Real wages fell again in September: inflation-adjusted hourly earnings -0.3% YoY, the 5th straight decline. The consumer is getting squeezed while the S&P 500 sits 0.5% from a record high. Which one breaks first?
10/10 The $22B 30Y auction this morning cleared at 5.618% with solid 2.54x demand. The bond rout has a ceiling for now, but 5.618% is still the highest long-bond yield in over two decades. Every mortgage, pension obligation, and small-cap valuation resets against that number. The bond market grades everything in real time now.
The bond market is breaking records nobody wanted broken. The 10Y hit its highest since 2002, the 30Y its highest in 24 years, and the Fed's minutes said another hike is likely this year. Here are 10 things that actually matter for your money. A thread:
9/10 The winners and losers are clear. Winners: savers earning 5%+ on cash, banks' net interest margins, energy ($XLE). Losers: growth stocks, small caps ($IWM), office landlords, anyone with a floating rate, and homebuyers. Yields this high pick winners without asking.
JUST IN: Oracle just lost 5% on somebody else's numbers.
$ORCL fell to about $136, down roughly 5% in a single session, after the FT reported OpenAI told investors its annualized revenue is near $50B, about $20B below the $70B run rate Axios reported on Sept 29. That is roughly $20B of Oracle's market value erased by news that isn't even about Oracle.
What it means for your money: Oracle is the landlord of the AI buildout. It has committed massive sums to datacenters built to host AI workloads, and the market is now asking what those commitments are worth if the biggest tenant's revenue growth is slower than expected. When the company valued at $1.4T reprices, the $427B company funding its infrastructure reprices too. $ORCL is down about 30% on the year for a reason.
The broader signal: skepticism over OpenAI and Anthropic valuations is now a trending topic across X. This is no longer one stock having a bad day. It is the market repricing the whole AI trade.
Watch whether the selling spreads to the other AI-capex names. The tenant missed. Now the market is checking the landlords.
This isn't a headline selloff, it's a leverage flush. $400M of levered longs liquidated in one hour, $1.1B across crypto. The mechanism: liquidated longs are forced sellers, so each leg down triggers the next. Tuesday's high was $86.6K, now $BTC is breaking below $82K, and $485M in ETF outflows shows the selling isn't just futures.
JUST IN: Uncle Sam just paid 5.618% to borrow for 30 years. That is the highest yield at a 30-year auction in over two decades.
The Treasury sold $22B of 30-year bonds at 1 PM ET. High yield 5.618%, up 31 basis points from September's 5.308%. Sounds terrible. But demand was solid: 2.54x bid-to-cover, above the six-auction average of 2.41x, and dealers got stuck with only 6.79%. The demand was real, so bonds caught a bid right after the results.
What it means for your money: a clean auction caps the bond rout. Winners are cash savers earning 5%+ and anyone who needed the long end to stop screaming. Losers are the same as ever: homebuyers with the 30-year mortgage at 7.63%, small caps ($IWM), and long-duration growth that keeps getting repriced by this yield.
Watch the tape. Strong auctions calm supply fear, and supply fear is what pushed yields to 2002 highs. If the week finishes without another scare, the market gets to breathe.
🇺🇸 US sells $22B in 30-Year Bond Auction 📊
📈 High Yield: 5.618% (prev 5.308%, six-auction avg 5.087%)
📉 Tail: +0.1bps (prev -2.7bps, six-auction avg -0.1bps)
✅ Bid to Cover: 2.54x (prev 2.61x, six-auction avg 2.41x)
🏦 Primary Dealers: 6.79% (prev 2.2%, six-auction avg 10.3%)
🎯 Direct Bidders: 20.89% (prev 18.3%, six-auction avg 20.6%)
🌍 Indirect Bidders: 72.32% (prev 79.5%, six-auction avg 69.1%)
Tail printed +0.1bps, a negligible concession after last month's 2.7bps stop-through, and the 5.618% yield is the highest at a 30-year auction since August 1999. BTC slipped from 2.61x but held above its 2.41x average. Indirects eased to 72.3% from 79.5% yet stayed above average, and dealers took 6.8% against a 10.3% average. Average auction.
Initial claims this morning is the highest-frequency labor read the Fed gets before its October meeting. The +29K September payrolls print already repriced the market once. If claims confirm the softening, the December-hike odds that the hawkish minutes just lifted start coming back down. One number, two policy paths.
OpenAI told investors it did $50B annualized. Last week the number was $70B. Somebody's math is $20B off.
What happened: Axios reported Sept 29 that OpenAI's run rate was approaching $70B, up 70%+ since July. Now the FT reports OpenAI's own disclosure to investors puts end-of-September revenue at about $50B annualized. That's a $20B gap between the signal and the books.
What it means for your money: OpenAI is in the market RIGHT NOW raising ~$30B at a ~$1.4T valuation. At $70B revenue that's 20x sales; at $50B it's 28x. The fundraising math just got harder, and the Anthropic race just got tighter. Anthropic filed to go public this fall after a $65B round at $965B, and it has been closing the revenue gap all year. For public markets: $MSFT is OpenAI's biggest backer and $ORCL is a core Stargate partner, so both have direct exposure if this repricing spreads.
Watch next: whether the $30B raise holds its $1.4T price tag, and what Anthropic's IPO filing says about its own run rate. The AI trade is moving from narrative to numbers.
JUST IN: Trump says the Iran crisis could be over "right after" the Nov 3 midterms, maybe even before, after what he called "very productive" talks with regional leaders at the UN. Markets spent this morning pricing $105 oil on the war premium. A credible deal timeline is the other side of that trade.
What happened: at a UN meeting with Erdogan and Gulf leaders, Trump said there's "considerable momentum" toward an Iran agreement. In the same breath he warned Iran faces destruction if no deal is reached, and The Atlantic separately reports he's weighing strikes before the election. Both headlines are live at once.
What it means for your money: this is the single biggest swing factor in oil right now. Brent jumped to $105 (+5%) on Hormuz, Houthi, and Iran escalation. A real deal timeline unwinds the war premium fast. Losers would be the long-oil trade ($USO, $XLE longs, anyone who bought the $100+ Brent fear). Winners would be airlines and trucking getting fuel-cost relief, and every equity index getting a breathing-again bid. The trap: "deal by midterms" and "strike before midterms" are both in the air. That is not de-escalation, it is optionality, and oil volatility stays bid.
Watch: whether tanker traffic through Hormuz actually restarts. Talk is cheap. Ships moving is the confirmation.
BREAKING: Oil just punched back above $105, up 5% in a day.
Brent is back above $105 and WTI is near $93 after Iran stepped up attacks on tankers in the Strait of Hormuz. A dozen attacks between Sept 28 and Oct 2, tanker traffic at its lowest since late July. Then the Houthis hit two Saudi airports, killing three people. And Axios reports the Pentagon told CENTCOM to finish preparations for resuming major combat operations in Iran.
What it means for your money: this is a supply shock, not just a headline. The EIA estimates 4.5 million barrels a day of Middle East production is shut in, roughly 4% of global supply, stuck behind shipping lanes that are now a war zone. Oil this high is a tax on everything and a problem for a Fed that just started hiking again. Winners: oil majors and energy ($XOM, $CVX, $XLE). Losers: airlines, trucking, and anyone with a commute.
Watch Hormuz shipping data and the White House on Iran. That decides whether $105 is the spike or the new normal.
Applied Digital beat big and the market shrugged anyway. Adjusted EPS of -$0.01 crushed the -$0.30 consensus and adjusted revenue tripled estimates to $300M, but the stock still closed down 6% on the day before bouncing in extended trading. The lesson: in a 5.3% world, even 322% revenue growth needs the tape's permission. $APLD
The auction that was supposed to fail just posted its best numbers in a decade.
The $39B 10Y reopening cleared at 5.300%, the highest auction yield since November 2000. But demand was the story: 2.77x bid-to-cover, the strongest since 2016, and foreign buyers took 80.3%, the most in a year. Primary dealers got a record-low 2.5%.
What it means for your money: 5.3% created its own demand. Big money is finding current yields attractive and unwinding shorts, which is why the 10Y fell to 5.28% after the print. The mechanism: strong auctions calm supply fear, and supply fear is what pushed yields to 2002 highs in the first place.
Watch the $22B 30Y today at 1 PM ET. That's the long end, and it's the one the market is scared of.
US Treasury yields continue to rise despite signs of strong demand at the latest auction:
The 10-year and 30-year Treasury yields both rose +4 basis points so far on Thursday, to 5.33% and 5.73%, their highest levels since 2002.
The $39 billion 10-year Treasury auction was awarded at 5.30%, the highest since November 2000 but slightly below the 5.33% market yield before the auction, signaling solid demand from investors.
The bid-to-cover ratio jumped to 2.77x, meaning investors submitted roughly $2.77 of bids for every $1 of Treasury securities available, its highest level since 2016.
Importantly, non-dealer investors took 97.5% of the auction, meaning pension funds, insurance companies, asset managers, mutual funds, hedge funds, foreign investors, and other institutions outside the primary-dealer network absorbed almost the entire sale.
For context, primary dealers are large banks such as JPMorgan, Citi and Bank of America that participate directly in Treasury auctions and can act as intermediaries between the Treasury and investors.
Yields above 5% are finally attracting serious institutional demand, but rising oil prices, inflation risks, as well as heavy government and corporate borrowing are still keeping long-term yields under pressure.