← News & Research
RESEARCH DIGEST · FRIDAY 25 SEPTEMBER 2026 · 3:11 AM EDT
Written by AI, which can make mistakes. Not financial advice.

These pages are written by AI from podcast transcripts, market news, macro releases and prediction-market pricing. It can misunderstand what a speaker meant, attribute a view to the wrong person, or get a figure wrong.

Where a ticker carries a + or −, that is our reading of the direction a source leaned — not a recommendation from AlphaDrift, and not necessarily a formal recommendation from the speaker either. Check the original before you rely on any of it, and speak to a licensed financial adviser about your own circumstances.

Odds of a third Fed hike this year reached 41%, while a US diesel export ban lost half its odds in a day

6 videos5 news & macro sources9 prediction markets5 things worth your time

1Fed officials are now talking about another hike, and the odds of a third this year reached 41%

Two Federal Reserve officials spoke on Thursday, and both leaned the same way. New York Fed President John Williams said it is "reasonable" to expect another rate hike by year-end, according to CNBC. Philadelphia Fed President Anna Paulson said "modest" rate moves are likely ahead to bring inflation down. The Fed's target range already tops out at 4.00%, after its 16 September hike.

Prediction markets moved with them. Polymarket prices a 25bp hike at the October meeting at 66.5%, up 2 points on the day and 15 points on the week, on $460k of 24-hour volume in that contract. Kalshi's matching contract last traded at $0.67, up from $0.66, with about 498,000 contracts open. The larger move was in the year-end count. Polymarket's contract for three hikes in 2026 rose 5 points to 41%, and 22 points on the week. Two hikes slipped 4 points to 49.5%. That market is thinner, at about $37k of daily volume, so its direction carries more weight than its size.

The long end kept rising. The Fed's own data put the 10-year Treasury yield at 5.11% on 23 September, up from 4.96% a day earlier. The 30-year reached 5.40%, from 5.29%. The gap between 10-year and 2-year yields widened to 0.31 points from 0.26, so long rates are rising faster than short ones. The long Treasury fund fell another 1.3% on Thursday to about $79.43. Stocks held flat, with down 0.1%.

The so-what is that a third hike is no longer a tail. It is close to an even bet, and two Fed voices on the same day made it easier to price.

2A 5.4% long bond is low by historical standards, which argues against a quick reversal

Howard Marks, co-chairman of Oaktree Capital, told Prof G Markets that today's rates are not high. "A 5.3 or 4% 30-year bond is very low relative to history," he said. It is "only high relative to recent history", and he argues the near-zero years from 2009 to 2021 were the aberration. The statement he would make with the most confidence is that rates "probably will not be going down much if at all" in the next year or two.

He gave three reasons for the rise. Inflation has sat above the Fed's 2% target for about five years, so lenders want protection. The deficit is near $2 trillion a year at a time of low unemployment, with no emergency to justify it. And AI is pulling hundreds of billions of dollars from the debt market, competing with the Treasury for the same capital. He expects rates to keep rising slowly rather than break through a failed auction.

That puts him against Treasury Secretary Scott Bessent. Bessent has said yields do not reflect the economy's fundamentals and that the inflation is about the war. Marks said he had not seen anybody agree with that. The Treasury's own tool is not settling it either. Reuters reported the Treasury accepted $4.078bn in a buyback operation on Thursday. Ross Givens, on his channel, put the earlier $6bn long-bond buyback at about 0.02% of a $32 trillion market.

Marks did not tell viewers to sell stocks. He said the fiscal problem is not a company problem. If anything, it argues for holding fewer dollar-denominated assets, and even then "I wouldn't do very much." The disagreement has a test. If Bessent is right, a Hormuz deal and lower oil should pull long yields down. If Marks is right, they should barely move.

3A US diesel export ban lost half its market-implied odds in a day

Polymarket prices a US diesel export ban announced by 31 October at 11%, down 10.5 points on the day, on about $26k of volume. The 30 September leg sits at 4.5%. Reuters had already reported on Wednesday that the White House denied a ban was under consideration.

Ross Givens, on his channel, laid out why the idea keeps returning. Diesel sells for more than $6.50 a gallon against about $4.50 for gasoline, a $2 gap that he says did not exist a year ago. By his figures the US makes about 5.3m barrels of diesel a day, uses about 3.6m, and exports roughly 1.8m. Drone strikes on Russian refineries and war damage in the Gulf left Europe short, so American diesel is priced against buyers in Rotterdam. He puts the diesel refining margin above $100 a barrel. He named Valero , Marathon Petroleum and Phillips 66 as the companies collecting it.

His case for a ban is political timing. The midterms are on 3 November, and both candidates in an Iowa House race have called for one. He said news broke on 23 September that the administration was preparing a 90-day ban. That account conflicts with the White House denial Reuters reported the same day. His trade was in futures, not stocks. He went long London gas oil, a bet that European diesel rises if American exports stop, and found almost no retail market in US heating-oil options.

The refiners traded as if the threat had faded. rose 1.9% on Thursday and 0.6%, while slipped 0.2%. If the ban odds keep falling, the $2 gap and the refiners' export margin stay in place into the election.

4Oil rose on a strike at Saudi Arabia's Hormuz bypass and fell on a phased reopening plan, in one session

On Thursday afternoon Reuters reported that the Houthis claimed strikes on Riyadh and on Aramco facilities in Yanbu. Saudi Arabia said it intercepted the missiles. Yanbu matters because it is the Red Sea outlet for the pipeline that lets Saudi crude bypass the Strait of Hormuz. Reuters separately reported Red Sea insurance costs soaring on the same route. The United States Oil Fund rose 2.9% on the day. Prof G Markets put Brent above $103, up more than 40% since the war began.

Later on Thursday Reuters reported, citing sources, that the US and Iran are discussing a phased deal to reopen Hormuz and end the US blockade of Iranian ports. Iran's president said it is up to the United States to choose to end the war. Overnight, Reuters reported oil falling on truce hopes.

Prediction markets moved a little, not a lot. Polymarket prices Hormuz traffic back to normal by 31 December at 23.5%, up 2 points, on $186k of volume. The 31 October leg is 8.5%. A contract on the Bab-el-Mandeb strait being effectively closed by 31 December sits at 18.5%, on $57k. Polymarket's September oil ladder cut the odds of WTI touching $100 this month to 21.5%.

Daniel Baer, interim president of the Carnegie Endowment, told Prof G Markets why a deal would not reset prices. Clearing the backlog and rebuilding trust in the strait takes time. Insurance or some form of tolling would add a lasting cost. He expects crude to stay above pre-war levels even after a full reopening. He also noted the President has said he expects a deal after the midterms, which he found hard to explain as a bargaining position. So the talks lift the odds of reopening, but the bypass route is now a target too.

5AI spending with revenue behind it was rewarded, and an unbuilt data-centre IPO was not

Meta rose 4.5% on Thursday to about $778. CNBC reported it is nearing its first new high in a year, on the success of its Muse AI agent. On the same day, Prof G Markets reported, citing the New York Times, that SB Energy has delayed its IPO. Bankers could not find enough buyers at a $50bn valuation.

The show's case against the valuation was specific. SB Energy has no data centres in operation, and fewer than a tenth of those planned are under construction. It claims a backlog above $400bn, but only about $1bn is expected within two years. More than 80% is due more than eight years out. Its existing solar revenue fell more than 8% last year.

The bull case on the chipmakers came from Business with Brian. He puts Nvidia at about 24 times next year's earnings against about 26 for Coca-Cola , with growth near 70% against 6.5%. Those are his figures, and they were not checked on the show. The hosts of The Compound and Friends called the big four hyperscalers' capital spending, estimated near $1 trillion next year against their free cash flow, their chart of the year. They asked what could end the rally, and their answer was the Fed raising rates. That links this claim to the first one. The market is still paying for AI that earns, and the thing most likely to test that is the hike path.

Wednesday's inflation and growth prints are the next test of the hike odds

Finnhub's calendar returned earnings only this run, so the economic dates come from the Fed's release schedule. The next data land on Wednesday 30 September: August personal income and spending, including the PCE inflation gauge the Fed targets, and the GDP report. Micron reports after that close, with analysts expecting about $32.32 a share. The September jobs report follows on 2 October. Before then, Nike and Carnival report on 28 September. If the PCE print cools while oil falls on the Hormuz talks, the third-hike odds are the first number to watch. If long yields stay put on that news, it is Marks's reading that holds, not the war explanation.

Mentioned today
SB Energyunlisted

How to read the tickers
TICKERa source leaned toward buyingTICKERleaned toward selling or trimmingTICKERmentioned, no direction givenNAMEunlistedprivate or pre-IPO — no symbol to chart

Sources scanned · Friday 25 September 2026

More research

How the analysis works

Views attributed to named sources are theirs, linked to the original in every case. AlphaDrift holds no position on the basis of anything published here.