SoftBank sells Japan's biggest-ever retail bond to fund OpenAI

Archived reading, published Tue, 25 Aug 2026 06:21:01 UTC (35 hours ago). This is not the current state of the meter.

See the live reading →

CRASH-O-METER

Cracking

how close are we

+1 since the last reading

Fragility89

how much tinder is stacked up — moves slowly

Ignition42

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 8 sources and rewrites this page.

Why it moved: Up one to 66: Guggenheim's $1.18bn loan — the only continuously traded instrument on Mark Walter's private insurance-and-credit empire — fell more than five cents to a record 72.5 on Monday, after a lender call last week failed to hold the price. That is a mark being tested and failing inside the exact channel we track, and it is not visible in index spreads (high yield 270bp, investment grade 81bp, VIX 15.85), so it belongs in ignition rather than fragility, which holds at 89.

Reporting from 23 Aug to 25 Aug

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 AI capex bubble24 Aughigh

Japanese households take the OpenAI duration risk

When bank credit committees decline AI-related duration and retail investors take it instead, the loss-absorbing layer of the boom moves to households who cannot mark it or sell it.

SoftBank plans a ¥1 trillion ($6.3bn) retail bond sale — the largest by any issuer in Japan — to fund its commitments to OpenAI. Seven-year paper, indicative coupon 4.3% to 4.9%, pricing 4 September, expected to be rated A by Japan Credit Rating Agency. It is SoftBank's third retail bond this year, after ¥418bn in April and ¥260bn in June. The shares fell 5.3% in Tokyo on the news (Bloomberg). The interesting part is why it is a retail deal. "Banks are finding it difficult to take on the risk given weak deposit growth, the credit rating and the seven-year duration, leaving the deal more reliant on retail investors," Yuuki Fukumoto of NLI Research Institute told Bloomberg. Read that plainly. A credit committee looked at seven years of unsecured holdco exposure to a company whose principal asset is a $60bn-plus commitment to OpenAI, and declined. The same risk was then repackaged at a headline yield a Japanese saver has not been offered in a generation and sold to people who will hold it to maturity because there is nowhere liquid to sell it. This is the migration thesis in a single transaction. Risk does not disappear when a bank passes on it; it relocates to a holder with no capital requirement, no mark-to-market discipline and no risk committee. SoftBank's OpenAI stake is illiquid and unlisted. The bondholder's recovery depends on SoftBank's ability to monetise or refinance it in 2033. The filing does not say what happens if it cannot. Context for the coupon: Alphabet sold $25bn of bonds this month against a $205bn 2026 capex plan, and Bloomberg data cited this week put total borrowing for data centres and AI at more than $410bn in 2026 so far.

Private credit and BDCs24 Aughigh

The only price on a private empire keeps falling

Private-equity-owned insurers hold assets nobody has to sell, so the single traded loan becomes the market's only referendum on whether the marks are real.

A $1.18bn first-lien loan issued by GIH Borrower LLC, the financing entity of Guggenheim Investments, was indicated down more than five cents on Monday at 72.5 cents on the dollar — the lowest since it was priced in November 2024 (Bloomberg). The sequence matters. Guggenheim reported second-quarter revenue down 38% year-on-year, with one measure of earnings collapsing 77%. The loan slid. The firm convened a call with lenders to explain that the drop was largely a delay in recognising advisory fees at Guggenheim Private Investments — the subsidiary whose accounting was the subject of a whistleblower report that helped trigger a federal probe into Mark Walter's empire. The loan briefly recovered. On Monday, Anne Walsh, the asset-management arm's chief investment officer, said on Bloomberg Television that the firm had "engaged with our auditors and we feel that the accounting treatment was appropriate." The loan then made a new low. None of Walter's businesses trade on a public exchange. That is the whole point of this dispatch. When an empire is private — insurers, asset management, sports teams, real-estate debt, loans to Carvana and Wendy's — the only continuous, adversarial price anyone can observe is a syndicated loan quoted by dealers. So the loan is doing all the price discovery for a structure that also sits behind insurance policies. 72.5 is not a default price. It is a price that says lenders do not believe they can verify the numbers. The New York Times reports the federal investigation is examining whether some of Walter's companies improperly characterised tens of billions of dollars of assets. That is a question about marks, and marks are the one thing this business model requires you to take on trust.

Hidden leverage and shadow banking24 Augmedium

Gates are working. That is not the same as fine.

A gate converts a liquidity problem into a duration problem for the investor and postpones the moment anyone finds out what the loans are worth.

Two numbers from this week's private-credit reporting sit oddly together. Redemption requests at the ten largest non-traded BDCs averaged 13% of assets in Q1 2026 and 14% in Q2, against typical quarterly gates of 5%, forcing most managers to restrict withdrawals. And listed BDC shares are at or near their period highs: Blackstone Secured Lending is 0.0% off its high, Ares Capital 0.74% off and up 6% in twenty days, FS KKR up 11%. One widely circulated commentary item claims listed BDCs trade "around 60 cents on the dollar" with quality names at 30% discounts to net asset value. We cannot reconcile that with the tape in front of us. Discounts to NAV exist across the sector, but the listed vehicles are not being sold; they are being bought. The mechanism worth understanding is the gate itself. A non-traded BDC promises quarterly liquidity against loans that take five years to repay. When requests run at 13–14% and the gate is 5%, the manager is not forced to dump loans to raise cash — which is genuinely good news, and the reason this is not a 2008-style run. But three things follow. The queue lengthens, so an investor who asked to leave in March is still in. The marks are never tested, because no loans change hands. And the portfolio left behind is progressively less liquid, since managers meet redemptions with the easiest assets first. Separately, payment-in-kind is now 11% of BDC portfolios, $15.7bn at cost, and the median non-accrual rate across the twenty largest listed BDCs rose to 2.8% of cost in Q2, the highest since 2017. PIK income is interest a borrower could not pay in cash. Both numbers describe the same thing: deterioration that shows up in accruals long before it shows up in a price.

The AI capex bubble24 Augmedium

Now they are selling the electricity

Power contracts were the most concrete evidence that AI demand was real; the market is now discounting them faster than it is discounting the chipmakers.

Last week the split in the AI complex ran along the financing line: debt-funded names fell, cash-funded ones did not. This week the selling moved upstream to power. Talen Energy is down 14.4% in five days and 19.5% off its high. Vistra is down 7.2% in five days and 13.6% in twenty. Core Scientific fell 14.2%, Applied Digital 13.7%, IREN 11.4%, CoreWeave 18.6%, Nebius 21.6% — the last of those after selling a $5bn convertible, one of the largest on record. Over the same five days Microsoft rose 1.6%, Alphabet 1.2%, Amazon 0.3%. High-yield spreads sat at 270bp and VIX at 15.85. The merchant power names were re-rated over the past two years on a straightforward premise: data centres would take every spare megawatt at prices set by scarcity, on contracts long enough to underwrite new generation. That premise is a derivative of the tenants' financing. If a neocloud's cost of capital rises, the megawatt it signs for in 2029 gets signed later, smaller, or not at all. Nvidia's reported decision to raise server prices by more than 15% on systems shipping early next year, driven by memory costs, pushes the same way. A neocloud buys GPUs with debt and rents them by the hour. If the equipment cost rises 15% and the rental rate is already contracted, the extra cost comes out of the equity — or, eventually, out of the lender. What we cannot tell from the tape is how much of the power selloff is AI demand doubt versus power prices or rates. It is not rates: long Treasuries rose 1.5% over the same five days. Nvidia reports on 26 August, which will resolve some of it.

Crypto and TradFi contagion24 Aughigh

Twelve people, half a trillion dollars, SpaceX perps

If a twelve-person offshore venue becomes the reference price for private companies, illiquid marks across the fund industry inherit its liquidity.

Trade.xyz, a market built on the crypto exchange Hyperliquid and run by a team of about a dozen, has generated roughly $500bn of trading volume since launching in October and accounts for more than 99% of activity on Hyperliquid's third-party market system, HIP-3. It lists perpetual futures on crude oil, precious metals, stock indexes and companies that have not yet gone public, including SpaceX (Bloomberg). When the US and Israel struck Iran on a Saturday in February and the oil futures markets were shut, traders there priced the conflict for hours before conventional venues opened. A perpetual future has no expiry. Instead, a funding rate paid periodically between longs and shorts pulls the contract's price toward a reference price. For crude oil that reference is a real, deliverable market, and the mechanism is disciplined by arbitrage. For a pre-IPO company there is no such thing. There is no continuous spot price for SpaceX shares, and no arbitrageur can deliver one against the contract. The price is whatever the positioning of the participants says it is, anchored to an index the venue itself defines. That is not a criticism of the design — it is the design. But it means the instrument is a betting market wearing the clothes of a futures market. The reason this matters beyond crypto is what happens if those quotes start being used. Venture funds, holding companies and secondary marketplaces all mark unlisted stakes with considerable discretion. A visible, continuous, 24/7 price is enormously attractive to anyone who has to justify a mark. Trade.xyz and Hyperliquid are lobbying US regulators to allow pre-IPO perps onshore. Hyperliquid does not currently serve US traders.

Household credit23 Augmedium

Student loans quietly went to 10.6%

Household stress is now concentrated in the two loan books — student and subprime auto — where the losses land outside the banking system entirely.

The number that moved this week and nobody wrote about: the share of student-loan balances 90 or more days delinquent rose to 10.6% in Q2 2026, from 10.3% in Q1, on New York Fed data. Alongside it, auto debt transitioning into serious delinquency hit 3.00%, the highest since 2010, and Fitch's July subprime auto 60-plus-day rate came in at 6.13%, a record for the month, against 0.49% for prime — a ratio of more than twelve to one. A warning about the credit-card figures circulating this week. Some summaries report card delinquency at 6.97%, others at 12.8% or 12.9% of balances. These are not the same measure: one is the rate at which balances transition into serious delinquency over a year, the other a stock of balances 90-plus days past due. Sources have been mixing them. Treat any single card number you see this month with suspicion unless the definition is stated. On the loss channel, the structure is what matters. Subprime auto paper mostly sits in asset-backed securities held by insurers, funds and, increasingly, private-credit vehicles — not on bank balance sheets. Deep-subprime lenders write loans at rates wide enough to absorb 6% delinquency provided recoveries hold, and recoveries depend on used-car values, which none of these releases measure. That is the variable we would want and do not have. The equity market has made its call. Credit Acceptance is up 8.8% over twenty days, Oportun 32.9%, Synchrony 6.8%. Ally is down 2.1%. Either the losses are already in the price at generous assumptions, or the recovery rate is about to matter a great deal.

Signs of the times

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

A dozen people, $500bn of volume

Trade.xyz, run by about twelve people on top of the Hyperliquid exchange, has traded roughly $500bn of perpetual futures on oil, metals and pre-IPO companies including SpaceX since launching in October. When conventional oil markets were closed on the Saturday the US and Israel struck Iran, it priced the war anyway.

cryptomarket structure24/7

24 AugBloomberg

Banks said no, savers said yes

SoftBank is selling ¥1 trillion ($6.3bn) of seven-year bonds to Japanese retail investors at 4.3–4.9% to fund its OpenAI commitments — the largest retail bond in Japanese history. An analyst told Bloomberg the deal leans on households because banks found the credit rating and seven-year duration hard to take.

ai capexretail creditjapan

24 AugBloomberg

Investment grade at a junk yield

QTS Realty sold $3.9bn of bonds to fund a Microsoft-linked data centre in Georgia at a yield of about 7.23% — above what many middle-tier junk bonds pay — and still drew roughly $23bn of peak orders, six times the deal size.

data centrescreditdemand

24 AugBloomberg (via Crash Lab web sweep)

Sold the Lakers to shore up the insurers

Mark Walter used insurance premiums to buy the Dodgers, the Lakers and a Formula 1 team; this month, weeks into a federal investigation into how his companies characterised tens of billions of dollars of assets, he sold a majority stake in the Lakers after owning it for barely a year.

insuranceprivate creditforced selling

24 AugThe New York Times

The rumour mill

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

Confirmed

“Companies have borrowed more than $410bn this year for data centres and other AI investments.”

Bloomberg's own compiled data carries this figure, and it is the cleanest single measure of how much of the buildout is debt rather than cash flow. For scale, Alphabet's entire 2026 capex plan is $205bn.

Claimed by Kitco NEWS

Confirmed

“Nvidia's stock has fallen after each of its last four earnings reports despite beating expectations.”

The per-quarter moves check out — roughly -5.5%, -3.2%, -0.8% and -1.8% after the last four reports, against a run of consecutive beats. Nvidia reports again on 26 August, with the shares already 7.4% lower over five days.

Claimed by Meet Kevin

Partly true

“On 19-20 August, $2.7bn of crypto positions were liquidated in 24 hours, almost entirely shorts — the largest one-sided short wipeout on record.”

Coinglass data support roughly $2.7–3.0bn of liquidations with shorts at about $2.74bn against $250m of longs, and several outlets describe it as the largest short-side liquidation on record. Bitcoin is up 25% in twenty days, which is what a squeeze of that size looks like from the outside; related claims of '$18bn erased in six minutes' are not supported.

Claimed by Coin Bureau

Partly true

“About 80% of major US metros are now buyer's markets and the number of home purchasers in July fell to its lowest level on record.”

Redfin counted 39 of 49 major metros as buyer's markets — a metro where sellers outnumber buyers by more than 10% — and a record-low buyer count of about 966,752, with sellers exceeding buyers by 51.3%. Sales themselves were at a two-year low, not a record low; the record applies to buyers, not transactions.

Claimed by Michael Bordenaro

Earlier developments

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

The dollar, gold and reserve status24 Aughigh

The pressure has to come out somewhere

When a government suppresses the price of its own debt, the adjustment does not disappear — it relocates to the currency, and gold is where you watch it happen.

Citadel Securities has put a name to what the Treasury is doing. Nohshad Shah, its head of EMEA fixed-income sales, wrote to clients that Scott Bessent's expanded buyback programme "amounts to financial repression at the margin," and that "preventing Treasuries from clearing at lower prices does not eliminate that pressure. It merely shifts it elsewhere." Elsewhere, in Citadel's argument, means the currency: lower yields reduce the appeal of the dollar, a weaker dollar raises import prices, and the inflation that pushed yields up in the first place gets worse (Bloomberg). The funding question is the interesting plumbing. Dealers assumed Treasury would pay for the doubled buybacks — $4bn per operation, 10- to 30-year maturities, 9 September to 4 November — by issuing bills. That is a duration swap: the government retires long debt and takes on rollover risk instead. Two senior officials have now told CNBC the near-$1tn Treasury General Account is available for it. The TGA stood at $935bn on 20 August, against the previous administration's stated target of $550–600bn (CNBC). That is a different animal. The TGA is cash sitting at the Fed. Spending it down moves money from the government's account into private hands and adds bank reserves. It is not quite money-printing, but it is the fiscal authority adding liquidity while buying its own long bonds, and our reading is that the market has noticed. The evidence is in the prices. The 30-year hit 5.34% on 21 August, its highest since 2007, having erased the intervention rally within a day. Gold is up 16.9% in twenty days to a record, with the broad dollar index down 2.19%. Bonds rejected the policy. Bullion accepted it.

Crypto and TradFi contagion24 Aughigh

A bitcoin company that now needs dollars

Digital-asset treasury companies convert a volatile asset into fixed dollar obligations, and the reserve tells you which side of that trade is currently under strain.

Strategy has created a second cash pool. The new one, called USD Cash, holds $1.59bn and can be spent on bitcoin, preferred dividends, debt payments, securities buybacks or general corporate needs. It sits alongside the original reserve — now $5.1bn — which without board authorisation may be used only to pay preferred dividends and interest on debt (Bloomberg). The mechanism worth understanding is why a company whose entire thesis is holding a volatile asset forever needs $6.7bn of ringfenced dollars. Strategy's preferred stock and converts are hard obligations: fixed cash out, on fixed dates, denominated in a currency it does not hold. The original flywheel solved this by issuing common stock above the value of the bitcoin behind it and using the proceeds to buy more bitcoin, so the coin count per share rose and the dividends were funded out of an expanding capital base. That only works while the premium exists. Bloomberg reports the valuation premium is still well below earlier-cycle levels and the financing model "remains impaired." So the company sold roughly $2bn of common stock in a week and repurchased $136.4m of its Stretch preferreds. Nansen's Nicolai Sondergaard put the trade-off plainly: "dilution in exchange for flexibility" — the equity raise strengthened the balance sheet but "did not immediately increase Bitcoin-per-share exposure." The share price disagrees with all of this. MSTR is up 25.6% in five days and sits at its period high, against bitcoin's 9.25%. The market is paying roughly three times the coin's move for a company that has just built a dividend reserve. One of those two readings is wrong, and the filing is the one with numbers in it.

Hidden leverage and shadow banking24 Augmedium

Two trillion of Treasuries, bought with borrowed money

The largest marginal holder of US government debt is a levered relative-value trade financed overnight, and the plumbing meant to make it observable has stalled.

The Office of Financial Research published a number on 19 August that deserves more attention than it got: hedge funds now hold about $2.0tn of cash Treasuries against $1.4tn of short Treasury futures. A significant share of that is the basis trade (OFR). Estimates of the trade's size from the CFTC, BIS, Fed and sell-side research span $350bn to $1.5tn — a range wide enough to tell you nobody actually knows. The trade itself is simple. Cash Treasuries trade slightly cheaper than the futures that promise delivery of them. Buy the bond, sell the future, collect the difference. The difference is tiny — a few basis points — so you scale it. You finance the bond in the repo market at a haircut of a fraction of a percent, the lender rehypothecates the collateral into another repo, and the short futures position needs only initial margin against a large notional. The OFR describes exactly this: chains of borrow-and-buy transactions that let funds "expand their positions dramatically." What breaks it is not credit. It is repo rates spiking or margin requirements rising abruptly, at which point the position is unwound by selling cash Treasuries — into the long end that the Treasury is currently spending money to support. Separate Form PF-based analysis puts large hedge funds at roughly 8.5% of privately held Treasuries and $4.0tn of gross exposure. One detail cuts against comfort. Central clearing of Treasury repo plateaued in the first quarter of 2026, and since May the SEC has not required affiliate repo transactions to be centrally cleared provided the affiliate clears everything else. The reform designed to make this leverage visible has stopped advancing.

Household credit23 Augmedium

Subprime auto is past 2008. The lenders are up.

The consumer data and the consumer-lender share prices are telling opposite stories, and only one of them can be right about how much loss is already provisioned.

Fitch's subprime auto 60-day delinquency index printed 6.13% for July, the highest July reading in a series that starts in 1994. One syndicated cut of the Fitch and Equifax data puts the rate at 6.9%, against a 2008 peak of about 5.0%. Prime auto is at 0.49%. The New York Fed's Q2 2026 household report has the share of auto debt transitioning into serious delinquency at 3.00%, the highest since 2010, with 5.49% of auto balances 90+ days late — and record originations of $211bn in the quarter. Credit-card balances 90+ days delinquent are around 12.8%, student loans 10.6%, and 4.7% of all household debt is in some stage of delinquency. Now look at the lenders. Credit Acceptance, which sits at the deepest end of subprime auto, is up 5.4% in five days and 8.8% in twenty, 5.4% off its period high. Oportun is up 32.9% in twenty days. Synchrony is 1.2% off its high. Only Ally, down 2.7% in five days and 8.7% off its high, looks like a stock reading the same data. The two facts are reconcilable, and it is worth being precise about how. Deep subprime lenders do not price loans expecting them to perform; they price a forecast of collections at origination and earn a spread over it. What hurts them is losses coming in worse than forecast, not losses being high. Record originations also flatter the ratio: a fast-growing loan book puts new, not-yet-delinquent paper in the denominator. So the equity market's view is that the loss rate is already in the price. That may be right. It is a bet on the accuracy of forecasts made two and three years ago, at a delinquency level nobody in the series has seen.

The AI capex bubble24 Augmedium

One quarter of capex, raised in one convertible

The financing for AI infrastructure is still available at scale — the price is being paid in the equity of the companies raising it, not in the coupon.

Nebius upsized a convertible bond sale to $5bn on 20 August, described by Reuters as one of the largest on record, to fund data centres and its AI platform. The company spent $5.7bn on capex in the quarter (Reuters). A record convertible therefore buys about eleven weeks of building. Since the deal, the stock is down 21.6% in five days and sits 24.1% below its recent high, still up 12.2% over twenty days. Convertibles are cheap for the issuer because the buyer is paid in optionality rather than coupon, and a large part of the buyer base hedges that optionality by shorting the underlying equity. The prospectus is not in front of us and the reporting does not identify who bought, so we cannot say how much of the fall is hedge flow and how much is sentiment — but the direction is what you would expect, and it is the cost that does not appear in the coupon. The wider number: companies have borrowed more than $410bn this year for data centres and AI investments, on Bloomberg's compilation. Hyperscaler debt issuance alone was $220bn by 10 August, per BNP Paribas. Carson Group puts incremental debt at roughly 32% of hyperscaler capex by mid-2026, against 9% in fiscal 2024. The split in the tape is the same one we flagged last week and it has widened. Microsoft is up 1.6% in five days, Alphabet 1.2%, Amazon 0.3%. CoreWeave is down 18.6%, Core Scientific 14.2%, Talen 14.4%, Applied Digital 13.7%, Nvidia 7.4%. Capital funded from operating cash flow is holding. Capital funded from someone else's balance sheet is not.

The AI capex bubble24 Aughigh

The cost of the asset just went up 15%

Every loan written against AI hardware was sized against a cost per unit that has just moved against the borrower, not the lender.

Bloomberg reported on 22 August, and Reuters and CNBC confirmed, that some of Nvidia's biggest customers have been told the price of servers built around its AI chips will rise more than 15% in many cases, driven by soaring memory costs. The increases apply to Vera Rubin and Grace Blackwell systems shipped early next year. This is not a headline about sentiment. It is a change in the unit cost of the collateral underneath the entire AI financing complex. Consider how a neocloud works. It signs a multi-year contract to supply compute at an agreed price, then borrows against that contract to buy the hardware. CoreWeave closed a $2.6bn delayed-draw term loan at SOFR plus 5.5% on 10 August; Lambda raised $926m of term loan B at S+300. Both were sized against an assumed cost per rack. If the hardware costs 15% more and the revenue contract does not move, the equity cushion in each project thins — and the equity cushion is what the lender is relying on. For the hyperscalers it is simpler and less dangerous: Alphabet's $205bn 2026 capex plan buys less than it did when the plan was written, and Alphabet has the cash flow to absorb it. The tape has already sorted the two groups. Over five days Nebius is down 21.6%, CoreWeave 18.6%, Core Scientific 14.2%, Applied Digital 13.7%, IREN 11.4%. Over the same week Microsoft rose 1.6%, Alphabet 1.2%, Amazon 0.3%. What we do not know is who absorbs the increase — whether neoclouds can pass it through, eat it, or delay deliveries. Nvidia reports on Wednesday with revenue expected above $92bn, and the gross-margin line will be the first place the answer shows up.

The AI capex bubble24 Aughigh

Broadcom's guarantees are now visible in its own spread

A guarantee that never appears as debt is still a claim on the guarantor, and the CDS market is the only place currently pricing it.

The price of five-year credit protection on Broadcom has climbed 28 basis points so far in August — more than either Oracle or SpaceX — and the yield on its 5.15% bonds maturing 2031 is up about 14bp over the same period, Bloomberg reports. The reason is not Broadcom's operating business. It is what Broadcom has promised. The company is in talks to raise more than $60bn of debt for an AI chip financing package expected to benefit Anthropic, and could end up guaranteeing part of a senior-secured tranche. That would follow an agreement earlier this year to backstop most of a $35bn debt package, in which Apollo and Blackstone financed the purchase of custom AI chips to lease to Anthropic. Here is the plumbing. Broadcom sells the chips. Private-capital investors put up the money to buy them. The chips are leased to a customer that is loss-making at scale. And Broadcom promises the investors that if the lease payments do not arrive, it will make them whole. The debt sits in the vehicle, not on Broadcom's balance sheet, so it does not appear in its leverage ratios. The obligation is real all the same. "The rise in Broadcom's CDS seems to me is more specific to their balance sheet than overall angst over AI investment," Impax portfolio manager Tony Trzcinka told Bloomberg, adding it was probably about expectations of further guarantees. This is the running argument of this site in its purest form: risk moved into a structure where nobody has to mark it, and the first place it becomes visible is a derivative market that only professionals watch. Bloomberg counted roughly $70bn of such shadow credit backstops across the sector on 15 August. The equity market has not repriced Broadcom. The credit market has.

Private credit and BDCs24 Augmedium

Fourteen percent want out through a five percent door

Gated redemptions at self-marked funds are how a liquidity mismatch becomes a slow queue instead of a fast price, which is exactly why nobody sees it in a spread chart.

Redemption requests at the ten largest non-traded business development companies averaged 13% of assets in Q1 2026 and 14% in Q2, forcing most managers to activate gates, according to research summarised this week citing Federal Reserve Bank of Boston work. Separate commentary tracks one large vehicle's withdrawal requests moving from about 1% to 4% to 14% of investors asking out. The arithmetic matters more than the adjective. A typical non-traded BDC offers to repurchase up to 5% of net asset value each quarter. If 14% of holders ask to leave, roughly a third get out and the rest are pro-rated back into the queue — and investors who have just been told they cannot have their money usually ask again, larger, next quarter. The gate does not reduce the demand for liquidity. It converts it into a backlog, and the backlog is itself a signal that generates more requests. The second thing to notice is what the leavers are paid. There is no market price for these vehicles. Investors are redeemed at the manager's own quarterly mark on loans the manager originated. Whoever stays owns whatever the mark was wrong about. Underneath, credit is deteriorating slowly rather than dramatically: median non-accruals across the 20 largest listed BDCs rose to 2.8% of cost in Q2, the highest since 2017, and payment-in-kind income is 11% of BDC portfolios, or $15.7bn at cost. We cannot reconcile one figure in this window. Commentary circulating this week says listed BDCs trade around 60 cents on the dollar with high-quality names at 30% discounts to NAV. The tape says otherwise: Blackstone Secured Lending closed at its period high today, Ares Capital is 0.74% off its own, and FS KKR is up 11.1% in twenty days. Someone is describing a different asset class.

Fed, Treasury and policy24 Aughigh

Lending Japan the dollars so it need not sell Treasuries

Three interventions aimed at the long end have moved the dollar and gold instead, which is what fiscal dominance looks like before anyone calls it that.

Scott Bessent is now intervening in three markets at once, and the connections between them are worth spelling out. First, the Treasury has directed US intervention in the currency market to support the yen. Second, Bessent has proposed the Fed raise the limit on its FIMA repo account. Third, the Treasury has at least doubled buybacks of 10- to 30-year debt, from $2bn to at least $4bn per operation, running 9 September to 4 November — and CNBC reports officials consider the $935bn Treasury General Account available to fund them. The FIMA piece is the one almost nobody explains. If the Bank of Japan has to defend a weakening yen, it needs dollars. The largest pool of dollars Japan owns is its Treasury portfolio, so the natural move is to sell Treasuries — into a long end already at its highest yields since 2007. The FIMA repo facility lets foreign central banks borrow dollars from the Fed against their Treasuries instead. Raising the limit means Japan can get the dollars without becoming a seller. Put plainly: the United States is arranging to lend Japan the money not to sell US bonds. As Bloomberg's columnist notes, each of these is defensible in isolation and small in effect. The market's verdict so far is that none of them addresses the deficit, the inflation rate or the Fed's credibility. The 30-year ended last week near 5.28% after briefly falling 10bp on the buyback news. The 10-year closed at 4.73%. What did move is the price of not holding dollars. Gold is up 7.9% in five days and 16.7% in twenty, at a record, while the dollar index is down 2.4% over the same twenty days. Suppressing a yield does not remove the pressure behind it; it relocates it to whichever price is still allowed to float.

Hidden leverage and shadow banking23 Augmedium

He had become the market, and the filings said so

The single largest buyer of the AI momentum complex was levered four times and publicly visible, and it was liquidated a month before the complex started falling.

A $45bn fund run by a 25-year-old, levered four times, up more than 400% for the year, fell 67% in a single month. Leopold Aschenbrenner's letter to investors in July, quoted in the Wall Street Journal, said: "We let you down this month." The portfolio was the AI momentum trade in undiluted form — chips, memory, energy, neoclouds. According to the WSJ's account, the mechanism of the blow-up was not a wrong thesis but a structural one. Everything the fund bought went up, so it bought more; because it was large enough to be the marginal buyer, its own purchases were producing the returns it was reporting. Its 13F filings made the positions public, so every other desk could see both what it owned and roughly how levered it was. When those names started falling, margin calls arrived, and the fund had to sell the exact securities everyone knew it had to sell. The WSJ says Citadel was on the other side. This is worth holding next to the price data. The names in that basket are the ones down hardest right now: Nebius off 21.6% in five days, CoreWeave 18.6%, Core Scientific 14.2%, Talen 14.4%. Whether that is continued liquidation, ordinary repricing, or the 15% server-price news is not something the tape can tell us. Two cautions. This is an opinion column, not a filing; we have not seen the investor letter, and the fund size and leverage figures are as reported there. And the loss was in July — this is a description of something that already happened, not a new crack. What it demonstrates is the oldest arithmetic in the business. Four times leverage means a 25% move against you is the whole fund. Concentration plus disclosure plus leverage produces a forced seller whose homework everyone else has already marked.

Crypto and TradFi contagion24 Augmedium

A record short squeeze, and the coin holders outran the exchanges

A rally driven by forced short covering, in vehicles that are issuing equity to hold dollars rather than coins, is a weaker foundation than the price implies.

Bitcoin is at $78,887, up 13.9% in five days and 23.2% in twenty. The move did not start with buyers. On 19 August, roughly $2.7–3.0bn of crypto derivatives positions were liquidated in 24 hours, of which about 92% were shorts — around $2.74bn against $254m of longs. Coinglass-based data describes it as the largest one-sided short liquidation on record. Three days later, on 22 August, a downward flash crash liquidated $1.6–1.8bn across roughly 283,000 traders in the other direction. Positioning going in was visible: on-chain trackers had Wintermute holding $146.19m of shorts against $13.85m of longs on Hyperliquid before the squeeze. Underneath that, real money did arrive — about $517m into US spot bitcoin ETFs on 19 August alone, and roughly $2.6bn across bitcoin and ether ETFs for the week, the strongest since October 2025. The interesting part is who benefited. The digital-asset treasury companies ran hardest: Strategy up 25.6% in five days and closing at its period high, BitMine up 28.9%, SharpLink up 27.1%. The intermediaries did not: Coinbase fell 3.9% today while bitcoin rose 1.5%, and Galaxy Digital fell 3.6%. And Strategy itself is behaving like a company managing obligations rather than accumulating coins. It disclosed a new $1.59bn "USD Cash" pool on Monday, sitting alongside a $5.1bn reserve that may only be used for preferred dividends and interest without board authorisation. It sold about $2bn of common stock over the previous week and bought back $136.4m of its Stretch preferred. "The new USD Cash pool gives Strategy more time and optionality, but it does not remove those underlying obligations," Nansen's Nicolai Sondergaard told Bloomberg. Bloomberg notes the valuation premium remains well below earlier-cycle levels — the flywheel is turning, but it is turning to service debt.

The AI capex bubble24 Augmedium

Nvidia's credit protection costs twice what it did in June

Nvidia has moved from selling chips to underwriting its customers' financing, and the credit market has started charging for that even as the equity trades 7% off its high.

The price of insuring Nvidia's debt against default has risen from around 42 basis points in mid-June to 85 on Friday, a record, according to the Telegraph. Protection on Oracle and Broadcom has also hit record levels. Nvidia reports on Wednesday and the sell side expects quarterly sales above $92bn, nearly double a year ago, on a $5.2tn market cap. Eighty-five basis points is not distress. It is roughly what a solid mid-investment-grade borrower pays. What matters is the direction and what the contract references. A credit-default swap pays out when the named company fails to pay its debts. On the face of it Nvidia has almost no debt and enormous cash, so this should be one of the cheapest names in the market. Two things have changed. First, it has actual borrowings: $25bn raised in June, its largest ever bond sale. Second, and more important, it has taken on obligations that are not debt. Up to $125bn of backstops across the $500bn of financing platforms announced with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR. Up to $105bn of credit support for an OpenAI data centre. None of that is borrowing. All of it is a claim on the same balance sheet if the customers falter. There is a second, duller explanation, and both can be true. CDS is one of the few liquid ways to hedge the whole AI complex in one trade. If you own neocloud bonds or data-centre paper and want cover, buying protection on the most liquid name in the chain is cheap and quick. Widening may be hedging demand rather than a default forecast. Either way, the node that was supposed to be the risk-free part of this structure is now correlated to it.

Bond market dysfunction24 Augmedium

Seventeen times more AI debt than this time last year

When a third of the AI buildout is debt-funded, the constraint on the boom stops being earnings and becomes the bond market's willingness to keep showing up.

AI hyperscalers have issued $220bn of debt in 2026 as of 10 August, against $12.5bn in the comparable period last year, on BNP Paribas figures cited by Reuters. Carson Group puts incremental debt at roughly 32% of hyperscaler capex on a trailing basis by mid-2026, up from 9% in fiscal 2024. That shift is the whole story. Capex funded from operating cash flow is self-limiting: you can only spend what you earn, and if earnings disappoint, spending falls the following quarter. Capex funded by borrowing is limited by something else entirely — the appetite of the marginal buyer of the bonds. It does not throttle back gradually. It stops. The pricing says the buyers are already charging for it. QTS Realty sold $3.9bn to fund a Microsoft-linked data centre in Georgia at yields around 7.23%, described by Bloomberg as junk-like on a supposedly high-grade deal, and still drew about $23bn of peak orders. Meta's Hyperion vehicle carries $27bn of debt rated A+, priced at 6.58%, amortising to 2049 — with PIMCO reported to have taken about $18bn of it. Hold that last one up to the light. A single asset manager holding roughly two-thirds of a data centre's debt, on a schedule running 23 years, secured on leases from tenants whose business models are younger than the loan's tenor and whose principal asset depreciates in three or four. We have not seen the documents; the $18bn figure comes from a market note rather than a filing, and we cannot verify how the residual value is treated. The Economist's point is the tail of this: private AI borrowing is now large enough to make capital costlier for everyone, governments included. Ten-year Treasury yields are up 0.6 points this year, bunds 0.4, JGBs 0.8.

The dollar, gold and reserve status24 Aughigh

Citadel names it: the pressure moves, it does not vanish

A calm long end achieved by intervention is not the same as a calm long end, and the difference is showing up in gold and the dollar.

Citadel Securities has told clients that the Treasury's expanded bond buybacks amount to "financial repression at the margin". Nohshad Shah, its head of EMEA fixed-income sales, wrote that "preventing Treasuries from clearing at lower prices does not eliminate that pressure. It merely shifts it elsewhere" — specifically into the currency, and from there into import prices, per Bloomberg. The mechanism is worth being precise about. Buying back long bonds and funding it with bills is a duration swap: the government's total debt is unchanged, only its maturity profile moves. The fiscal arithmetic that pushed 30-year yields to 5.34%, the highest since 2007, is untouched. So if the long yield is held down, the adjustment has to appear in some other price. There is a second channel, and it is the one to watch. CNBC reports the Treasury may fund the purchases from the Treasury General Account, its cash balance at the Fed, which stood at $935bn on 20 August. Spending down the TGA moves money out of the Fed and into private hands. Mechanically, that raises bank reserves. It is an easing of funding conditions delivered by the fiscal authority rather than the monetary one. The prices are behaving as if both channels are live. The dollar index is down 2.3% over twenty sessions. Gold is up 16.3% over the same period and sits at its high. Silver is up 16.8%. Meanwhile the ten-year is back to 4.69% and long-end rate volatility is at a twenty-day low. That is not a bond market in distress. It is a bond market that has been quietened while the stress relocated to the currency, exactly as described.

Crypto and TradFi contagion24 Aughigh

Strategy raises equity to hold cash, not coins

The treasury-company model only works while the premium holds; Strategy is now managing obligations rather than compounding coins, and saying so in its own filings.

Strategy has created a second pool of money on its balance sheet. The new "USD Cash" pool holds $1.59bn and may be used to buy bitcoin, pay preferred dividends, service debt, repurchase securities or meet general corporate needs. The original reserve, now $5.1bn, may only fund preferred dividends and interest without board authorisation, Bloomberg reports. To see why this matters, remember what the digital-asset treasury model is. Issue shares above the value of the coins you hold, buy more coins, bitcoin-per-share rises, and the premium is retrospectively justified. It is a machine that runs on its own premium. Strategy sold about $2bn of common stock in the past week. Nansen's Nicolai Sondergaard notes the raise "strengthened the balance sheet, but did not immediately increase Bitcoin-per-share exposure" — the trade-off is "dilution in exchange for flexibility". In other words, equity is being issued to fund fixed obligations rather than to buy the asset. The company also repurchased $136.4m of its Stretch preferred, which has been trading below par. Buying back your own preferred below par is deleveraging. It is prudent, and it is also an admission: the market is discounting the dividend stream, and management would rather retire it cheaply than defend it. On our reading this is a small reduction in fragility at the company level, not an increase. Cash raised against a live equity bid reduces the chance of a forced coin sale. The shares agree: Strategy is up 25.3% in five days and at its period high, with bitcoin at $78,807, up 13.8% over the same stretch. SharpLink and BitMine are up 25.8% and 29.9%. The flywheel is still impaired. It has just been given more time.

Household credit23 Augmedium

Record subprime delinquency, record subprime lending

Household credit stress at 2008-plus levels is currently being absorbed by pricing rather than by capital, which works until the ABS market disagrees.

Sixty-day delinquency on US subprime auto loans is running at 6.9% on Fitch and Equifax data, against a 2008 peak of 5.0%. A separate Fitch cut puts July at 6.13%, the highest July on record, with prime at 0.49%. The New York Fed's Q2 report shows serious auto delinquency at 3.00%, the highest since 2010; student loan balances 90+ days delinquent rose to 10.6% from 10.3%; 4.7% of all household debt is in some stage of delinquency, against balances of $18.8tn. In the same quarter, auto originations hit a nominal record $211bn. One caution before the mechanism: the summaries we have give two different figures for 90-plus-day auto delinquency, 3.00% and 5.49%, presumably different denominators or transition-versus-stock measures. We cannot reconcile them from what is in front of us, and would not lean on either alone. The shares of the lenders doing this business are not falling. Credit Acceptance, which lends at the deepest end of subprime, is up 7.8% over twenty sessions. Oportun is up 32%. Synchrony is 1.2% off its high. Capital One is up 5.1%. That is not necessarily complacency. Deep subprime lenders price for loss: at APRs in the high twenties, a great deal of default is already in the contract. Equity holders do fine as long as realised losses stay inside the priced band and the securitisation market keeps funding new originations. So the number to watch is not the delinquency rate. It is subprime auto ABS spreads and used-car residual values — the two things that determine whether the losses stay inside the band and whether the funding stays open. Writing record volume into record delinquency means the new book is underwritten to loss assumptions no cycle has tested at this level.

Fed, Treasury and policy24 Augmedium

Paying for the buybacks out of the cash tin

If the government's own liquidity buffer becomes a market-management instrument, the shock absorber and the shock are drawn from the same pot.

Two senior Treasury officials told CNBC the department could draw down its cash pile — the Treasury General Account, $935bn as of 20 August — to fund the expanded buybacks of long-dated debt that Scott Bessent announced last week. Ten-year yields fell as much as four basis points to 4.69% on the report (Bloomberg). Dealers had assumed the buybacks would be funded by selling more bills. That version is a swap: retire long debt, issue short debt, no change in the total. Funding it from the TGA is not a swap. The TGA is cash held at the Federal Reserve. When the Treasury spends it, that money moves out of an account at the Fed and into the banking system as reserves. Buying long bonds with it therefore does two things at once — takes duration out of private hands and adds reserves. That is, mechanically, what quantitative easing does, executed by the fiscal authority rather than the monetary one. That comparison is our inference, not the officials'. What has to stay true for it to be harmless: that the Treasury can rebuild the buffer through ordinary auctions whenever it wants. The buffer exists precisely because that assumption occasionally fails. Since 2015 the department's own policy has been to hold at least five days of expenditure, minimum $150bn, in case it is locked out of debt markets. The officials would not say how much of the $935bn might be used, and nothing has been decided. A cash reserve held against the possibility of a failed auction is being considered as a tool for improving the price at auctions. Those are not the same job.

The AI capex bubble23 Augmedium

Nebius sold $5bn of converts. Then fell 24%.

The marginal AI borrower is financed by instruments whose price depends on its own share price, and that share price has fallen a quarter in a week.

On 20 August Nebius upsized a convertible bond sale to $5bn, which Reuters described as one of the largest on record, to fund data centres and its AI platform. The company spent $5.7bn on capex in the quarter. Since then the shares have fallen 24.1%, including 6.9% today. Nebius now trades 26.5% below its recent high. A convertible is cheap debt bought with equity upside. The coupon is low because the buyer gets an option to convert into shares. Most of that paper is bought by funds that immediately short the stock to isolate the option value, which means a large convert issue tends to put mechanical selling pressure on the underlying. If the equity keeps falling, the conversion option goes worthless and the instrument reverts to what it always was underneath: a fixed obligation with a hard maturity, sitting on the balance sheet of a company burning $5.7bn a quarter. The rest of the debt-financed build-out has moved the same way this week. CoreWeave is down 20.8% in five days, Applied Digital 15.0%, Core Scientific 13.8%, IREN 10.8%, Supermicro 10.2%. The power suppliers have gone with them — Talen down 13.9%, Vistra 7.6%. Nvidia, which is now guaranteeing parts of its own customers' financing, is off 7.0%. The hyperscalers funding capex from operating cash are untouched: Microsoft +1.2% on the week, Alphabet +1.1%. Meanwhile the bond window is wide open. QTS placed $3.9bn against roughly $23bn of orders. Equity investors and credit investors are now looking at the same buildout and pricing two different things.

What would change our mind

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

Whether the Guggenheim loan stabilises above 70 or keeps falling — a slide through the 60s would be the first genuine repricing of a private insurance-and-credit balance sheet, and we would raise ignition several points.

Would move the number

Any non-traded BDC that meets redemptions by selling loans rather than gating: an actual transaction price on a private loan book is the single most informative event available to us.

Would move the number

Whether the SoftBank retail bond prices inside its 4.3–4.9% range on 4 September, and whether other AI borrowers follow it to the retail market — repeated retail funding of hyperscaler duration is new fragility, not just new measurement.

Would move the number

High-yield spreads widening through 320bp, or Talen and Vistra stabilising: the first would say the AI selloff has escaped the sector, the second that it never left it.

Would move the number