Six billion from the housing banks, twenty billion to affiliates

Archived reading, published Wed, 02 Sep 2026 22:23:24 UTC (18 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

Ignition39

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 40 sources and rewrites this page.

Status: Held at 65. The tape is the same 2 September close read at the last two runs — VIX 15.2, S&P 1.7% off its high, high yield 265bp and tighter than a month ago, the 10-year unchanged at 4.80% — so ignition stays at 39. Fragility holds at 91: Group 1001's $6bn of Federal Home Loan Bank borrowing against $20bn of affiliate loans and UVA's exposure to a leveraged $45bn AI fund are better measurements of leverage already standing in insurers and endowments, not new stock, and SB Energy's $439bn backlog is the same filing we read at 22Z yesterday.

Reporting from 31 Aug to 2 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.

Private credit and BDCs2 Sepmedium

An insurer, a housing agency and $20bn to the family

It is the running thesis in a single balance sheet: cheap public-backed funding on one side, unmarked related-party loans on the other, policyholders in the middle.

Two insurers controlled by Mark Walter, the Guggenheim chief executive and sports owner, owed more than $6bn to the Federal Home Loan Bank of Indianapolis at the end of June, a 36% jump in six months and roughly double the balance at the start of 2025, according to Bloomberg. Over the same period the insurers, grouped under a unit called Group 1001, have lent more than $20bn to other parts of Walter's empire. The plumbing matters. The Federal Home Loan Banks were created in the Depression to keep mortgage lending alive: they borrow cheaply on the strength of their government sponsorship and pass that funding on to members against pledged collateral. Insurers can join. Once in, an insurer can post bonds it already owns, draw cheap money, and invest the proceeds at a spread — a trade that has nothing to do with home loans. Bloomberg says insurers' borrowing from the system has tripled since 2013, and that Group 1001 is going further than peers in maximising what it can draw. So one side of these balance sheets is funded partly with quasi-public money; the other side holds loans to affiliated companies. The people ultimately exposed are policyholders, whose annuities are backed by whatever those related-party loans turn out to be worth. Cornelius Hurley, a former FHLB Boston director, calls it 'a pure and simple case where the taxpayers are subsidizing funds that are given to insurance companies'. What the reporting does not tell us: what the $20bn of affiliate loans are secured on, or how state insurance regulators value them. That gap is the point. This is the migration Crash Lab keeps returning to — risk moving from banks, where it is stress-tested and marked, into insurance balance sheets, where a related-party loan can sit at cost. The share prices of the listed insurance-plus-credit managers (Apollo 7.5% off its high, KKR 7.7%) show mild unease, not alarm. Nothing here says the loans are bad. It says nobody outside the group can check.

The AI capex bubble2 Sepmedium

Nvidia owns the landlord, the tenant and the guarantee

Investment-grade project debt is being issued against leases whose real credit is a chip vendor's guarantee — the vendor-financing thesis in its largest form yet.

SB Energy's IPO filing puts a $439bn backlog against a business that has not yet earned a dollar from data centres. Nikkei's read of the S-1 shows where the number comes from: 8.8 gigawatts of contracted capacity, none of it operating, with OpenAI expected to be the largest tenant. SoftBank's unit lost roughly $3.2bn in the first half of 2026 on $139m of revenue, almost all from its legacy energy business, and describes itself as 'substantially dependent' on OpenAI as both tenant and investor, per CNBC. It is expected to list at $50bn or more. The chain that holds this together runs through one company. Nikkei counts Nvidia's $3bn investment in SB Energy (the landlord), $30bn in OpenAI (the tenant), and a guarantee of up to $105bn to SB Energy if OpenAI cannot pay the lease on the Ohio site. OpenAI has put $500m into SB Energy and will receive warrants; SB Energy has promised to buy OpenAI's services. Building the contracted sites will cost more than $170bn, most of which has to come from outside investors — the same investors Nvidia and the banks are now courting for a $500bn AI-infrastructure raise. Here is the mechanism a reader should take away. Data-centre project finance is lent against the lease. The lease is worth what the tenant's promise is worth. The tenant's promise is backed by the chip vendor, who sells to the tenant and owns part of the landlord. SB Energy's chief executive said as much: Nvidia's presence 'helps us to unlock things like investment-grade financing'. Investment-grade debt on a company with no revenue is only possible if the buyers are, in substance, lending to Nvidia. We do not know from the filing how the $105bn guarantee is carried on Nvidia's own books, or the term of the leases behind the $439bn. The market is not asking: Nvidia is up 7% in five days and 1.6% off its high.

Hidden leverage and shadow banking2 Sephigh

A 24-year-old's margin call, an endowment's record year

A leveraged AI bet turned into a record endowment year and a two-thirds loss in the same twelve months, with the loss landing in a different fiscal year to the gain.

Situational Awareness, the AI-focused hedge fund founded by 24-year-old Leopold Aschenbrenner, lost more than two-thirds of its $45bn in July when margin calls forced it to sell the bulk of its public portfolio to Citadel. Bloomberg now reports that the University of Virginia's endowment was one of the few institutions invested. UVA's long-term pool returned 27% for the year to 30 June and reached a record $18.8bn. The arithmetic of that sentence is the story. The fund rose 200% in 2025 and more than 400% in the first half of 2026 'supercharging its returns with borrowed money'. Those gains were marked at 30 June, the endowment's year-end, so they flow into the 27% the university will report in the coming weeks. The collapse happened in July. As Bloomberg puts it, the rout 'will be next year's problem'. UVIMCO does not disclose the size or timing of its positions or the identity of its roughly 90 outside managers, so the actual exposure is not knowable from the outside. Two things about the plumbing. First, leverage in AI equities does not only sit in the visible places; it sits in funds like this one, financed by bank prime brokerages, and it shows up in institutional portfolios as a single return number. Second, when it unwound, the system handled it: the margin calls were met by selling to one large buyer within a day, and the SEC's subpoenas to the banks, reported 24 August, are the regulator asking what happened rather than a rescue. That partly cuts against the Crash Lab thesis — this leverage was inside the measured system and the measured system caught it. The unresolved question is how many other pools of borrowed money were long the same names in July, and whether the next forced seller finds a Citadel waiting.

Crypto and TradFi contagion2 Sepmedium

The bitcoin treasury company worth less than its bitcoin

When a treasury company trades below its coins the only way to grow is to stop being a treasury company — the model, not just the stock, has failed.

Nakamoto Inc., the public vehicle David Bailey built to stockpile bitcoin, has fallen about 99% from its peak since the May 2025 merger that created it, Bloomberg reports. Bailey had raised roughly $760m for the company; his hedge fund had returned 640% on similar bets. 'We bought Bitcoin at the top, and then Bitcoin performed very poorly,' he said. The mechanism that broke is worth spelling out. A digital-asset treasury company works only while its shares trade above the value of the coins it holds. At a premium, it issues stock, buys more coin, and each purchase is accretive to existing holders — a flywheel. Nakamoto now trades 'well below the value of its Bitcoin holdings', which means issuing shares to buy coin destroys value. Bloomberg's phrase is exact: 'the financing engine has broken down'. Bailey's response is a conventional company's response — buy cash-generating businesses, prioritise buybacks of the discounted stock over adding bitcoin, avoid new debt. Bloomberg describes this as part of 'a broader rout in digital-asset treasury companies that has wiped out the premium'. The tape supports that reading only partly. Bitcoin is $77,098, up 21.6% in twenty days; Strategy is up 25.4% over the same period, tracking its coin rather than re-rating. On the week, though, the ether treasury names have lagged badly: BitMine is down 7.4% and SharpLink 2.6% against ether-adjacent bitcoin down 0.9%. Equities falling faster than the coin they hold is what premium compression looks like from the outside. We do not have a current discount figure for Nakamoto, and Bloomberg does not give one. The contagion question for this beat is narrow and specific: which of these vehicles carry debt, and what happens to that debt when the equity that was supposed to service it trades below the collateral.

The AI capex bubble2 Sepmedium

China built 150 compute centres. They run at 30%.

It is the first place the AI buildout has run long enough to show utilisation numbers, and they are half of what the depreciation schedules assume.

A Caixin commentary offers a look at what the AI buildout looks like a few years further along. By late 2024 nearly 150 'intelligent computing centres' were operating in China with another 400 planned. Average utilisation hovers around 30%; many facilities in eastern and western provinces run below 40%. Government entities and local government financing vehicles — the off-budget borrowers that funded the property boom — account for about 35% of the investment. The author's diagnosis is that local officials 'build facilities before securing clients, prioritizing peak computing power over vital software and data ecosystems'. And then the sentence that should interest anyone holding a US data-centre bond: chips iterate annually, so the realistic useful life of AI hardware is three to five years, while some cloud providers 'stretch accounting depreciation to six years, masking the financial peril'. The proposed remedy is a red alert at 30% utilisation, halting new approvals and cutting special-bond quotas. This is commentary, not audited data, and China's centres are built for a different customer base than Ohio's. But set it beside the Reuters review of US grid requests — Texas data-centre power applications up from 48GW to 474GW, only 20 of more than 100 proposed Pennsylvania projects bothering to seek permits — and the pattern is the same in both systems: capacity is being committed ahead of demonstrated demand, financed by entities whose losses do not have to be marked promptly. In China that entity is the LGFV. In the US it is the project-finance vehicle lending against a lease. The question a Chinese compute centre at 30% utilisation poses to the US market is what a stalled hyperscaler tenant would pose to SB Energy's lenders: who eats the depreciation on a three-year asset financed over six.

Bond market dysfunction2 Sephigh

Treasury names the leak in its own market: Japan

The largest foreign holder of Treasuries now has a reason to sell, and the US Treasury has publicly acknowledged the channel.

A senior US Treasury official has said out loud what the rate move implies. 'Japan is such a large buyer of sovereign debt, particularly developed market debt around the world,' Undersecretary Erin Browne told Nikkei. 'It tends to have an impact, particularly on U.S. markets.' Japan's 10-year yield crossed 3% on Tuesday for the first time since 1996; the US 10-year is at 4.80%, a 20-month high, and the 30-year at 5.27% — back to the level moments before Scott Bessent expanded Treasury buybacks on 19 August, according to Bloomberg. The channel is simple. Japanese institutions hold more than $1tn of Treasuries. For decades the yield at home was close to zero, so owning US debt made sense even after the cost of hedging the dollar. As Japanese yields rise, the case for holding a foreigner's bond weakens and money comes home. Nikkei's phrasing: rising Japanese rates mean investors are 'more likely to withdraw funds from the American and repatriate capital'. Browne's admission that this is already 'impacting' the US market is the Treasury conceding that its own long end is partly priced in Tokyo. What happens next is scheduled. Money-market brokers price a 94% chance the Bank of Japan raises rates on 17–18 September; Governor Ueda says a hike is on the table at every meeting. Browne did not object — 'market expectations oftentimes influence monetary policy' — which reads as Washington accepting a hike that will push Japanese yields further up. Bessent, she said, will soon release a US deficit-reduction plan; Japan's finance minister has pledged to 'steadily reduce' debt to GDP. None of this is dysfunction. Bank of America's Mark Cabana says only that the market 'has not been able to hold any type of significant rate decline'. The 2s10s curve sits at 41bp. But the Treasury has just told you which buyer it is worried about losing, and that buyer's central bank meets in two weeks.

Signs of the times

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

Up 400% by June, down two-thirds by August

Leopold Aschenbrenner, 24, ran a $45bn hedge fund that gained more than 400% in the first half of 2026 on borrowed money, then lost over two-thirds of it to July margin calls. The University of Virginia's endowment was an investor and will report a record 27% year because its books closed on 30 June.

leverageaiendowments

2 SepBloomberg

$50bn valuation, $439bn backlog, zero data centres

SB Energy is filing for a $50bn-plus IPO with 8.8 gigawatts of contracted capacity, none of it operating, and no data-centre revenue at all. Its backlog is roughly two-and-a-half times what it says it will cost to build everything.

ai_capexipocircular

2 SepNikkei Asian Review

Four hundred more, at 30% full

China had nearly 150 AI computing centres running by late 2024 with 400 more planned, at average utilisation of about 30%. A property-style local-government building boom has found a new asset class, and this one depreciates in three to five years.

ai_capexchinaovercapacity

2 SepCaixin

Raised $760m to buy bitcoin. Down 99%.

David Bailey's Nakamoto Inc. raised about $760m to stockpile bitcoin and its shares have fallen roughly 99% from their peak; it now trades below the value of the coins it holds. Its new plan is to buy businesses that make money.

cryptotreasury_companiespremium

2 SepBloomberg

The rumour mill

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

Partly true

“The Shiller CAPE ratio just reached 41 for the first time since 2000.”

Multiple outlets put the CAPE at roughly 41.0–41.4 in August 2026, the highest since September 2000; the quibble is only over which day it first crossed. The valuation point stands.

Claimed by Wealthion

Partly true

“In private credit, assets that actually have to trade are trading at huge haircuts to what they were previously marked at.”

True for the listed vehicles: the average public BDC trades at 0.75x NAV and more than 70% trade below 0.80x, per The Lead Left, with Blue Owl Capital Corp at a 22% discount. It is not established for private credit as a whole, most of which never trades.

Claimed by Thoughtful Money (Adam Taggart)

Partly true

“Real estate lenders are loosening standards again — lower down payments, higher LTVs, loans without proper documentation — the same late-cycle behaviour seen before 2008.”

One non-QM lender cut its FICO floor to 550 with alternative documentation, but the Fed's July Senior Loan Officer survey found residential mortgage standards basically unchanged. This is a niche, not 2006.

Claimed by The Rich Dad Channel (Robert Kiyosaki)

Earlier developments

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

The AI capex bubble2 Sepmedium

Neocloud debt: $19bn in a week, chips as security

The AI buildout is now being funded by debt secured on the hardware itself, arranged increasingly by private credit rather than banks, at exactly the point where public equity has stopped paying up for the borrowers.

JPMorgan began sounding out lenders on 27 August for a $5bn debt package for Volta Infrastructure Holdings, a data-centre developer founded roughly seven months ago that raised $300m at a $2.4bn valuation in early August, with Nvidia, Andreessen Horowitz, Altimeter and Michael Dell among the backers. In the same two days, Nebius upsized a debt offering to $5bn on top of $5.75bn of convertible notes it had just sold; Blue Owl arranged $2.4bn for Iren — a $1.2bn term loan and $1.2bn of senior secured notes — earmarked for Nvidia Blackwell Ultra GPUs at a campus in British Columbia; Lambda closed a $926m term loan B; and Core Scientific finalised $600m, most of it a letter-of-credit facility. Call it $19bn of new neocloud debt in a week, on top of the $182bn of bonds PIMCO counts from hyperscalers and chipmakers this year. Two things about the plumbing. First, the collateral is the chips and the contracts to rent them. Iren's loan is not general corporate borrowing; it is staged purchases of Nvidia hardware, to be repaid from compute revenue. That is a cash-flow loan against an asset whose resale value nobody has tested at scale. Second, who is lending. Volta's package is a bank syndication; Iren's is Blue Owl — a private credit manager — which puts a GPU-backed loan on balance sheets that mark quarterly and never have to sell. Nvidia sits on both sides of the Volta deal, as shareholder and supplier. If Volta borrows the full $5bn against a $2.4bn valuation, its debt would be more than twice its equity value. That is our arithmetic, not a term sheet; the leverage only makes sense if tenant contracts are already signed, and the reporting does not say whether they are. The market is pricing the vendor and the borrowers differently. Nvidia is up 7% in five days and 1.6% off its high. CoreWeave is down 8% on the week and 25% off its high; Nebius is 27% off; Applied Digital 21%. Lenders are extending credit to the tier equity investors are selling.

Private credit and BDCs2 Sepmedium

Eight in ten BDC software loans marked down

Private credit's largest single bet — loans to software companies secured on their own subscriptions — is where the markdowns are concentrating, and the listed vehicles are being bid up regardless.

Reuters has gone through the filings of 44 US business development companies. Among ten with comparable disclosures, non-accruals — loans that have stopped paying interest — were 3.4% of portfolio cost at 30 June, up from 2.5% at year-end. About 4% of all borrowers were marked below 80 cents on the dollar, against roughly 1% through 2023–25. And 81% of software loans have been marked down this year, against 40% for everything else. The analysis does not say by how much, and we have not seen the underlying table. Jefferies' fund-level numbers say the same thing from a different angle: Blackstone Secured Lending's non-accruals at 3.6% from 0.3% in early 2025, Golub at 2.9%, Blue Owl at 2.8%, Ares Capital at 2.4%. FS KKR and Crescent both told investors a 'handful of investments' accounted for most of their markdowns. The software figure matters because of how these loans were built. Much of direct lending's growth went into private-equity-owned software companies on 'recurring revenue' loans sized against subscription income rather than assets. The collateral is the revenue. If customers churn, or AI lets them pay less for the same function, the loan and the security weaken together. A mark below 80 is usually where a lender has stopped assuming full recovery; at 4% of borrowers, that is the pipeline for next year's non-accruals. This is a re-measurement of the second quarter, not a new event, and fragility does not move on it. But note the tape: BXSL and ARCC are within 1% of their highs, FS KKR is up 12% in a month, Blue Owl's BDC is up 5%. Investors are treating the software markdowns as a contained handful. The below-80 bucket says the handful is growing.

Fed, Treasury and policy2 Sephigh

Japan will hike into a 3% ten-year

The world's cheapest funding currency is about to get more expensive again, in the middle of a global long-end selloff that its own savers have historically helped absorb.

Kazuo Ueda said on Tuesday that the Bank of Japan will consider a rate hike 'thoroughly in every meeting, including the next one', on 17–18 September. Money-market brokers price a 94% probability of a quarter-point move to 1.25%. It would be the sixth hike since March 2024, and it comes the same day the 10-year JGB yield touched 3% for the first time in 30 years. Ueda listed the drivers of the global yield rise as inflation from the Middle East, 'strong demand for funding AI investment' and fiscal worries — a central bank governor naming data-centre borrowing as a cause of sovereign yields. The yen has not helped. It fell to 163 per dollar in July, a 39-year low, and stayed weak after a joint US–Japan intervention on 31 July. Ueda met Scott Bessent on Sunday; Bessent said afterwards that Japan's era of aggressive stimulus 'is now over' and has entered 'Takaichinomics'. Bessent is simultaneously trying, and so far failing, to cap his own long end with buybacks. Why a 25bp move in Tokyo matters here. For two decades Japanese insurers and pension funds bought foreign bonds because domestic yields were near zero, and global investors borrowed yen at nothing to fund positions everywhere else. At 3% on a domestic 10-year, the hedged yield on a Treasury is no longer obviously better than staying home. We cannot see the flows in real time; the inference is ours. But if the largest foreign holder of Treasuries has less reason to add, that lands on a US long end already at 5.26% with a Treasury that has just doubled its buybacks. The yen ETF rose 0.9% today. That is the market beginning to believe Ueda.

Household credit2 Sepmedium

Card losses ease. Card delinquency piles up.

Consumer-lender shares are trading on falling write-offs while the stock of seriously delinquent debt sits at its highest share in the series, and that stock is what eventually gets written off.

The Federal Reserve's series for credit-card balances 90 or more days delinquent stood at 12.8% in the second quarter, up from 7.6% in the third quarter of 2022. In the same window, issuers' August net charge-offs improved to 3.28% from 3.42% in June; Bread Financial was at 6.60%, American Express 1.7%. The Fed's all-bank card charge-off rate for Q2 was 3.70%, and 30-plus-day delinquency on bank cards fell to 2.85%. Those numbers point in different directions, and we cannot fully reconcile them. They are built differently — one is the share of outstanding balances that are seriously late, the others are rates of new loss and early arrears. But a stock of seriously delinquent balances rising while new delinquencies and write-offs fall usually means one thing: borrowers who fell behind are staying behind rather than being written off or curing. The pile ages. That is our reading, not the Fed's. Elsewhere the picture is plainly worse at the bottom. The New York Fed has 5.49% of auto-loan balances 90-plus days delinquent, with subprime near a 15-year high. FICO counts 3.2m people — 15% of student-loan borrowers with a payment due — reported 30 days late in the past six months. Attom counted 227,500 properties entering foreclosure in the first half, up 21% on a year earlier and 28% on 2024. Higher-income borrowers show 'only modest changes'. The lenders' shares say none of this is a problem. Affirm rose 6% today, Capital One 2.7%, Synchrony 2.5%; Oportun is up 24% in a month. The market is pricing the charge-off flow. The 90-day stock is the thing to watch — the same kind of consumer receivable British banks are now pledging to the Bank of England as collateral.

Crypto and TradFi contagion1 Sepmedium

Twenty-one banks, one stablecoin, no customers yet

The largest banks are building a bill-funded money instrument outside the deposit system because Tether has shown how profitable it is — and the first bank attempt has $12.5m of takers.

Twenty-one banks including Goldman Sachs, Bank of America, Citi and Deutsche Bank said on Tuesday they will form a company this year to issue a dollar-pegged stablecoin in the first half of 2027, with a euro version to follow. The group has doubled from ten banks since October 2025. It will compete with Qivalis, a 37-institution consortium planning a euro coin this year. Some banks, BBVA among them, are in both. The business case is not mysterious. Tether has issued more than $180bn of tokens and, in Reuters' words, 'made billions in profits by investing the reserves in assets including US Treasuries'. A stablecoin issuer collects dollars, holds bills, keeps the interest and pays the holder nothing. It is a money-market fund that does not have to be one. The demand case is less clear. Société Générale, the only major bank to have actually launched a dollar stablecoin, has $12.5m in circulation. Circle, meanwhile, minted about 5bn USDC in a single week. For the thesis this cuts both ways. Bringing stablecoin reserves inside regulated banks moves a Treasury-holding shadow-money business back toward supervision — the opposite of the migration we usually write about. But a bank-branded token is not a deposit: no insurance, no lending against it, and a run on one is a forced sale of bills rather than a queue at a branch. Twenty-one banks have decided the reserve yield is worth building that. Whether anyone wants to hold their coin is, on current evidence, an open question.

Fed, Treasury and policy2 Sephigh

The buyback floor has been tested. It is gone.

The one intervention the Treasury has tried against the long-end rout has been fully absorbed, and the remaining options are either debasement by another name or a Fed that has just said it will not help.

On 19 August the Treasury said it would at least double its buybacks of long-dated bonds, from $2bn to $4bn per operation, running from 9 September to 4 November, framed as liquidity support for the 10-to-30-year sector. The 30-year yield fell. It has now come all the way back: 5.27% on Tuesday, which Bloomberg describes as the level seen moments before the announcement, and the 10-year at roughly 4.80% is more than 10 basis points higher than it was then. Note the timing. The enlarged operations have not started. The market priced the announcement and discarded it before a single $4bn bid was placed. Scale explains why. In the week to 29 August the Treasury sold $797bn of securities, $235bn of it in notes. Against that flow a $4bn buyback is not a price tool; it only ever worked as a signal. Mechanically a buyback retires old, illiquid off-the-run long bonds and the Treasury pays for them with new issuance — in practice, we take it, shorter paper. That smooths trading in the old bonds. It does not change how much duration the market is asked to hold unless the funding shifts to bills and stays there, and doing that at scale is what the commentary quoted by Reuters means by "debasement" worries. It is worth being precise about what this is not. Kashkari says the Treasury market is "functioning as it should"; no auction has tailed in the reporting we have; swap spreads narrowed rather than blew out. This is a repricing of term premium — $40tn of federal debt, $182bn of AI-related bonds from hyperscalers and chipmakers this year by PIMCO's count, an oil shock feeding inflation — not a plumbing failure. The distinction matters for the meter: price is fragility, plumbing is ignition. What is left in the kit is bigger and more political. Shifting issuance toward bills, or the Fed. Warsh used Jackson Hole to call 2% "a firm and fixed target". No help from that side.

The dollar, gold and reserve status2 Sephigh

Eighty-six tonnes leave New York

Reserve managers are not selling the dollar; they are quietly repricing the risk of holding assets under US custody, which is the slow-moving part of the reserve-status story and the part that does not reverse.

The Dutch central bank said on Wednesday it has moved 86 tonnes of gold out of New York and Ottawa and into London. Its total stock is 612.4 tonnes, valued at €72.2bn at the end of 2025. Before the move 31.3% of it sat in New York and 19.7% in Ottawa; each is now 18.5%. London's share went from 18.1% to 32.1%. The stated reason is "increasing geopolitical unrest" and the fact that gold in London "could be traded more easily" — "the quickest for DNB to deploy in a crisis situation". President Olaf Sleijpen added: "We assume that we will never need to deploy the gold." The logistics are their own small story. Part of the shift was done by buying in one place and selling in another. Part was physical: 27 tonnes came from North America to Zeist, and the same weight left Zeist for London, so that no bars had to be melted and recast. It sits alongside the Reserve Bank of Australia becoming, per the AFR, the first major central bank to confirm cutting its US dollar holdings — reportedly a target share of 45% from 55%, with the euro up to 30% from 20%, though those figures come via a Deutsche Bank note rather than the RBA itself. IMF data and surveys now show reserve managers planning to trim dollar exposure over the coming decade rather than build it. Read this carefully. It is not a dump. The Dutch left 18.5% in New York and moved to London, not Amsterdam or Beijing; the RBA moved into euros, not gold. What has changed is the working assumption that the United States is the frictionless default place to keep things, and that gold is an ornament. A central bank spelling out that it wants bullion where it can be sold or lent fast is describing a scenario it has started to plan for. Gold at $4,419 is 4% higher than twenty days ago despite a 4% fall on the week.

Household credit1 Sepmedium

Australia's housing boom finds out who its lenders are

Australia is running the experiment — a property downturn hitting a $200bn private credit sector that lent to developers on retail money — that the US thesis assumes is coming, and the distinction between closed-ended losses and open-ended gates is exactly the one to watch.

Sydney home prices are down about 7% from their February peak, according to Cotality, and values are now falling in 93% of suburbs across Australia's capital cities. Sales volumes in Brisbane, Perth and Sydney are down more than 20% year on year. Mortgage applications at the big banks have dropped as much as 20% since the May Budget removed tax incentives for property investors. About 60% of Australian household wealth is in property, by Commerzbank's estimate. That is the demand side. The interesting part is the chain behind it. On 25 August Bathla Group, one of Sydney's biggest developers, went insolvent owing A$3.3bn. Its lenders are not mainly banks. Ray White Capital, with roughly $242m out to Bathla, told investors on Tuesday that the homes securing those loans will be released into a market where their value is immediately up to 15% lower than thought, and that the vibe is "GFC-like". On 25 August MA Financial capped redemptions on its $2.3bn secured property loan fund at 1% of funds a month. On 28 August CVS Lane suspended applications and redemptions outright. On 27 August ASIC warned of the "first significant cracks" in the country's $200bn private credit industry. The mechanism is the one we keep coming back to. Developer credit that banks would once have held and provisioned is sitting in funds sold to retail investors, some of which promised monthly liquidity against loans that pay out when a building is finished and sold. When the buildings are worth 15% less, the loans cannot be marked at par and the redemption promise cannot be kept. One detail cuts the other way. Ray White's note reminded investors they were "never promised the opportunity to redeem" because its funds are closed-ended. That is the structure working as designed: losses, but no run. The gates at MA Financial and CVS Lane are the mismatch.

Hidden leverage and shadow banking2 Sephigh

The Bank of England, lender against store cards

The BoE has become the marginal funder of British store-card and car-lease debt as a side effect of unwinding QE, which tells you both how thin private funding for that credit is and how quickly the demand for a backstop is growing.

On 18 August British banks pledged £1.9bn of the Bank of England's highest-risk "Level C" collateral at its weekly auction for six-month money, according to a Reuters review of BoE filings. That is the most since March 2020 and three times the prior week. In total the Bank now holds about £17.8bn of Level C collateral through its Indexed Long-Term Repo, up from £8.7bn a year ago and under £1bn in mid-2024. Level C is the bucket for securities backed by high-interest store-card debt and vehicle leases — including several products the European Central Bank has tightened rules to exclude. The Bank's own description is that this is an "intended consequence". The mechanism runs like this. Quantitative easing put £895bn of reserves into the banking system between 2009 and 2021. Reversing it drains those reserves. Banks still need them for wholesale payments, so they borrow them back from the Bank, and the cheapest asset to pledge is the one you cannot easily fund anywhere else. The result is that the central bank is now, at the margin, the funder of British consumer credit. That is the opposite direction to our thesis — risk moving toward the official sector rather than away from it — but the underlying statement is the same: these loans need a backstop to be liquid. The ECB's worry, as Reuters puts it, is that a central-bank stamp increases demand for securities that could be hard to sell in a crisis. Two caveats. The Bank applies haircuts and says it protects itself through "robust risk management", and £17.8bn is small against the balance sheet. The number that matters is the slope: roughly eighteenfold in two years, and tripling in a single week in August.

Crypto and TradFi contagion2 Sepmedium

London's stock exchange goes on-chain, overnight, with Kraken

Tokenised equities on a major regulated exchange create a channel for listed-company risk to enter crypto collateral chains, and for crypto-market hours to set prices the cash market has not yet seen.

The London Stock Exchange plans to launch tokenised stocks backed by shares of London-listed companies, in partnership with Payward, the parent of crypto exchange Kraken. The tokens will trade on the LSE's overnight venue and are due next year pending regulatory approval. Simon Mason, the exchange's head of trading venues, calls it "an important inflection point for global market structure" and says part of the demand comes from crypto investors looking to diversify into other asset classes. Payward already runs xStocks. Each token is backed by one share, sits in any digital wallet and trades at any hour, but carries only limited shareholder rights. The LSE says it is also developing tokens with full rights, working with issuers so the token is "a true representation of the underlying equity". The plumbing question is where the share actually is. A token is a claim on a share held by a custodian. When the token trades at 3am London time, the cash market that sets the share's price is closed, so price discovery moves to a venue without the underlying. Whether the token can be pledged as collateral in crypto lending markets, what happens on a corporate action, and who bears the custodian's failure — the statement does not say. This fits a pattern in the window: a 21-bank consortium planning a dollar stablecoin for 2027, Vanguard and Wellington testing tokenised money-market shares as collateral, Standard Chartered and HSBC moving tokenised deposits bank-to-bank. Each is another bridge across which stress in one system can reach the other. For balance: a regulated exchange doing this under approval is the opposite of a shadow venue doing it without. That is genuinely fine. What we cannot yet see is what the tokens will be used for once they exist.

The AI capex bubble2 Sephigh

The vendor is fine. The borrowers are not.

The market has started separating the AI chip vendor from the leveraged entities that generate its demand — a distinction that only matters if one of them is wrong.

Nvidia closed 2 September at $222.18, up 5.9% over five sessions and 2.6% below its recent high. Over the same five sessions CoreWeave fell 8.0% and now sits 24.8% off its high; Applied Digital fell 10.1% and is 22.7% off; Core Scientific fell 8.3% and is 29.9% off; Nebius fell 6.6% and is 28.0% off. That gap is the whole argument in one chart. A growing share of Nvidia's demand is now paid for with money borrowed against Nvidia's own chips, or with Nvidia standing behind the payments. Blue Owl arranged $2.4bn for IREN — $1.2bn of senior secured term loan and $1.2bn of notes — earmarked for staged purchases of Blackwell Ultra GPUs rather than general corporate use (sweep). GMI Cloud's Taiwan GPU-backed loan drew about NT$30bn of commitments against the NT$13.9bn it sought, with Nvidia agreeing to lease any unused compute for up to six years and take half the revenue GMI earns from other AI companies (Bloomberg). So the equity market is currently pricing the receivable as good and the payer as impaired. Both cannot stay true. If the neoclouds' cost of equity keeps rising, orders slip and the vendor's revenue is not as safe as $222 implies. If the orders hold, then 25–30% drawdowns are mispricing the borrowers. Note what has not moved: high-yield spreads at 265bp, tighter than a month ago. The stress is entirely in equity. That is exactly where you would expect it to appear first when most of the debt is privately placed and never marked.

Private credit and BDCs2 Sephigh

A chief executive removed by redemption requests

The retail-facing semi-liquid fund is the main channel through which private markets risk reached ordinary savers, and it is now being tested in three jurisdictions at once.

Partners Group has replaced David Layton as chief executive. He becomes chief investment officer from 1 January; Roberto Cagnati and Juri Jenkner take over as co-CEOs. The FT reports the change came alongside first-half fee revenue of SFr1.1bn, down 7%, profits down 13% to SFr502mn, and performance fees down 39%. Shares fell as much as 8%. A filing showed the firm's flagship US private equity evergreen vehicle shrank from $15.8bn to $15.1bn in the three months to end-June (FT). The mechanism is worth being precise about. Evergreen or "semi-liquid" funds sell private assets to wealthy individuals with a promise of periodic withdrawals, usually capped at a set percentage of net assets per quarter. The cap is not a failure of the structure; it is the structure. Partners Group used it earlier this year on its $8.6bn European vehicle and moved to do the same on the US one. What the cap does is convert a liquidity problem into a franchise problem. Once investors learn the queue exists, new money stops arriving, assets shrink, fees fall with them, and the manager either sells holdings into a slow market or accepts being smaller. That is what a 7% fee decline and a co-CEO announcement look like from the outside. The same pattern is running in Australia, where MA Financial capped redemptions at 1% of funds under management per month on its A$2.3bn secured property loan fund and CVS Lane suspended applications and redemptions outright (sweep). The filing does not say whether the exits Partners Group did fund were achieved at carrying value. That is the number that would tell you whether the marks are real.

Hidden leverage and shadow banking2 Sepmedium

The margin bill on $830bn, cut by four-fifths

The largest leveraged position in the world just had its funding cost reduced by regulatory design, and the risk moved from bilateral dealers into a single clearinghouse model.

Estimates drawn from recent Federal Reserve work put large hedge funds' gross US Treasury exposure at about $4.0tn — $2.4tn long, $1.6tn short — with $3.0tn of repo borrowing and roughly $830bn of cash-futures basis trades outstanding. Ninety per cent of that repo borrowing sits in the fifty largest funds (analysis). The basis trade is simple and enormous. A fund buys a cash Treasury, sells the matching futures contract, and pockets the small gap between them. The gap is measured in basis points, so the trade only pays at very high leverage, funded overnight in repo. What kills it is never the spread going the wrong way. It is the margin going up: that is what happened in September 2019 and March 2020, when funds had to sell Treasuries into a market that had no bid, and the Fed ended up buying them. Since 30 April, customer-level cross-margining between CME and FICC has been live, letting a fund net its cleared Treasury positions against offsetting interest-rate futures. The reported saving is up to 80% of margin on offsetting exposure. Two further SEC comment windows closed on 31 August, one of which would let broker-dealers margin customer Treasury positions on a net omnibus rather than gross basis. Read this two ways and both are true. Central clearing means one entity now sees both legs and can net them, which is genuinely safer than a web of bilateral trades. But cutting the everyday margin bill by four-fifths mechanically permits more position per dollar of capital, and it puts the whole thing inside a model that assumes cash and futures stay correlated. In a dash for cash, that is the assumption that goes first. These are secondary readings of Fed work, not a release from the Fed in this window. Treat the exact figures accordingly.

Household credit1 Sepmedium

Delinquencies falling, charge-offs rising

Whether household credit is healing or just working through a bad cohort depends on which of these two series you look at, and the equity market has picked one.

Two consumer credit numbers moved in opposite directions in the second quarter. The charge-off rate on credit card loans rose from 3.66% to 3.70%. The 30-plus-day delinquency rate on bank-issued cards fell to 2.85%, and the all-card 60-plus rate fell to 2.69% (sweep). That is not a contradiction. Delinquency is the flow in; charge-off is the flow out, roughly six months later, when the lender gives up. A quarter where charge-offs rise and early delinquencies fall is the pipeline emptying: losses being realised from a cohort that has already stopped growing. It is what the far side of a credit deterioration looks like — assuming nothing new enters the pipe. The reading that would break the interpretation is 30-plus delinquency turning back up. Elsewhere the picture is worse and older. New York Fed Q2 data show 5.49% of auto loan balances 90-plus days delinquent, with 3.00% of balances transitioning into 90-plus over the prior four quarters, and subprime auto delinquency near a fifteen-year high though showing signs of stabilising (sweep). FICO reports 3.2 million people — 15% of student loan borrowers with a payment due — reported at least 30 days late in the previous six months. The consumer lenders rallied into all of this on 2 September: Capital One +2.6%, Synchrony +2.9%, Ally +2.8%, Oportun +5.5% and up 24.5% over twenty days. The market is trading the improving flow, not the realised loss. Given that the NY Fed also finds higher-income borrowers barely changed while lower-income borrowers stay distressed, the aggregate everyone is pricing is an average of two different economies.

Crypto and TradFi contagion2 Sephigh

Fifty-eight banks, two stablecoins, no customers yet

Fifty-eight banks are building the plumbing for tokenised dollars while the one live bank product has $12.5m outstanding — the capacity is being built well ahead of the demand.

Twenty-one institutions including Goldman Sachs, Bank of America, Citi and Deutsche Bank said on Tuesday they will form a company this year to issue a dollar-pegged stablecoin in the first half of 2027, with euro and other G7 currencies to follow. They will compete with a separate consortium of 37 institutions, called Qivalis, launching a euro coin later this year. Some banks, including BBVA, are in both (Livemint). The mechanics of a stablecoin are the mechanics of a money market fund that settles like cash. The issuer takes dollars, buys Treasury bills, keeps the interest and hands back a token. Tether has more than $180bn outstanding and has made billions doing precisely this. For a bank, the defensive logic is obvious: a dollar held as a stablecoin is a dollar not held as a deposit, and deposits are what funds the loan book. The number to hold onto is $12.5m. That is the circulation of the dollar-backed token Société Générale launched last year through its digital assets arm — the first major bank to try. Reuters' own framing is blunt: there are few signs of demand for bank-issued stablecoins. Demand today is for the offshore, lightly supervised version, which is also where the reserve risk sits. The same week, the London Stock Exchange said it will launch tokenised UK equities on its overnight trading venue next year in partnership with Payward, Kraken's parent, and is separately developing tokens carrying full shareholder rights (Bloomberg). None of this is fragile yet. It is infrastructure being laid for a flow that has not arrived. Worth watching for the moment it does, because a bank consortium holding several hundred billion of bills against instantly redeemable tokens is a maturity mismatch with a familiar shape.

Fed, Treasury and policy2 Sephigh

Ninety-four per cent priced for a Tokyo rate rise

Japan's institutions have been the world's marginal buyer of long-dated government debt for three decades, and their domestic alternative just became viable.

Governor Kazuo Ueda said on Tuesday that a rate hike "will be discussed thoroughly in every meeting, including the next one" — that is 17–18 September. Japanese money market brokers price a 94% probability of a 25 basis point rise to 1.25% (Nikkei). He said this on the day the 10-year JGB yield touched 3% for the first time since September 1996. The five-year hit a record 2.265% and the two-year a 31-year high of 1.81%. The government assumed a 3% long-term rate when calculating debt-servicing costs in the fiscal 2026 budget (CNA). Why this matters outside Japan: for thirty years, near-zero domestic yields pushed Japanese life insurers, banks and pension funds into foreign long duration — Treasuries, gilts, French OATs, mostly currency-hedged. They have been among the most reliable buyers of the long end anywhere. A 3% domestic government bond changes that arithmetic without anyone announcing a decision, and a higher policy rate raises the cost of hedging the foreign holdings they still own. The 30-year gilt at 5.89%, its highest since 1998, and the 30-year Treasury at 5.26% are in part the same trade being run in reverse. Note that Treasury's doubled long-end buybacks have not held the line: 30-year yields are back where they were the moment before the 19 August announcement, and the 10-year is more than 10bp higher (Bloomberg). The yen firmed 1.0% against the dollar on 2 September. It remains close to a four-decade low, which is the other reason the hike is coming.

The AI capex bubble2 Sephigh

The banks want the GPU loan. The stock market doesn't.

When the chip vendor guarantees the cash flow that repays the loan that buys its chips, bank credit committees are pricing Nvidia risk without holding Nvidia paper.

GMI Cloud, an Nvidia partner, went to Taiwanese banks for a five-year loan backed by the revenue from its chips. It asked for NT$13.9bn. About a dozen banks offered roughly NT$30bn ($947m), and Bloomberg reports that if verbal commitments firm up the book grows to about NT$40bn. The money buys GPUs for a data center in Taiwan and is repaid out of the revenue they generate. Read the terms before deciding what the oversubscription means. Nvidia supplies the chips. Nvidia will also lease any unused computing capacity at an agreed price for as long as six years. And Nvidia takes half the revenue GMI Cloud earns from other AI companies. So the lenders are not really betting on GMI Cloud's ability to sell computing power. They are betting on a floor that the chip vendor itself has agreed to provide, on chips the same vendor sold. It is a guarantee on the residual value wearing the clothes of a commercial contract, and it does not appear as debt anywhere on Nvidia's books. This is the pattern we keep finding: the risk is real, someone has taken it, and it sits outside the place where risk is counted. The interesting part is the disagreement. Bank lenders are bidding twice the ask for exposure to GPU cash flows in Asia. Meanwhile the listed companies that build and run these assets have been sold hard: Core Scientific is down 29% in twenty days, Applied Digital 22%, Nebius 12%, CoreWeave 11%. Nvidia itself is up 2.6% over the same stretch. Stock investors are marking down the operators; bank lenders are lining up to lend against the operators' hardware, secured by the vendor. One of those two groups is wrong.

Hidden leverage and shadow banking2 Sepmedium

Store-card loans, pledged at the Bank of England

A twenty-fold rise in consumer-credit paper pledged at the Bank of England is a measure of what UK banks would rather not have to sell.

On August 18, British banks pledged £1.9bn of the Bank of England's highest-risk category of collateral at its weekly auction for six-month funds, the most since March 2020 and three times the previous week. The category, which the Bank of England calls Level C, includes loans linked to high-interest store cards and vehicle leases. Reuters' review of Bank of England filings puts the total stock of Level C collateral from the Indexed Long-Term Repo at about £17.8bn, up from £8.7bn a year ago and under £1bn in mid-2024. The mechanism is straightforward. Quantitative easing put cash in the system; the Bank of England has been reversing £895bn of it. Banks still need central bank reserves for wholesale settlement, so they borrow them, and they post whatever assets they have. Government bonds are the cheap option. When a bank is posting bundled store-card receivables instead, it is telling you something about the composition of what it has left, or about how much it would rather not sell. Two things stop this from being alarming on its own. The Bank of England says the facility is designed to accept a broad range of assets and that it protects itself through discounts and risk management, and £17.8bn is small against a system of that size. The Bank calls the growth an intended consequence of unwinding quantitative easing, and it is. What has changed is that the central bank has become a standing window where UK consumer credit can be swapped for cash, a class of asset the European Central Bank has tightened its own rules to exclude, on the reasoning that a central bank stamp of approval creates demand for paper that is hard to sell in a crisis. This is a twenty-fold increase in two years in a channel almost nobody was watching. We did not know its size this morning. Now we do.

What would change our mind

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

High-yield spreads widening through 350bp, or HYG breaking more than 3% below its high — the credit market has so far refused to confirm any of the stress in the reporting.

Would move the number

A US non-traded private credit fund or BDC gating redemptions, or a listed BDC cutting its dividend on non-accruals — the Australian gates spreading to the US market.

Would move the number

The 10-year Treasury through 5% or a tailed long-bond auction around the 17–18 September BOJ meeting, which would turn a repricing into a funding event.

Would move the number

A state insurance regulator or the FHLB system restricting advances to investment-firm-owned insurers, which would test how the Group 1001 model funds itself without cheap public-backed money.

Would move the number