Chinese banks are paying 4% for dollars to buy Treasuries

Archived reading, published Mon, 07 Sep 2026 00:36:34 UTC (4 days ago). This is not the current state of the meter.

See the live reading →

CRASH-O-METER

Cracking

how close are we

Fragility91

how much tinder is stacked up — moves slowly

Ignition37

how close a spark is — moves fast

Crash Lab watches the machinery under the current boom: the debt paying for AI data centres, private credit, the leverage that has moved out of banks into places nobody has to mark, the bond market and the dollar.

Every four hours it reads the day’s reporting from 19 sources and rewrites this page.

Status: Held at 64. It is Sunday night and nothing in the share or corporate-bond markets has traded since Friday, so we are reading the same September 4 close we read at the last four updates: the VIX, the market's gauge of how much turbulence traders expect over the next month, at 14.53 (under 20 is calm; it passed 80 in March 2020); the S&P 500 just 1.0% off its high; shaky companies paying 2.65 percentage points more than the government to borrow; the government paying 4.78% to borrow for ten years. Bitcoin is up 0.4% to $80,200, gold is at $4,477 and the dollar index, which tracks the dollar against a basket of other currencies, is flat. So ignition, how close a spark is, stays at 37. Fragility, how much dry tinder is stacked up, holds at 91. The two genuinely new facts from the sources we trust most both point toward things steadying rather than toward a new pile of risk: Beijing is putting $54bn of capital into the insurers and banks it has been leaning on, and Chinese commercial banks have quietly become a buyer of US government bonds at the margin. But the first is a state admitting strain, and the second is paid for with dollar deposits costing 3 to 4% that can be pulled within months and sit on the banks' own books, so neither is anyone paying down risk. The rest of the week's material re-reports things we have already counted: the prices private-lending funds put on their loans, hedge funds' borrowed bets on government bonds, and the debt behind the AI buildout.

Reporting from 3 Sep to 6 Sep

Crash points

Eight places it could go, scored 0–10. Tap one for the explainer.

What’s moving the needle

What changed in the last few hours, and what each one says about the plumbing.

The dollar, gold and reserve status6 Sepmedium

The new buyer of America's debt is funding it month to month

Who holds US government bonds, and with what money, decides how those bonds behave when the market is under strain. The holders are shifting toward hands that have borrowed short-term and are betting with borrowed money.

Chinese commercial banks have spent the past few months buying US government bonds, paying for them by raising the interest they offer on dollar deposits, according to people who spoke to Reuters. Nobody had reported it before. It matters because the buyers America relied on for years are leaving, and their replacement is holding a ten-year IOU with money that can walk out the door in months. The Big Five state lenders have capped what they pay on dollar deposits at 2.8% since 2023, which is the ordinary state of affairs. What is not ordinary is the rest: customers with more than $50,000 have been able to negotiate above 3% since June, and some smaller and foreign banks have offered close to 4% since August, a few of them advertising on social media to find the money (Business Times). The arithmetic is simple. The interest rate the US government pays to borrow for ten years has risen more than 0.3 percentage points since June (to 4.76%); the rate China's government pays to borrow for ten years has fallen to its lowest in over a year; and the gap between them briefly reached its widest since 2006 on Wednesday (about 3.13 percentage points, per Nikkei). A bank that gathers dollars at 3.5% and holds US bonds paying 4.75% pockets the difference without ever touching the yuan. That matters, because the sources say the banks are reluctant to change currency themselves while regulators are watching. What the story does not say is how much. Reuters could not establish the amount, or whether it changes China's overall holdings of American debt. The mechanism is what interests us. The buyers the US bond market has traditionally depended on are leaving or shrinking. Japanese investors have sold a net ¥3tn ($18.7bn) of foreign bonds this year through August 22, the biggest outflow at this point in a year since 2022, because Japanese government bonds now paying above 3% are pulling money home (Reuters analysis via Astro Awani). Norway's sovereign fund is proposing to cut roughly $80bn. Their replacement, on this evidence, is banks borrowing short from depositors who can renegotiate or leave, to own a bond that pays a fixed rate for ten years. Anyone who watched the brokerage Schwab in 2023 knows what happens to that trade when the cost of the deposits moves. For the meter, none of this is a spark. A buyer showing up is better than no buyer, and the money is real. But the holder stepping in is smaller, less patient and more reliant on borrowed money than the one steppin

Fed, Treasury and policy6 Sephigh

Beijing writes checks to the insurers it used to prop up its stock market

When a state has to put fresh capital into the companies it used to hold up its own stock market, it is telling you what those companies were absorbing.

China's finance ministry is putting a combined $54bn into state-owned insurers and banks, the companies said on Sunday. China Life gets 35bn yuan ($5.2bn), China Taiping 7bn, China Export & Credit Insurance 10bn, China Re 3bn, and PICC plans to sell the ministry up to 15bn yuan of new shares. Separately, three state lenders announced 290bn yuan of fresh capital between them, including 30bn for the Export-Import Bank (Business Times). A government does not put sums like that into companies that are fine. The plan itself is not new; it was unveiled at the parliamentary session in March. What is new is the money actually landing, and the reasons given for it. The report says the state insurers were 'directed to support the stock market with medium- and long-term funds', that years of low interest rates have eaten into the industry's profits, and that many small and mid-sized insurers are reporting a shrinking cushion of capital against their promises to policyholders. The injection also positions the big groups to 'help regulators manage' the weaker ones. Read that as plumbing. Beijing used the insurers' balance sheets, the list of what each company owns and owes, as the buyer that shows up for shares when nobody else will. Those balance sheets, sitting on low domestic interest rates and on the shares they were told to buy, have worn thin. So the state is now writing checks to the institutions it used as shock absorbers, and asking them to absorb the smaller insurers as well. This is the Crash Lab thesis in state-directed form. In the West, risk has drifted onto insurers through lending to companies by investment funds rather than banks, and through reinsurance; in China it was put there by instruction. Either way the insurer is the layer where losses are slowest to be admitted in the prices firms put on their holdings, and where the rescue, when it comes, is a policy decision rather than a market event. For the meter, capital going in is capital going in, and it does not raise fragility. But look at the order of events: banks reported more bad loans to households last week, and insurers get their capital topped up this week. Our read is that this is measurement of strain that was already in the system, not new strain. The world did not get more dangerous on Sunday; we got a clearer look at it.

Crypto and TradFi contagion6 Seplow

Twenty-one banks decide to issue the dollar token that has been draining them

The banking system is copying the product that has been pulling deposits out of it, while bitcoin quietly becomes something you can pledge to buy a house. Both widen the pipe between crypto and the regulated system.

Twenty-one financial institutions, among them Bank of America, Capital One, Citi, Goldman Sachs and Wells Fargo, say they intend to issue a digital token meant to be worth exactly one dollar, a stablecoin, in the first half of 2027. Thirty-eight state bankers' associations have announced a separate network, the BankChain Alliance, for deposits turned into digital tokens and for stablecoins, with the same launch year (TheStreet). The coverage is explicit about the motive: to stop competitors that are not banks pulling deposits out of the banking system. Banks have spent years watching that happen; deciding to build the thing themselves is the change. That motive is the whole mechanism. A stablecoin is a bank deposit that has left the bank and become a short-term US government bond held by somebody else. Tether, one of these issuers, has more than $183bn of tokens outstanding, backed by reserves parked in US government debt; Circle minted 5bn of its USDC token in a single week; and products that turn US government bonds and money-market funds into tokens passed $16bn in August, according to congressional testimony cited in the same sweeps (Bitcoin.com). Every dollar of that is a deposit a bank no longer has to lend against. The banks' answer is to build the drain inside their own house. Two other items this week matter more for contagion than for competition. Coinbase and Better have launched mortgages backed by bitcoin, letting borrowers pledge their coins as security for the down payment rather than sell them; a waitlist reportedly points to more than $260m of demand. And on August 27 Virtu, M1X and Tradeweb completed a loan of cash borrowed overnight against bonds put up as security, with every step recorded on a blockchain ledger (the Canton network) and a government's digital bond as the security. A fall in bitcoin that becomes a demand from a lender for more cash, right now, against somebody's house, and the market's short-term plumbing rebuilt on a ledger that has never lived through a March 2020, are new channels from crypto into places it did not previously reach. Share prices say the market is not worried: Coinbase fell 4.2% on Friday but is up 20% in twenty days, and Strategy, a listed company whose shares are largely a wrapper around bitcoin, is up 43% over the same stretch against bitcoin's own 24%. All of this reaches us through secondary sources rather than filings, and none of it is under stress yet. For the meter this is wiring being laid, n

The AI capex bubble6 Sepmedium

Investors are rewarding the AI builders who borrow to build

Once the AI buildout is paid for with debt rather than profits, a shortfall in demand stops being a fall in a share price and becomes loans that do not get repaid, with named lenders holding them.

Oracle spent about $56bn building data centers and the like in its 2026 financial year and, by one account circulating this week, covered a $23.7bn hole in its cash by raising $43bn from bond investors and $5bn from selling shares (Inforcapital, via web sweep). We have not checked those figures against Oracle's own filings and the source is secondhand, so treat the shape, not the decimals, as the fact. The shape is that one of the biggest builders is paying for the AI boom with borrowed money, and the stock market is rewarding it for doing so. The same sweeps put the big technology companies' AI-related debt above $350bn and this year's new AI-related corporate borrowing at roughly $220bn, alongside a week of fresh secured loans further down the chain: Nebius enlarging a deal to $5bn, Nscale closing up to $3.05bn of loans backed by its projects in Texas and North Carolina with lenders who stand first in line, and Lambda's $1bn facility to buy GB300 chips that Microsoft will lease (Yahoo Finance). Morgan Stanley is quoted estimating some $800bn of lending from investment funds and separate off-the-books companies for this buildout by 2028. Now look at Friday's close. Oracle, the most debt-funded of the giants, sits at its high for the period, up 8% in twenty days. Amazon is 9% off its high and Alphabet 10%; Microsoft fell 2.1% on Friday. Investors are currently rewarding the builder that borrows to build and marking down the ones paying out of their own earnings. That is not irrational on its own terms. Borrowing does not hand new shares to outsiders, and if the computing power earns more than the debt costs, existing shareholders keep all the upside. But it changes what kind of story this is. Spending your own cash is a risk to shareholders: if demand disappoints, they take the loss. Spending borrowed money is a risk to lenders: if demand disappoints, bondholders, the firms on the other side of the leases and whoever bought the paper issued by those separate companies take the loss first, and the question becomes who admits it and when. For the meter this is fragility, and it makes a crash more likely, not less. A share price at its high is the market saying that question will not be asked. The $43bn is the market saying there are already a great many people to ask.

Household credit6 Seplow

Household debt: write-offs up a touch, late payments down a touch. Fine.

The consumer is the part of the system that would turn a problem in the market's plumbing into a recession, and this quarter's data says that is not happening yet.

The second-quarter numbers on American household debt, as relayed in this week's sweeps of data from the Federal Reserve, America's central bank, and its New York branch, are not alarming, and it is worth saying so. The share of credit card debt banks gave up on rose a fraction (from 3.66% to 3.70%). The share of card payments running late eased (from 2.80% to 2.74%). Household debt at least a month behind fell (from 4.8% to 4.7%), seriously overdue debt fell (from 3.4% to 3.3%), and seriously overdue car loans fell too (from 5.6% to 5.5%, per Motoring Chronicle, citing NY Fed data). The page carrying those numbers is headlined as a 32-year high in late car payments, which the figures on it do not support; we are using the data, not the headline. The pattern, write-offs up a touch and late payments down a touch, is what steadying looks like. A write-off is a loan the lender has given up on. A late payment is a loan that has just started to go wrong. When the first rises while the second falls, lenders are cleaning up last year's late payers while fewer new ones fall behind. That is the back end of a deterioration, not the front end of one. The share prices mostly agree. Capital One, Synchrony and Ally, all big lenders to consumers, are up on the week and within 3% of their highs; Credit Acceptance, which lends for cars to people with the weakest credit, is 0.4% off its high. The exception is Affirm, down 6.9% on the week and 7.9% off its high. Its buy-now-pay-later customers skew younger and poorer, and the sweep found no fresh loss data on that kind of lending at all, so the market is guessing without us. What is still standing from earlier: a January reading of roughly one in every fourteen car loans to weak-credit borrowers two months behind (6.90%), and roughly 3.6m federal student-loan borrowers in default since the payment pause ended. Those are piles of stress we continue to carry. But nothing in this quarter's flow adds to them. For the meter this adds nothing to the pile, and we say so plainly; the household beat is one place where the machine is, for now, running as designed.

Signs of the times

Stuff you wouldn’t have believed was possible until 2026.

Hong Kong takes 100 tonnes of Russian gold

Hong Kong imported almost 100 tonnes of gold from Russia in the first seven months of the year, a record and nearly three times last year's pace, and buyers there have taken about HK$276bn ($35bn) of Russian gold under Western sanctions since 2022. The output of the world's second-largest gold producer has not gone away; it has been rerouted through a city that began testing its own system for settling gold trades in July.

goldsanctionsreserves

6 SepFinancial Times

Chinese banks are advertising for dollars online

Smaller and foreign banks in China have been advertising on social media for dollar deposits, offering close to 4% since August, so they can buy US government bonds paying 4.76%, while the Big Five state lenders stay capped at 2.8%. Part of the money holding up America's debt is now coming from Chinese savers who answered an ad.

treasurieschinacarry

6 SepBusiness Times Singapore

A mortgage with bitcoin as the down payment

Coinbase and Better have launched mortgages backed by bitcoin, letting borrowers pledge their coins as security for the down payment instead of selling them, with a waitlist reportedly pointing to more than $260m of demand. The equity in some American houses is now partly a bet, made with borrowed money, on an asset that moved 24% in the last twenty days.

cryptomortgagescollateral

5 SepBitcoin.com

Gold now outweighs US bonds in the world's official reserves

Coverage of IMF data this week put gold at 27% of what the world's central banks held in reserve at the end of 2025, ahead of US government bonds at 22%, the first time a metal has been the largest single piece, even as the dollar's share of the currency portion edged up to 57.1% in the first quarter. Central banks are keeping their dollars and buying gold with the rest.

goldreservesde-dollarisation

6 SepMSN / Perplexity sweep

The rumour mill

What the crash-callers are saying, checked against real reporting.

Confirmed

“Three FOMC members dissented at the July meeting in favour of raising rates.”

Hammack, Kashkari and Logan, three of the officials who vote on interest rates at the Federal Reserve, America's central bank, each preferred a quarter-point rise at the July 28–29 meeting, which left rates at 3.50–3.75%. With markets pricing roughly 60% odds of a rise on September 15–16 and the White House campaigning against one, the number of dissenters is the number to watch.

Claimed by Thoughtful Money (Adam Taggart)

Partly true

“The 10-year and 30-year Treasury yields are at their highest since 2007, spiking partly because the government announced it would double the amount of Treasuries it buys.”

The interest rate the government pays to borrow for thirty years did hit its highest since June 2007 in mid-August, and the Treasury Department, which borrows the money the government spends, did double its purchases of its own long-dated bonds to at least $4bn per operation from September 9. But the ten-year rate is at the highs it reached in 2023 to 2025, not 2007 levels. And the cause and effect are backwards: the buybacks are a small prop under prices, not what pushed them down.

Claimed by Michael Bordenaro

Partly true

“US homebuilders currently have over 9 months of supply, a level previously reached only during 2008–09, 1980–82 and 1973–75.”

The Census Bureau's July figure is real: builders had 9.6 months' worth of new single-family homes on hand at the current sales pace. D.R. Horton's rate of buyers cancelling rising to 20%, and its lowered forecast for 2026, are confirmed. The comparison to three specific past recessions is not sourced.

Claimed by Reventure Consulting

Earlier developments

Dispatches from previous readings. The same argument, no longer the news.

Private credit and BDCs6 Sepmedium

Four in five software loans are worth less than the books said

The individual loans we have watched lose value one at a time are now a number for the whole sector, and four in five software loans have started moving the way Loparex's did.

Reuters went through the filings of ten American funds that borrow money, lend it to mid-sized private companies, and pass the interest to shareholders, and found that four in five of their software loans are now carried at less than they cost (81%, against 40% of loans to every other kind of business). Software was the industry these lenders liked best. Loans where the borrower has stopped paying, or is unlikely to pay in full, have risen to about one in thirty from one in forty at the start of the year (3.4% of cost at June 30, from 2.5% at the end of 2025), and the share carried at less than 80 cents for every dollar lent has quadrupled from where it sat for three years (about 4% now, around 1% from 2023 to 2025). Every one of those figures is the lenders' own admission that the loans are worth less than the books said, and until this week we had only seen it one company at a time. A wider cut of 44 such funds puts the value they place on everything they hold at $92.88bn, against $95.19bn it cost them: a discount of $2.31bn, or about 2.4%. That number is small, and it is built to be small. These funds decide for themselves what their own loans are worth, and until a borrower actually misses a payment there is little that forces them to admit a loss. That is why the 81% figure is the one to hold on to. A software loan that has been written down is still, in most cases, carried at 80 or 90 cents on the dollar. Loparex was carried at 88 cents in December and at 5 cents last week. The sector data says the queue behind it is long and concentrated in one industry. Software was the collateral private lenders liked best: subscription income that arrives every month, little need to spend on buildings or equipment, and easy to borrow against. The reporting does not say why it has turned, and we will not guess. Reuters also found 88 companies where two funds valued the same loan more than 10 cents on the dollar apart, which tells you these valuations are opinions. And KBRA, a firm that grades how likely borrowers are to pay you back, found the typical rate of stopped-paying loans among the funds set up with an end date jumped in a single quarter (its median rose to 2.75% from 1.81%). The market is not pricing a queue. Shares in two of the big listed lenders of this kind closed within a whisker of their highs (Ares Capital 0.8% off, Blackstone Secured Lending 1.1%). The firms that manage the funds are softer (Blackstone down 4.4% on the week and 8.8% off its high,

Hidden leverage and shadow banking4 Sephigh

The world's largest bond manager sees 2007 in two products

The largest bond manager in the world is saying, on the record, that two of the structures moving loan risk from banks to insurers are being priced the way the middle slices of bundled subprime mortgages were in 2007.

Christian Stracke, president of Pimco, the largest bond manager in the world, told Bloomberg this week that some buyers of two structures now carrying loan risk out of the banks are not getting paid enough for the risk, and that the structures share "at least some similarities" with the subprime mortgage paper of 2007–08. Fund managers compare things to 2007 for a living, but Pimco is the firm whose whole business is deciding whether a bond pays enough, so it does not say so lightly. He also said the problem is not systemic and that Pimco is not bracing for a crisis. Both halves matter. Here is how the two products work. The first takes stakes in private investment funds, the ownership slice that gets paid last when anything goes wrong, pools a lot of them together, and cuts the pool into slices ranked by who gets paid first. The top slices get a safe-looking grade from the rating firms and are sold, mostly, to insurers, whose rules let them hold rated debt cheaply but not fund stakes. That trade has a name, a collateralized fund obligation, and Bloomberg puts the wider market these sit in at over $1tn. The second runs the other way. A bank keeps a portfolio of loans on its own books but pays an outside investor to take the losses on the middle slice, and the amount of its own money the rules make it hold against those loans falls. That one is called a significant risk transfer, and the investor has, in effect, sold insurance on the middle of a loan pool. "Very similar to what we saw before the GFC," Stracke said. "Examples of people under-pricing the risk to get leverage." In other words, like the run-up to the 2008 crisis: people accepting too little pay for a risk because it lets them borrow more against it. This is our whole thesis in two products. The second moves loan risk out of the bank, where regulators charge for it and stress-test it, and into a fund or insurer whose exposure appears in no bank filing. The first turns fund stakes that cannot easily be sold into something an insurance regulator counts as a bond. Neither is priced against a market, because there is not one. What the interview does not tell us, and what nobody publishes, is how large the pile of those middle slices held by insurers is, or what the holders say it is worth. Stracke's "not systemic" is a judgment about size. Ours is that the size cannot be established from outside, which a regulator would call the point. This adds nothing new to the pile. It is one more credible perso

The AI capex bubble6 Sepmedium

Whose money is actually in the chips?

When the owners put in almost nothing and the thing pledged to the lender wears out, the lenders funding the AI buildout are the ones betting that a three-year-old chip will still cover their loan.

The chief executive of IREN, one of the companies filling data centers with Nvidia's Blackwell Ultra chips, said this week that its customers' advance payments can cover about half the cost of those chips and that lenders can fund most of the rest. Read that twice. A company is buying some of the most sought-after hardware in the world mostly with its customer's money and its lender's money, and what its own shareholders own is whatever is left over. Funds run by Blue Owl arranged $2.4bn for IREN's British Columbia campus, split between a $1.2bn loan and $1.2bn of bonds, both of which get paid first and have the assets pledged against them. Chinese-language coverage picked up by our sweep put the interest rate at a fixed 9% and said the money covers about 90% of the related chip spending. We have not seen the term sheet. The same stack is everywhere. Nebius raised its debt sale to $5bn. Nscale closed up to $3.05bn of loans, secured on its sites in Texas and North Carolina, and Bloomberg reports it wants $3.5bn more before listing its shares: $1.5bn of bonds that can turn into shares and about $2bn from Nvidia, the vendor whose chips the money buys. Lambda's $926m loan sits in a separate company set up to hold the servers off the parent's books, pays an interest rate that floats 3 percentage points above the market's benchmark overnight rate, comes due in December 2030, and is secured on the servers and the cash they earn. Big Tech's AI-related debt has passed $350bn by one estimate. The mechanism to hold on to is this: a chip is collateral that loses value on a schedule nobody agrees on. If the customer pays half up front and the lender funds most of the rest, the party carrying the risk of what the chips are worth at the end is whoever holds the bond when the contract ends and the hardware is three years old. The debt is secured. The security wears out. The market's answer this week was to buy the shareholders' slivers. IREN rose 26% in five days and sits 0.6% off its high; Nebius rose 7.5% on Friday, CoreWeave 5.6%. Oracle, which has raised $43bn of debt against $55.7bn of building, closed at its high. Prices say the question of what used chips are worth has been answered. The filings say it has not yet been asked. This is fragility, and the slow kind: nobody is going to be forced to sell a chip this month. But every one of these loans rests on a price for used hardware that nobody has had to name, and the more of them there are, the more it matters when

The dollar, gold and reserve status6 Sephigh

Central banks are moving the gold, not selling it

Where a central bank keeps its gold is a vote on how far it trusts the United States to hand the bars back, and two European ones cast that vote this year without moving the price.

The Dutch central bank moved about 86 metric tons of gold out of vaults in New York and Ottawa and into London between March and August, citing "increasing geopolitical unrest" and a wish to be better prepared for a severe crisis. Central banks do rearrange where they keep their gold. Saying in public that the reason is getting ready for a severe crisis is not the routine part. New York used to hold nearly a third of Dutch gold and now holds less than a fifth (31.3% to 18.5%); London's share has nearly doubled (18.1% to 32.1%). France has shifted 129 metric tons out of New York, saying the bars did not meet its purity standard. India's figures show 77% of its central bank's gold is now held at home. And the FT reports that Hong Kong imported almost 100 metric tons of Russian gold in the first seven months of the year, a record and nearly three times last year's pace, as the city tries out its own system for settling gold trades. None of this moves the price. Gold closed Friday at $4,477, down 0.3%, about 5% off its high. It is not about price. Location is a statement about which legal system you trust to hand the bar back when you need it. A central bank keeps gold in New York because that is where it can be sold or swapped for dollars fastest in a crisis. Moving it to London, or home, trades that speed for distance from American law. Two European central banks made that trade in the same year that Australia's central bank confirmed it was cutting its dollar holdings and Norway's sovereign fund proposed cutting $80bn of US government bonds. One item in our sweep, from a single source, says gold reached 27% of central banks' official reserves at the end of 2025 against 22% for US government bonds. We cannot confirm it and quote it only as a claim. The IMF, which lends to countries in trouble and keeps the official count of what currencies central banks hold, has the dollar's share of those reserves rising to 57.1% in the first quarter. That cuts the other way, and it is a reminder that this is a drift, not a run. What it adds to the machine is this. The market for US government debt has leaned on foreign official buyers who held the bonds and stored their gold in New York because both were convenient. The convenience is being repriced, one vault at a time. It moves the odds only slowly, and it moved them months ago: these transfers happened between March and August and we count them as already in the pile. But the pile is what a bad auction of government de

Fed, Treasury and policy6 Sepmedium

Three dates: September 9, 15 and 18

Three policy decisions in nine days land on the exact seam where a mistake would show first: Japanese money coming home from US government bonds just as the cost of America's longest-term borrowing rises.

Between September 9 and September 18, three governments make decisions that all touch the same nerve. On September 9 the Treasury Department, which borrows the money the US government spends, doubles the amount of its own long-dated debt it buys back to keep that part of the market trading, to at least $4bn per operation on bonds with between ten and thirty years left to run, running through November 4, and insists this is "not monetary policy". On September 15 the Federal Reserve, America's central bank, meets with traders pricing a rate rise as more likely than not, after Warsh's speech at Jackson Hole, the central bank's annual August gathering, and a jobs report showing 162,000 new jobs. On September 18 the Bank of Japan meets, with Japan paying more than 3% to borrow for ten years for the first time since 1996 and its five-year and forty-year borrowing costs at records. Japan's finance minister, Satsuki Katayama, spent Friday saying there was "absolutely no sense whatsoever" that Scott Bessent, the US Treasury Secretary, had pressed her for a Japanese rate rise. Bessent, at the same meeting of the G20, the club of the biggest economies, said he was sure the Bank of Japan would "do the right thing". Traders read that as pressure. The yen swung between ¥160 and ¥155.3 to the dollar in the week. Here is the wiring. Japan's budget for the year to March 2027 assumes it pays an average of 3% to borrow for ten years, according to Deutsche Bank's Kentaro Koyama; it is already there, so every hundredth of a percentage point above that is interest the budget did not plan for. Higher rates at home pull Japan's life insurers back from foreign bonds (they sold a net ¥3tn of overseas debt through August 22), and Japan is the largest foreign holder of US government bonds. So the US Treasury Secretary appears to be encouraging the one policy move most likely to shrink the pool of buyers for his own bonds, in the same month his department starts buying back its long-dated debt to keep that market working. The buybacks are paid for by selling very short-term IOUs, so the Treasury is shortening how long its debt runs at the moment long-dated bonds are being sold off. It has swapped the risk that long bonds fall in price for the risk of having to come back and borrow again, more often, at whatever rate is on offer. It is small: $4bn against a market where hedge funds alone carry $4tn of gross exposure. But it starts Tuesday, and it is the first support operation on the c

Crypto and TradFi contagion6 Sepmedium

The companies that hold bitcoin ran 20 points ahead of bitcoin

The premium investors pay for companies that hold coins is the froth that goes first, but the reserves behind digital dollars and the mortgages backed by bitcoin are the pipes that would carry a crypto slump into government debt and housing.

Bitcoin has risen by nearly a quarter in twenty days (23.7%, to $79,800), and the listed companies whose whole business is holding it have risen far more: Strategy is up 42.7%, BitMine 32.8%, SharpLink 35.3%, Galaxy 30.6%, Coinbase 20.2%. On Friday, with bitcoin flat, BitMine fell 5.5%, Coinbase 4.2% and SharpLink 3.3%. That gap is the story. Investors are paying more for these companies than the coins they hold are worth, and that extra is what borrowed money looks like without any borrowing: it swells when the coin rises and is the first thing to vanish when the coin stalls. Meanwhile the plumbing between crypto and the banks thickened in ways that will outlast any premium. Twenty-one institutions including Bank of America, Capital One, Citi, Goldman Sachs and Wells Fargo say they intend to issue a digital token meant to be worth exactly one dollar, a stablecoin, in the first half of 2027. Circle, which runs one of the two biggest, created 5bn new USDC in a single week. Citi ran live transactions on September 2 in which bank deposits moved as digital tokens across Swift's ledger, the network banks use to message each other about payments. Virtu, M1X and Tradeweb settled a loan of cash borrowed overnight against bonds put up as security entirely on a digital ledger, the Canton Network, against a government-issued digital bond, with the bond, the cash and the return all settling at once or not at all. Tether, the other giant, has more than $183bn of tokens in circulation and makes its profit by holding US government debt against them; one sweep puts that holding at about $141bn. And Coinbase and Better launched mortgages backed by bitcoin, with a waiting list one recap put at $260m. The mechanism deserves spelling out. The borrower pledges bitcoin as the security for the down payment rather than selling it, so the loan is secured on a house and on an asset that moved 24% in a month. If the coin falls, the borrower must put up more or the lender sells the coins into a falling market. That is a demand from a lender for more cash, right now, because a bet has moved against you, the same dynamic that took down the crypto lenders of 2022, now attached to residential property. The route into the wider system we watch is the reserve pool. Tether plus Circle is a very large holder of the government's shortest-term IOUs, and it shrinks when token holders cash out, which happens in days, not quarters. The premiums are froth, and froth popping does not move our numbe

Private credit and BDCs5 Sephigh

From 88 cents to 5 in eight months

Private credit's stability rests on marks nobody tests until a borrower stops paying, and this is what a test looks like.

At the end of last year, Blue Owl Capital Corp — OBDC, the listed BDC — carried its first-lien loans to Loparex, a maker of adhesive liners owned by Pamplona Capital, at 100 cents on the dollar and its second-lien debt a little below 90. This week Moody's deemed the company in default and said Chapter 11 may be coming. The second lien is now marked at about 5 cents. OBDC put the name on non-accrual after the second quarter, which a BDC must do when full repayment is in doubt (Bloomberg). This was not a surprise borrower. S&P said in 2024, after a distressed exchange it treated as tantamount to default, that Loparex's capital structure was "unsustainable". The company skipped a June interest payment on the second lien and is under forbearance through September. So the question is not why the loan went bad. It is why a loan to a company already judged unsustainable was marked at 88 in December. The mechanism is the point. A BDC's loans do not trade. The manager values them quarterly, and Bloomberg notes the same stressed loan is routinely carried at different values by different lenders. The mark is a claim the manager makes about its own portfolio, and it is only tested when the borrower stops paying. Then it moves in one step. How much this matters for OBDC is limited: non-accruals are 0.8% of fair value. The sector numbers are where it bites. Reuters' review of 44 BDCs put aggregate fair value at 97.57% of cost at end-June, from 99.25% at end-2025, with about 4% of borrowers marked below 80% of par against roughly 1% in 2023–25 (via web sweep). OBDC rose 0.2% on Friday and is 4% off its high; Blue Owl the manager fell 0.8%. The market is treating Loparex as one bad name. It is one bad name. It is also a demonstration of how far a private mark can sit from where the loan would clear, for how long. Bloomberg does not give the current first-lien mark; that is the number we want.

Hidden leverage and shadow banking5 Sepmedium

Jefferies' invoice fund finds a second fraud

Lending that used to sit on a regulated balance sheet now sits in a fund whose collateral turns out to be paper, twice.

Point Bonita, the private credit fund inside Jefferies' Leucadia Asset Management that lent to the collapsed auto-parts group First Brands, has almost $500mn of exposure to the iron ore trader Radiant World and a linked company. Jefferies obtained a freezing order for up to $499mn from London's High Court against Radiant, its founder Pinkesh Nahar, an allegedly affiliated firm called Sapphire Minmetals and its chair Rakesh Sethi. Radiant and Nahar deny wrongdoing. This comes on top of the $715mn of First Brands exposure Jefferies disclosed last October (Financial Times). Point Bonita does invoice-backed lending. The fund advances cash against receivables: a piece of paper saying a customer owes the borrower money. The loan is as good as the paper. If an invoice is invented, or pledged to two lenders, the collateral is nothing, and the lender does not find out until it tries to collect. Several of Radiant's lenders cut ties in recent weeks over alleged inconsistencies in the documentation behind its financing. Point Bonita's exposure had grown from $35mn in 2021 to hundreds of millions over four years. Two alleged frauds in one fund is not two pieces of bad luck. It says something about an underwriting model that grew fast against documents it did not independently verify, and that appears to have been the same in both cases. The market's view: Jefferies rose 1.8% on Friday, is up 4% on the week and sits 2.75% off its high. That is a price that says this is fund investors' money, not the bank's. The FT's report does not say how much, if any, of the exposure is Jefferies' own capital, and that is the number that decides whether this is a fund problem or a Jefferies problem. Trade finance is what banks used to do with a credit committee; it is now done in a fund with redemption terms, and the FT notes Point Bonita has faced a wave of redemption requests. That is the migration this site is about, in miniature.

Bond market dysfunction5 Sephigh

The junk index says 265. The tail says 1,053.

The number everyone watches for credit stress is built so that it is the last to notice the weakest borrowers breaking.

Our own crash trigger watches the high-yield index spread. It is at 265 basis points, tighter than a month ago, and needs to reach 800 to fire. Meanwhile the spread on companies rated triple-C or lower has climbed to 10.53 percentage points, from 8.08 a year ago, the highest since the aftermath of last year's tariff shock (Financial Times). Both numbers are true. They describe different markets. The index spread is a weighted average dominated by double-B issuers — large, mostly profitable, and, in the words of Ninety One's John Stopford, "as expensive as they have ever been". Triple-C names are a small slice by market value, so their spread can go to 1,053 without moving the headline much. Higher Treasury yields — the 10-year hit 4.82% this week, its highest since 2023 — do nothing to a double-B company paying a fixed coupon until it refinances. A triple-C company refinancing now pays roughly the base rate plus ten and a half points. The damage shows in the default data. JPMorgan counts $40.1bn of default actions this year, up 9% on 2025, led by cable and satellite after Dish DBS's $9.75bn default in June. The recovery rate over the past twelve months is 29%, against a 25-year average of 40%. Corbin Capital's John Cocke attributes part of that to aggressive restructurings that have taught investors to expect less back. Why this matters for private credit: BDC borrowers are unrated but mostly single-B and triple-C equivalents, and their loans float. They do not wait for refinancing to feel SOFR at 3.66% plus a spread; they feel it every quarter. Loparex, above, is what the tail looks like inside a BDC. We read this as observable stress, confined for now to the weakest borrowers. HYG is 0.4% off its high. The thing to watch is whether single-B spreads follow. If they do, the index will finally say what the tail has been saying.

Fed, Treasury and policy5 Sepmedium

Japan's regulator follows the megabank money offshore

The leverage under private credit runs on bank money, and the regulator of the cheapest bank money in the world has started asking where it went.

Japan's Financial Services Agency will examine whether the country's banks are properly assessing the credit risk in their lending to overseas nonbank lenders and real estate firms, Nikkei reports. It will also review lending policies and project screening for data centres, and study what rising interest rates do to deposit stability and to unrealised losses on government bond holdings. The item is going into a draft of the agency's annual priorities plan, to be finalised by June next year (Nikkei). This matters because of where the private credit machine gets its leverage. The loans have left bank balance sheets. The funds that hold them borrow — fund finance, back-leverage, NAV loans — and Nikkei says overseas nonbank intermediaries are "increasingly tapping Japanese banks for funding", mainly the megabanks. So the corporate loan is not on the bank's books; the loan to the lender is. The risk did not leave the banking system. It went one layer up. Pimco's president Christian Stracke described the other end of the same pipe this week. Collateralised fund obligations and synthetic risk transfers — structures that shift bank and fund exposure to insurance companies — show "people under-pricing the risk to get leverage", and writing protection on a mezzanine tranche is "very similar to what we saw before the GFC". He says it is not systemic (Bloomberg). What has changed the economics is the 10-year JGB at 3% for the first time since 1996. The yen funding that made lending to foreign nonbanks attractive is no longer nearly free, and the FSA's own list — deposit stability, bond losses — is a list of a bank under rising rates. One analysis circulating this week puts Japanese investors' net sales of overseas debt at ¥3tn ($18.7bn) through 22 August, the largest year-to-date outflow since 2022; we have not been able to corroborate that figure from a paper of record. None of this is an unwind. A priorities plan finalised in 2027 is measurement. A megabank actually pulling a facility from a fund would be something else, and that is what would move our ignition number.

Crypto and TradFi contagion5 Sepmedium

The Treasury buyer Bessent promised is shrinking

A marginal Treasury buyer that sells exactly when markets are stressed is not a substitute for the ones that are leaving.

Tether's USDT, the largest stablecoin, contracted by nearly $3bn to about $184bn in the first half of the year, putting it on course for its first annual shrink since the 2022 crypto crash. Circle's USDC fell by a similar amount to roughly $72bn (Bloomberg). Last year Treasury Secretary Scott Bessent said the stablecoin market could grow tenfold to $3tn by the end of the decade, which would make issuers a major new buyer of the government's debt. The mechanism is simple and that is the problem. A stablecoin issuer holds Treasury bills against every coin, so a dollar of coin is roughly a dollar of bill demand. But coins are where traders park money between trades. Supply tracks crypto trading volume. When trading dries up, holders redeem, and the issuer sells bills. This buyer is procyclical: it shows up when risk appetite is high and leaves when it is not, which is the opposite of what a Treasury market losing its old buyers needs. Norway's oil fund proposed cutting roughly $80bn of Treasuries this week; gold now outranks Treasuries in official reserves. There is a contradiction in the window worth stating. A crypto-industry weekly recap, surfaced by a web sweep, says Circle minted 5bn USDC in a single week and that 21 firms including Bank of America, Citi and Goldman Sachs plan a joint dollar stablecoin for the first half of 2027. That is a single tier-2 source and we cannot reconcile it with the first-half attestation figures — though both could be true if the rebound is recent. Bitcoin is up 24% in twenty days to about $80,000 and Strategy is up 43%, and Bloomberg itself says the stall may prove temporary. As a payments product, stablecoins are fine. As a pillar of Treasury demand, this is a fair-weather friend, and the weather has been changeable.

Hidden leverage and shadow banking4 Sephigh

The buyer holding up the government bond market borrows its money every night

When the extra buyer the US government bond market now depends on is a hedge fund that has to renew its borrowing every morning, a cash squeeze becomes forced selling within days, not quarters.

Hedge funds owned about $2.6tn of US government bonds at the end of last year, and their total bets on those bonds, counting the ones they own and the ones they have bet against, came to $4.0tn ($1.6tn of it bets that prices fall). They paid for the bonds with $3.0tn of cash borrowed overnight against bonds put up as security. The figures come from a paper the Federal Reserve, America's central bank, published in June, and Reuters' Jamie McGeever dug up this week. He adds that the bets, the overnight borrowing and the amount changing hands each month have all more than doubled since the start of 2023, and that the 50 largest funds account for roughly 90% of it. Fifty firms, then, carry most of a $4.0tn position in the debt of the US government on money they have to borrow again every single morning. The trade behind it is hedge funds borrowing heavily to profit from a small gap between two nearly identical prices. A fund buys a government bond, sells a contract to deliver that same bond at a fixed future date, and pockets the small difference between the two prices. The bond is paid for with cash borrowed overnight against bonds put up as security, so the position has to be financed afresh every day. The New York branch of the Fed published its own map of this market on September 1. It says the corner where dealers lend to 'levered accounts, most of which are hedge funds executing Treasury relative-value strategies' is the one where loans are arranged privately between two parties with no clearing house standing in the middle, and where borrowing costs the most of the three segments. In other words, the banks' customers in that corner are the funds running exactly this trade. This is not new risk. The paper describes the end of last year, and it has been in our number. What changes the picture is who sits beside these funds. Norway's government fund has proposed selling roughly $80bn of US government bonds. China's reported holdings, per a Reuters figure that survived our fact-check, are at their lowest since 2008. The Treasury Department, which borrows the money the government spends, starts buying back old government IOUs that do not come due for twenty or thirty years at $4bn a time from September 9, because Bessent, who runs it, describes trading in 30-year bonds as 'especially poor'; in other words, they cannot be sold in size without moving the price. Patient holders who never borrowed against their bonds are being replaced, at the edge of the market

Private credit and BDCs5 Sepmedium

The lender that got out of Bathla nine months before it collapsed

Australia is running the whole cycle in miniature, including the part where the best-informed lender got every dollar back and the loss landed on whoever put up the new money.

Alceon, an Australian fund that lends to companies in place of a bank, had staff from one of its own businesses working on-site at Bathla Group projects for more than a year. Then, in December 2025, it sold its way out of nearly A$460m of loans to Bathla when the company borrowed again to pay off the debt coming due. Nine months later Bathla collapsed owing A$3.2bn to A$3.5bn, with what ASIC, the agency that polices Australian markets, calls 'substantial exposure to private lenders'. Alceon told investors it no longer has any. The lender that knew the most got out whole, and somebody else is holding the loss. The word that matters is refinancing. In a book of private loans, a refinancing is recorded as the borrower paying back the full amount. The value the lender said the loan was worth was tested, and it held, for the lender who left. The loan did not get any better. It moved to whoever supplied the new money, and the Australian Financial Review's report does not say who that was. What it does say is that the lender with people inside the building chose to leave, and the lenders who replaced it did not have that view. This is the thing we keep coming back to: in a market with no public price, the exit of the best-informed creditor looks exactly like a healthy repayment. The sequence has now run its full course. ASIC and Australia's central bank, the RBA, describe the 'first significant cracks' in a market of roughly A$200bn after Bathla and the A$1.8bn collapse of Jon Adgemis's hospitality group. Four funds, Merricks, Longreach Credit, Centuria Bass and MA Financial, have limited how much money investors can take out. Centuria Bass stopped both withdrawals and new money on August 14, and MA Financial's Secured Loan Series now lets out 1% of the fund a month. The sums are small next to the US market. The order of events is not: the borrower fails, the stated values move, the exits narrow. In the US the values have moved and BCRED, Blackstone's fund that lends to private companies, is capping withdrawals at 5%, but Blackstone's shares (BX) are only 8.8% below their high and Ares Capital's (ARCC) less than 1%. This moves nothing in the US number, because Australia was already in it. What it gives us is the order of events and a fix on where each market sits in it. Australia has reached the stage where the exits close. The US has reached the stage where the stated values move. The pile is the size it was; we can now see one step further down the sequence tha

Crypto and TradFi contagion4 Sepmedium

Three-quarters of a trillion dollars of borrowed stock bets on crypto exchanges

Bets on stocks made with borrowed money, on exchanges no US regulator supervises, have grown 33-fold in nine months, and that is a new place for a forced seller to appear.

Contracts that let traders bet on stocks, stock indexes and commodities with borrowed money traded $778bn in August on the major crypto exchanges, according to Fasanara Digital via Bloomberg. In November the figure was $23.5bn. Such contracts are now 23.48% of everything traded in that form on those platforms, up from 0.5%. Borrowed bets on American shares have found a home outside the system built to supervise them, and they got there in nine months. The contract is a perpetual future: a bet on a price that never expires. Instead of settling on a date, those betting up and those betting down pay each other a small fee, the funding rate, to keep the contract's price close to the real one. The security a trader puts up is usually a stablecoin, a digital token meant to be worth exactly one dollar. The exchange decides how much borrowing it allows, and if a trader's security runs out an automatic program closes the bet. There is no bank standing behind the trader, nobody phones with a demand for more cash, and the US rule limiting how much a broker can lend against shares (Regulation T) does not apply. Exchanges built on a platform called Hyperliquid, such as Trade.xyz, launched a run of these contracts earlier this year on silver, oil and shares of companies not yet listed. 'The volatility of some stocks is multiple times higher than Bitcoin,' says Fasanara's Nikita Fadeev. 'Here you can also trade these kinds of assets ... with much more leverage.' In other words, traders can borrow far more against something that swings far more. Two cautions. The $778bn is the total that changed hands in a month, not the amount of bets open at any one moment, and Bloomberg does not say how much borrowing the average bet carries, so the capital actually at risk is unknown. And a chain of forced closures on an offshore exchange does not hit the New York Stock Exchange directly. But this is risk moving house in its purest form. Borrowed bets on US shares and commodities are being built on exchanges that file nothing with US regulators, where the exchange is also the clearing house that guarantees the trades and the insurance fund that pays when they fail. Traders who profit from tiny gaps between the crypto price and the real one keep the two tied together, which means they also carry trouble back across the link. Set beside the stablecoin figures (Tether shrank $3bn in the first half of the year as crypto trading dried up), the picture is of crypto's traders leaving crypto

The dollar, gold and reserve status5 Seplow

Central banks now hold more gold than US government bonds

The central banks that used to buy US government bonds no matter what are shifting into an asset that pays no interest and cannot be frozen.

Gold made up 27% of the reserves the world's governments and central banks hold at the end of 2025, ahead of US government bonds at 22% and the euro at 15%. That makes gold the largest single holding for the first time since 1996, according to reporting that surfaced this week. Central banks bought 51 tonnes in June, Poland and China leading. Australia's central bank has become the first major one to confirm it has cut its dollar holdings. And the Dutch central bank said this week it is moving 86 tonnes out of New York and Canada: 59 tonnes sold in New York and bought back in London, 27 tonnes physically moved to a vault at Zeist, and New York's share of Dutch reserves cut from about 30% to 18%. The buyers the US bond market could once take for granted are choosing metal over paper. The honest caveat is arithmetic. Gold is at $4,477 an ounce. A share measured by value rises when the price rises, and we cannot split the 27% into 'bought more gold' and 'gold went up'. The same figures show the dollar's share of the currency reserves central banks hold ticking up to 57.1% in the first quarter of 2026 from 56.4%. Norway's proposed $80bn cut in US government bonds leaves its dollar holdings, in its own words, 'essentially unchanged'. The dollar index, which tracks the dollar against a basket of other currencies, is at 99.2 and down less than 1% on the month. This is not a flight from the dollar. It is something narrower: a shift away from US government bonds specifically, by the holders who never needed to sell them. That ties straight into the hedge funds above. Every tonne of gold a central bank buys instead of a bond is a bond that ends up, at the edge of the market, with someone paying for it each night with cash borrowed against bonds put up as security. The Dutch move is the tell. It is not a bet against America. It is a decision that gold is easier to use 'in crisis situations' if it is not in New York. For the meter this is fragility, and the slow kind. Nothing sparked. The patient holders are thinning out, which leaves more of the market to the borrowed kind, and that is one more thing on the pile. It is not enough on its own to move the number this week; it is the kind of thing that makes the number hard to bring down.

Household credit5 Sepmedium

Car loan arrears at a 32-year high; the lender's shares at a record

Households at the bottom are falling behind, and the market has decided the lenders were built for exactly that. The verdict has held so far; a second wave of missed payments would test it.

Car loans to Americans with poor credit were going bad in January at the worst rate in the 32 years that Fitch, one of the three firms that grade how likely a borrower is to pay you back, has kept count (6.90% of them 60 or more days late). It eased to 5.8% by the end of the second quarter, then worsened again in July, 'particularly in subprime', Fitch says, concentrated among lower-income borrowers carrying a lot of debt. The New York branch of the Federal Reserve puts car loan balances three months or more behind at 5.5% in the second quarter, and credit card balances that far behind at 12.9%, down a hair from 13.1%. Some 3.6 million people with federal student loans have defaulted since the payment pause ended. The people at the bottom of the credit ladder are missing payments in numbers not seen in a generation, and their lenders' shares are near records. Credit Acceptance, which lends to the riskiest car buyers of all, closed Friday at $605.64, 0.4% below its high and up 4.3% on the month. Ally is 2.8% below its high. Affirm is the exception, down 6.9% in five days. The mechanism squares those two facts, at least for now. Lenders in this corner never expect everyone to repay. They charge rates that assume a large share will default, then fund the loans by bundling them together and selling slices of the income to investors: the loans go into a trust, the investors holding the top slices get paid first, and the lender or a junior investor holds the slice that takes the first loss. A 6.9% delinquency rate is a disaster only if total losses exceed what the extra interest and that first-loss slice were built to soak up. A share price at a record says the market thinks they were. What we cannot see is whether Fitch's July turn is the start of a second wave. The January figure is old, and it was a measurement, not news. July is the piece worth watching, because it arrives as the Federal Reserve leans toward raising interest rates on September 15 and diesel sits at a record $5.85 a gallon, costs that land first on exactly the borrower Fitch is describing. For now this changes the odds very little, and we say so plainly. The stress is real, it was known, and the structures built to absorb it have absorbed it. If July turns into a trend, it becomes fragility, and the first sign will be a first-loss slice getting eaten through faster than the lender assumed.

Fed, Treasury and policy4 Sephigh

Ten days to a Federal Reserve meeting nobody agrees on

An interest-rate decision argued over in public ten days before it happens is the nearest thing on the calendar to a policy mistake in progress, and the rates the government pays to borrow for twenty or thirty years are already the highest in years.

The August jobs report showed 162,000 jobs added, nearly three times the roughly 55,000 that forecasters had expected, and July was revised from a loss of 23,000 jobs to a gain of 21,000. A miss of that size matters because it pushes the Federal Reserve, America's central bank, toward raising interest rates at a meeting ten days away, at the very moment the president is demanding it cut them. Within hours he posted: "LOWER THE RATE OR I'LL STOP TRADING WITH COUNTRIES WITH WHICH WE HAVE A DEFICIT", named Switzerland, Mexico and the EU as the targets, and said the Fed should be "at 1% or half-a-percent" rather than 4% (Bloomberg). The collision now has a date. The Fed's chair, Kevin Warsh, told the central bankers' annual gathering at Jackson Hole on August 28 that "two percent is a firm and fixed target" and that unless inflation is clearly heading there "we still have work to do". Traders read that as a signal that rates could go up rather than down, and put the odds of a September rise above 50%. The meeting starts September 15. On the same afternoon as the president's post, his own economic adviser Kevin Hassett argued the other way, saying consumer prices over the past three months are rising at an annual pace of 1.6%, and that a favorable inflation reading would be a "strong argument" to hold rates where they are. The part that matters most is what the government pays to borrow for a long time. It pays 4.78% to borrow for ten years and 5.25% for thirty, and the Treasury Department's enlarged bond buybacks, the program Scott Bessent announced to fix "especially poor" trading conditions in thirty-year bonds, had not carried out a single operation as of August 31. Ordinarily a rate rise into a market worried about inflation and deficits lowers the long-term rate, because bondholders need less compensation for the risk that inflation eats their money. The president rejected that explicitly: "It doesn't reassure the bond market." So three outcomes. A rise defies the president and tests the Fed's independence in the open. A hold, with Hassett's inflation number as cover, leaves the market asking whether it was the data or the post. A cut is implausible on this data, and the market for long-term government bonds would say so. So far investors have priced the Fed doing what the data says: rates flat, the dollar up 0.16%, and the VIX, the market's gauge of expected turbulence, at 14.5, which is calm. Where it fits our argument: policy is the one area where risk

The AI capex bubble4 Sepmedium

Nvidia lends $2bn to a customer that says it has $103bn of contracts

Each new computing company Nvidia funds adds another loop to the circle in which the chipmaker finances the demand for its own chips, and the contract backlogs that justify it rest on customers who are themselves still raising money.

Nscale, a London company that rents out computing power for artificial intelligence, is in talks to raise as much as $3.5bn before it lists its shares on the stock market: up to $1.5bn of convertible notes, debt that can later be swapped for shares, led by Daniel Loeb's Third Point, and about $2bn of financing from Nvidia, according to Bloomberg. Nvidia makes the chips Nscale will buy with the money, so the supplier is funding its own customer's order. The listing could raise another $3bn on top. Nscale is telling investors its contracts total about $103bn, following a $45bn agreement to supply computing power to Anthropic, and that it could generate about $18.1bn of annual revenue. Take the pieces in order. Nvidia is the company whose chips Nscale will buy with the money; Nvidia is also putting up more than half of the money. That is vendor financing, a supplier lending its customer the money to buy its product, and it is the pattern we track from CoreWeave to Lambda to IREN. Nvidia's share price closed at a record on Thursday. The convertible notes are offered at the listing price minus a double-digit percentage discount, with the discount adjusted up to a $30bn valuation. That means the lenders who get in before the listing pay less than the public will, and their money counts as debt until they choose to turn it into shares. Then the $103bn. Contract value is not revenue and it is not cash. It is a promise by customers, chiefly Anthropic, to pay for capacity over years, contingent on that capacity being built, powered and delivered on time, and on the customer still being able to pay. Anthropic's ability to pay $45bn depends on its own fundraising, which depends in part on the same giant cloud companies and chipmakers. Bloomberg's sources do not say what the cancellation rights are or how much of the $103bn the customers must pay whether or not they use the capacity; we cannot tell. None of this is unusual by 2026 standards, and that is the point. It is a company valued at $30bn whose customer list depends on the same handful of balance sheets, the lists of what each company owns and owes, that are financing it. The smaller computing-rental companies rallied on the day (Nebius +7.5%, IREN +7.3%, CoreWeave +5.6%) while Microsoft fell 2.1% and Alphabet 1.1%. The market is paying more for the borrowers and less for the customers. For the odds of a crash this is fragility rather than a spark: one more loan in a circle that holds only as long as everyone in

What would change our mind

The specific, observable things that would move the number - in either direction.

The Federal Reserve's rate-setting committee meets September 15–16 and either raises rates over the White House's open objection or visibly holds because of it. The first tests hedge funds' heavily borrowed bets on tiny gaps between nearly identical government bond prices; the second tests whether anyone still believes the Fed decides for itself. Either would move ignition several points.

Would move the number

The Treasury's first $4bn purchase of its own long-dated bonds on September 9 failing to steady the thirty-year rate, or a ten- or thirty-year auction where buyers only show up at a higher rate than expected. That is the point at which the sell-off in government bonds stops being investors demanding a fairer price and becomes a market that is not working.

Would move the number

A fund that lends to private companies and cannot be sold on an exchange (a non-traded BDC or interval fund) breaking its 5% cap on withdrawals outright, or an offer to buy back shares in BCRED or a peer settling at more than 15% below what the fund says they are worth. Either would mean the queue of investors waiting to get out has turned into a forced seller.

Would move the number

Data on the size of Chinese banks' purchases of US government bonds, or evidence that Japanese life insurers' selling of foreign bonds has run past this year's ¥3tn pace. Either would tell us whether the government's long-dated debt has a real buyer base or just a trade that lasts as long as the interest gap does.

Would move the number