $10bn of spending: about $500m into Lucid alongside an order for up to 35,000 vehicles and a stake above 11%, up to $1.25bn into Rivian, about $500m into Nuro, $300m into Wayve, and 3,300 jobs cut, about 10% of the company. Brian names his own falsifier — if those commitments grow faster than cash flow, his first test stops passing. This is the second channel to call inside ten days.
PayPal failed one of the three. Free cash flow of $6.6bn, up 25%, against a market capitalisation near $45bn, and sales up 6%, but operating margin fell about 1.7 points to 16.4%. Management attributes that to deliberate spending to win back branded checkout. The shares are 32% below their high and sit below the $60.50 a share that Stripe and a private equity firm offered in late August before walking away, after which the stock fell 13% in a day. Of 43 analysts, 34 revised in the last month and 31 raised — yet the average price target sits a couple of percent above the current price, and 33 of the 43 rate it hold. Brian's caveat is his own: he would take a full position rather than an oversized one, and says he is wrong if the margin is still falling two quarters from now.
Intuitive Surgical passed all three tests and he still did not buy it, which is the most informative decision in the video. Sales grew 21%, operating margin went from 28% to 31%, free cash flow grew 62% to $3.2bn, and the shares are down 42%. Two things happened six days apart in July. The company beat its numbers but reported US procedure growth of 12% against a usual 14%, and the shares fell 14%. Then the FDA cleared Johnson & Johnson's Ottava robot for ten general surgery procedures, ending twenty years without a serious competitor in soft-tissue surgery. CNBC covered Ottava on 11 September as leading into a new market. Short interest in rose 5% to about 7m shares over the month, while it fell 38%, 22% and 10% in , and — the only one of the four where the bet against the company grew after bad news. He is waiting for $333, and says plainly that waiting is not selling.
The video carried a disclosed paid segment for a small gold and silver producer, read between the company sections. It is a promotion rather than something the sources recommended, and it is recorded as such.
The consequence: these four sit directly against the argument above them. Three are in the seven sectors whose earnings Paulsen puts at roughly 2.5% real — which is the point if the drawdown has already priced that, and the risk if it has not.
What the sources recommended
Five directional calls came out of three videos, and the split is unusual. Every buy came from one channel, every sell from another, and the video with the most detailed macro case named no stock at all.
The buys are , and , all from Business with Brian, all framed as quality bought into a drawdown, and each carries a falsifier from the person making it — second-half guidance on , fleet commitments outgrowing cash flow on , a margin still falling in two quarters on . He explicitly declined to buy despite it passing every test he applies, and treats rising short interest as his reason to wait.
The sells are both portfolio-construction calls from Henry Jennings of Marcus Today, reviewing a listener's holdings on Equity Mates, and both are hedged. On , the geared Nasdaq fund held on top of an ungeared one, he said the leverage stacking is "something that you may want to address" — and a host then suggested continuing to contribute to it to capture the upside, so the segment contradicts itself and the listener was left holding both views. On , at about 22% of the portfolio, he said he would hold a smaller Australian weighting than that, on the grounds that Australia is 2% of global market capitalisation. Both are Australian-listed funds. He told the same listener to hold the three individual stocks — Catapult, and , all ASX-listed — while disclosing that he had recently sold Catapult himself at a 20% profit.
Two conflicts of interest are disclosed and both bear on how to read that segment. Equity Mates states it is part of the Betashares group and the episode's segments were Betashares-sponsored, while several of the funds under review are Betashares products. Business with Brian ran a paid promotion for a listed gold producer inside a video about buying undervalued stocks.
Jim Paulsen named no single stock across 27 charts and an hour. His calls sit at the asset-class level — technology and telecoms down 20% or more, the ten-year breaking 4% before 5%, and a stretch in which bonds at a 5% yield could compete with equities — and an asset-class view is not a single-name call.
The hike thesis gets marked to market on Wednesday
The Federal Open Market Committee meets 15–16 September, and both venues put the decision at roughly 78% for a 25 basis-point increase. The 22% branch is the one that would move prices most, because it is the one fewest people are positioned for.
Two earnings reports land the same day and both read on the argument above. FedEx reports against a consensus near $4.05 a share, and freight volume is the cleanest read on whether the weakness Paulsen sees in the broad economy is showing up in shipments. Lennar reports against consensus near $1.32, with the ten-year at 4.95%; housing is where a 40 basis-point move in eight weeks shows up first. General Mills reports Tuesday, a read on whether food inflation is being passed through or absorbed.
Separately, Treasury Secretary Bessent said the administration would sanction an unnamed "large" bank on Monday. An unnamed target is a whole-sector risk until it is named.
