1The rate market prices a September increase at 30% and a cut at 1%
Jeffrey Schmid, president of the Kansas City Fed, called inflation "stubborn" and "sticky" this morning and said the policy rate is not restrictive. CNBC's write-up notes he stopped just short of calling for a rate increase.
The arithmetic behind that is on the tape. Core PCE printed 130.658 for July, released 26 August, which is +0.25% on June and +3.34% over twelve months. Headline CPI ran +3.30% over the same span. The effective fed funds rate is 3.63%. Subtract one from the other and the real policy rate is roughly a third of a point — restrictive is a hard word to defend at that level, which is Schmid's point.
Polymarket's September FOMC contract agrees and has for weeks. No change prices at 67.5%, up three points on the day. A 25bp increase prices at 30.5%, down four. Every cut leg combined sits near 1.5%. The book is deep: about $55m of total volume across the five legs and roughly $1.2m in the last 24 hours. The separate full-year contract prices no cuts at all in 2026 at 88%, up half a point, on $7.6m of volume.
The 24-hour direction and the level point different ways, and both are real. The hike leg fell four points today while the hold leg rose three — a dovish drift inside a hawkish regime, not a turn. Nothing in this window explains the drift: weekly claims printed 203,000 for the week to 22 August, near the low end of the last two months, which argues against a labour-market crack. The ten-year sits at 4.64% against 4.74% a week earlier, and the two-to-ten spread is unchanged at 0.47. Reuters reports the European Central Bank is set for a September hike of its own with no appetite to signal more.
What would settle it is Warsh. A CNBC survey published 26 August found 80% of respondents want the Fed Chairman to say more about his economic view at Jackson Hole, and Reuters reports Bessent and Warsh diverging on who should set the price of money. Until he speaks, the hold-versus-hike split stands where the market put it, and the one-in-three chance of a rise is the price of a 3.3% inflation rate meeting a 3.63% policy rate.
2The written-off half of technology moved further today than the winning half
closed yesterday at 205.75 and traded 252.17 this afternoon, +22.6%. is +27.6% at 172.24. is +19.1% at 225.73. , whose results after yesterday's close set the whole move off, is +9.1% at 228.77. The chip maker beat; the software names it dragged up beat harder.
That landed the same morning as an argument that the software complex is structurally impaired. Kai Wu, on the Intangible Economy strand of Excess Returns, describes technology as a K: "the software index is down 50% or whatever, and relative to the market the semiconductor index is up by 200%." His test for which side a company falls on is one line. "If your moat was your code, that's gone — AI has basically obliterated that moat." What survives, on his reading, is a moat made of something else: embedded enterprise workflows, customer relationships, regulatory capture. He is explicit that the bigger companies serving the biggest customers are the safer half of that trade.
Dan Rasmussen, of Verdad Advisors, takes the same observation somewhere harder. By his estimate more than 40% of private equity by 2021 was software or healthcare software, bought at multiples near 20 times EBITDA with about half the purchase price in debt. "If the public software companies are down 50 to 80% and these private-equity-backed things are 50% levered, that's an annihilation." His conclusion is a relative one and he states it plainly: over the next five years he would rather own the private credit than the private equity on the same assets, because "there's a chance the cash flows are fine for some of these businesses for a while — it's just the equity value and the valuation multiples have gotten absolutely smoked."
Both men are describing a multi-year re-rating, and one session does not refute it — a stock that fell 70% and rose 22% is still down 63%. But the two readings do point opposite ways, and the thing that separates them is measurable rather than rhetorical. If the moat really was the code, the beats now showing up in identity, security and CRM should decay into falling renewal rates and shrinking net revenue retention over the next few quarters. If the moat was the enterprise workflow, they will not. That number is reported, so this disagreement has an expiry date.
Note what the argument is not about. Neither man disputes the capex side: and SK Hynix and the chip suppliers are the half that has been working, and Wu's own framing has trillions flowing to data centres, utilities and power. Rasmussen's objection there is about crowding rather than mechanism — he puts the Goldman Sachs VIP index at roughly 90% correlated with the AI trade, and his read on the people running it is that they are "maximising expected value" by taking enormous risk on a future that may simply arrive late. His summary of what happens to capex booms comes in four clauses: "everyone always overinvests. The returns are always lower than people anticipated. The returns always come later than people thought. And there's this gap between the investment and the realisation that's always too long for most people to survive."
3A Hormuz agreement was announced and the market pricing it fell 14 points
Reuters reported on 26 August that Iran and Oman had reached agreements on shares of the Strait of Hormuz and its revenues, citing the Revolutionary Guards. A senior Iranian source told Reuters separately that the two are "still working on" the accord.
The market that prices exactly this event went the other way. Polymarket's "Iran-Oman Hormuz Agreement by September 30" contract sits at 42.5%, down 14 points in 24 hours on $41.8k of volume in that span and $202k lifetime. The 31 August leg is at 8.5%, down six points, on $38.6k of 24-hour volume. Neither book clears the $100k daily threshold where a price carries real weight, so read the direction rather than the decimal — but the direction is unambiguous, and it is away from the headline.
The deeper book says something narrower and harsher. "Strait of Hormuz traffic returns to normal by 15 September" prices at 1.35%, down 1.7 points, on $120k of 24-hour volume — the most liquid of the three contracts checked, and the one closest to zero. Reuters reported this morning that shipping traffic through the strait has risen slightly. Both can be true: traffic is recovering at the margin, and the market prices a return to normal within three weeks at 1.35%.
Today marks six months of the war, and Reuters' own framing of it is a "costly stalemate". The rest of the day's reporting fits that shape rather than the announcement's: OPEC+ losing market sway as China gains influence, oil companies avoiding blacklisted ships, Dubai airport traffic down more than 30% in the first half, British energy price caps up 4%, and Europe and Asia accelerating renewables because of the war. The consequence for anyone holding energy exposure is that the risk premium is not being unwound on an announcement — it needs the traffic data, and the traffic data has not moved enough.
4Risk appetite is priced at 268 times earnings in one market and 30% below par in another
Bryce Leske and Alec Renehan, on Equity Mates, walked through Shanghai's STAR 50 index, which they describe as China's Nasdaq-100 equivalent, 86% technology, up almost 30% this year and trading at 150 times earnings — against roughly 35 for the Nasdaq 100. New listings on that market have been debuting at an average of 268 times earnings.
The individual cases are more striking than the index. Unitree, a humanoid-robot maker, ran an IPO reported as 8,000 times oversubscribed, listed at 219 times earnings, and closed its first day up 460% — a $9bn company that ended the session near $50bn. , for comparison in the same segment, trades at 33 times. CXMT, China's largest DRAM maker and the world's fourth largest, raised $8.6bn and jumped 466% on debut to just under half a trillion dollars, briefly the most valuable company listed on the mainland. Their own caution is that Chinese listing rules push first-day pops by design, and Jay Ritter's data has these IPOs averaging 227% first-day jumps — so the pop is partly mechanical. It does not explain 268 times earnings at the offer.
At the other end of the same risk spectrum, the assets you cannot sell are marked down. Rasmussen and Richard Andis looked at London-listed vehicles that hold direct private equity fund interests — actual LP stakes, not manager stocks like Blackstone or KKR. Their realised volatility runs 24-25% annualised, "a little bit more volatile than small caps, which makes sense because it's levered micro cap". Those vehicles used to trade near par. They now trade at 30-40% discounts to net asset value, which Rasmussen argues "is probably where private equity should trade in aggregate".
So the same broad asset class — small, levered, technology-adjacent equity — is being marked at a premium of 268 times earnings where it is newly issued and at a 30-40% discount where it is redeemable. The variable that separates the two is not fundamentals. It is whether a price is set by people who can get out.
Two things worth holding onto from the same conversation, because they are the part with a long horizon. Rasmussen's largest position is Japanese sub-book small caps, on a mechanism rather than a mood: the Tokyo Stock Exchange requires companies trading below book value to get to book, and because Japanese book value is almost entirely tangible, that resolves into selling securities and non-core real estate and returning the cash. And the robotics cost curve Equity Mates traced is real regardless of the multiple attached to it — Unitree's humanoid went from $90,000 in August 2023 to $16,000 in May 2024 to $5,900 in July 2025, at a reported 60% gross margin.
What the sources recommended
Only three single-name calls came out of three videos, all from one person, and all three sit inside an argument about the same thing — the long end of the Treasury curve.
Ross Givens recommends buying , the zero-to-three-month Treasury ETF yielding about 4%, as the only government paper he wants held, and , Weyerhaeuser, the largest US timberland REIT, at $24 a share and "near its lowest price to book value in decades". The stock traded 23.84 this afternoon, so the level is current. His broader preference is a category rather than a list — gold, copper, energy, infrastructure and productive land, on the grounds that "the things that hold their value tend to be the things nobody can print more of".
Against that he recommends exiting and anything else tied to long-dated Treasuries, including target-date funds, balanced funds and the 60/40. His case: the US crossed $40tn of debt on 18 August, the 30-year auction six days before that cleared at 5.2% — the most expensive since 2001 — and the Treasury doubled its debt-buyback operations from $2bn to at least $4bn the morning after the 30-year pushed through 5.3%. He is careful about what is and is not happening, which is the strongest part of the segment: the Fed is buying roughly $10bn a month of short bills for reserve management, and that is not the same instrument or the same purpose as buying long bonds to suppress the long end. He expects the second thing eventually and says so as a forecast, not a fact.
Two caveats attach to those calls. The events he builds on are a week old and were covered here on 26 August from two other channels, so this is a reiterate of a case rather than new information — the specific numbers are his framing of it. And his own read of the mortgage pass-through has stalled rather than accelerated: he cites 6.65% for the week of 20 August against about 6% six months ago, and Freddie Mac's survey printed 6.66% this morning, one basis point higher over the week. itself was −0.1% on the session.
Nike came close to a fourth call and did not get there. Renehan noted is still down 75% from its high with its price-to-earnings ratio back near 19 or 20 from a peak of 70, and said "at some point it becomes a buy again, and I reckon it's probably pretty close to that". Close to a buy is not a buy, so it is recorded as a mention.
The thesis gets marked to market tonight, and again at Jackson Hole
reports after today's close, on a consensus of $0.939 a share and $2.76bn of revenue, and was flat at 244.70 into it — the one remaining large semiconductor print of the week, and the cleanest read on whether 's beat generalises or was specific. , , , , , and also report after the close, which puts four more enterprise-software results against the moat argument in claim two within hours. and reported this morning and both traded lower.
The bigger test is Warsh. With the September contract split 67.5% hold against 30.5% hike, and 80% of a CNBC survey asking the Chairman for more of his economic view at Jackson Hole, his remarks are the single input that can move that split more than a data print can. The claims above are priced on the assumption he does not.
