1Oil's war premium shrank after Washington chose diesel supply over new Russia sanctions, but Hormuz is still shut
On Sunday afternoon President Trump posted that he would not sign a Russia sanctions document "at this time". He gave two reasons. He does not want to drive Russia away from peace talks. He also said Ukraine's strikes on Russian diesel refineries are hurting world diesel markets, while Russia keeps shipping cheap diesel. Reuters reported oil falling on the waiver, and Turkey's energy minister said his country will get some diesel from a US-Russia supply deal.
The chain runs from refined fuel to crude. More Russian diesel on the water eases the tightest part of the energy market. That takes some pressure off fuel inflation. It also trims the premium in crude. US crude futures traded at $97.44 in Asian hours, down 1.6%, per Yahoo Finance. Bloomberg notes oil touched $110 last week for the first time since May.
Prediction markets moved with it. Polymarket's contract on crude hitting $100 in October fell to 72.5% from 86.5% a day earlier, on about $14,000 of 24-hour volume. The $110 strike fell to 15.5% from 22%. Those are thin books, so read the direction rather than the level.
The waiver does not touch the other supply problem. Polymarket gives Strait of Hormuz traffic only a 0.85% chance of returning to normal by 31 October, and 12.5% by year-end, on about $646,000 of liquidity. Iran is urging Washington to accept its proposal to reopen the strait, per Reuters. Egypt said it is ready to back Saudi Arabia against the Houthis. Peace in Ukraine is not close either: Polymarket prices a ceasefire agreement by year-end at 13.5%.
So the two forces pull opposite ways. Near term, the waiver dominates, because it changes how much diesel moves this month. The Hormuz closure is the larger loss of supply, and nothing this weekend made it shorter.
2The AI sell-off is now a test of whether revenue can outrun interest bills
Bryce Leske and Ren Tan, on Equity Mates, put a number on the worry. CoreWeave sold bonds due 2032 only four months ago. They now trade near 78 cents on the dollar, and the yield touched 15%. Its interest bill was about $640 million in the second quarter, roughly a quarter of revenue. Management guided that bill to $860–940 million for the third quarter, up about 40%. Revenue is guided up 37%. For now the two are growing at the same pace. The hosts' point is that the margin for error is gone if revenue slows.
The stress is spreading through new listings. Firmus pulled a planned $7 billion IPO at a $44 billion valuation. AirTrunk delayed its Singapore listing, citing volatility and soaring borrowing costs. The hosts counted one-day falls of 15% in CoreWeave, 10% in Nebius and 12% in Iren .
Demand is the second question. The hosts showed a chart of Anthropic's annualised revenue run rate. It climbed from $30 billion in April to $80 billion in August, then slipped to $76 billion in September. They read it as customers reining in token spending and switching to cheaper models. That is one chart from a private company, so treat it as a signal to watch, not a trend.
Paisley Nardini, head of investment strategy at Tema ETFs, on Excess Returns, framed the same problem at index level. She counts about $450 billion of AI-related debt issuance this year through September. A year ago, she said, hyperscalers had cash to spare; now their free cash flow is negative. She does not see a price bubble. The S&P 500 trades at roughly 19–20 times earnings, nothing like 1999. Her risk is an "earnings bubble": growth that stays positive but slows, which markets would reprice hard.
A clip from The Compound and Friends sized the downside. One panellist said an OpenAI bankruptcy would push the US into recession "immediately", while stressing he was not predicting one. The comparison offered was a General Motors failure, not a bank run. It would hurt suppliers and jobs, but it would not freeze deposits.
The damage so far is concentrated in chips. Yahoo Finance reports chip stocks on an eight-day slide that matches their worst streak since the financial crisis. The triple-leveraged chip fund fell 12.7%. Futures were higher early Monday, with the Nasdaq up 1.3%. That bounce shows the trade still has buyers. It does not answer the funding question, which only earnings can.
3Higher rates are squeezing everything under the index, yet the Fed is still priced to hold in October
Brian, who hosts Business with Brian, opened with the tension. The 10-year Treasury yield hit its highest level since 2002 last week. The S&P 500 fell only about 1% on the week, while chip stocks fell about 9%. His explanation is profits. Expected S&P 500 earnings are up about 10% since June, and prices only about 3.5%. So the market got cheaper as it rose. He also cites 72 companies guiding above forecasts this season, the most since tracking began in 2006.
He does not wave the risk away. Profit growth is expected to slow to about 3% by the middle of next year. He notes the S&P 500 was higher a year later in eight of the last ten profit-growth peaks. The damage came when profits actually fell.
Nardini sees the same split from the other side. Rates are not hitting the index, she argued, because they are hitting what sits under it. Smaller, capital-hungry companies have been "penalised" for months. She cited a figure that 73% of S&P 500 companies sat below their 50-day average in early October. Her view of the index's real breadth is blunt: its "effective number" of holdings is about 50. She expects a slow hiking path, and notes markets have usually gained in the twelve months after a first hike.
Ben Carlson, on The Compound and Friends, went further. He argued today's rates are worse for the economy than the zero-rate 2010s, because housing activity is "absolutely dead".
The data back the pressure but show some relief. FRED has the 10-year at 5.10% on 15 October, down from 5.21% two days earlier. The effective federal funds rate averaged 3.75% in September, up from 3.63% in August, reflecting the hike already delivered. Unemployment ticked up to 4.2% from 4.1%.
Here sources and markets disagree. Brian says the Fed "could" raise rates again this month. Kalshi prices an October hold at 86 cents and a 25-basis-point hike at 14 cents, unchanged on the day, with about 1.8 million contracts open on the hold. Polymarket has a hold at 87.5%, up a point, and a hike at 11.5%, down two. Cheaper oil makes the hold easier to defend. The decision lands on 28 October.
4In weight-loss drugs, Eli Lilly has won the share battle and price cuts are now the main risk
Answering a listener question, the Equity Mates hosts walked through Eli Lilly . Both own it. Lilly holds about 61% of the US GLP-1 market against Novo Nordisk at 39%. In the latest quarter, Lilly's revenue grew 48% and Novo's 7%. Mounjaro is now the world's best-selling drug, at $8.7 billion in the first quarter against $7.9 billion for Keytruda. About 11% of US adults now take a GLP-1, up from 3% in 2024.
The risk is price, and the hosts say so in the same breath. Lilly's realised prices fell 13% globally in the second quarter and 9% in the US. Internationally they fell 36%, after China added Mounjaro to its national reimbursement list. Generic semaglutide launched this year in India, Canada, China and Brazil, and Indian versions are reported at about 90% below Ozempic. Volume has outrun those cuts so far, and the hosts note falling manufacturing costs are lifting margins.
Valuation is the other check. The stock trades at about 40 times earnings. The hosts' reverse discounted-cash-flow work implies 9–11% annual profit growth for a decade. Actual profit went from $5 billion to $10 billion to $20 billion over three years, with $32 billion guided for 2026. Lilly reports at the end of October. That report will show whether volume is still beating price.
What the sources recommended
Six single-name buy calls came out of two videos, and every one of them came with a condition: wait for the next earnings report, keep the position small, or both.
Brian, on Business with Brian, screened for companies whose profit forecasts rose since July while their share prices fell. He named five. Vertiv makes power and cooling gear for data centres. Profit per share jumped about 60% last quarter, yet the stock is about 45% off its high. He would wait for its 28 October report and then buy gradually, small. nVent sells power and liquid-cooling equipment. He flags competition from Schneider Electric and debt taken on near 6% for an acquisition, and would wait for its report. Onto Innovation makes chip-inspection tools and has a record backlog above $1 billion. Its price has fallen below the bottom of his buying range, and he flags heavy exposure to memory makers. He would keep it small, after the report. Comfort Systems builds data-centre interiors and has a $14 billion backlog. He calls it priced near the top of its own range, and would wait for numbers due this week. Innodata labels and tests training data for AI models. He puts it in the "riskiest slice" only, because two customers make up about 71% of sales.
The Equity Mates hosts' Eli Lilly call is a reiterate, not a fresh one. Both already own it. Their stated rule is to hold while the thesis holds, and they would add if the price fell sharply.
This week's earnings mark the funding question to market
The arguments above get their first hard test over the next ten days. Capital One and General Motors report on Tuesday. Meta, Tesla, IBM, Lam Research and Texas Instruments report on Wednesday after the close. Comfort Systems reports on Thursday and nVent on Friday. Vertiv follows on 28 October, the same day as the Fed decision. Eli Lilly and Onto Innovation report on 29 October.
Two numbers matter most. One is whether data-centre backlogs are still growing as fast as the debt behind them. The other is whether the Fed's statement validates the 86% hold that prediction markets now price.
