The third dollar-system scan adds a fourth mechanical Treasury bid that the prior two scans had not surfaced, and confirms a Fed research programme now in its fifth annual iteration that treats dollar-role durability as a permanent institutional question rather than a cyclical worry. Both developments compound the architectural-durability case; neither moves a falsification threshold.
A fourth conduit
The load-bearing new fact is the FEDS Note of 22 June 2026, "Decomposing Hedge Funds' U.S. Treasury Exposures" (https://www.federalreserve.gov/econres/notes/feds-notes/), which documents a doubling of hedge-fund Treasury positions between 2023 and 2025. The mechanism is the basis trade: long cash Treasuries, short Treasury futures, financed through sponsored-repo leverage to arbitrage a small but scalable carry. If the 2023 base was approximately $1 trillion — consistent with public estimates — a doubling places current aggregate hedge-fund Treasury exposure in the $1.5-2.0 trillion range. The dollar figures are inferred; what is verified is the note's existence and its core finding.
This matters because the prior two scans enumerated three non-informed, mechanical bid conduits for US Treasuries: sovereign central banks rotating into gold at roughly 1,000 tonnes per year on a four-year rolling average (per the June 16 WGC Central Bank Gold Reserves Survey); stablecoin issuers mandated by the GENIUS Act to hold high-quality reserves, on a trajectory toward $150 billion per year in front-end Treasury purchases; and US life-insurance private-credit channelling roughly a third of the $8 trillion US life-insurer asset base into investment-grade credit. The hedge-fund basis-trade tier is now a fourth. All four conduits share one structural property: position sizing is driven by spread, regulatory mandate, or leverage availability — not by any informed assessment of US sovereign creditworthiness. The marginal Treasury bid is increasingly divorced from fiscal-credibility views.
The basis-trade tier is also the most fragile of the four, and that fragility is precisely what connects the dollar-system durability claim to the AI-capex and private-credit thesis. The same prime-brokerage and repo desks inside universal-bank FICC franchises that warehouse hedge-fund basis-trade leverage are the desks financing hyperscaler debt. JPMorgan Chase is the largest-scale expression of this shared infrastructure, which is why it is being added to the company slate as the dealer and prime-brokerage picks-and-shovels expression of the fourth conduit. A doubling of hedge-fund Treasury positions is, mechanically, a near-doubling of the repo-financing gross balance flowing through FICC desks — with LTCM and March 2020 as the operative stress-unwind analogues.
Institutional permanence
The FEDS Note of 16 July 2026 announcing the Fed's Fifth Conference on the International Roles of the U.S. Dollar (https://www.federalreserve.gov/econres/notes/feds-notes/) is useful evidence precisely because it is the fifth in the series. The Fed now runs dollar-international-role research as a standing programme with its own conference cycle, and the flagged themes — stablecoins, digital payments, dollar dominance — are the same themes the GENIUS Act operationalises. Combined with FEDS Notes from 30 March, 8 April, 1 May ("Banks in the Age of Stablecoins," https://www.federalreserve.gov/econres/notes/feds-notes/), and 22 June, the 2026 count of dated Fed-research artefacts touching stablecoins or dollar international roles stands at five in roughly five months — a density qualitatively unlike anything in the pre-2025 baseline.
The May 1 note's framing deserves attention in its own right. It positions banks as adapting to stablecoin expansion rather than supervising a contained perimeter — stablecoins are treated as a permanent institutional feature of the credit system. Circle's launch of Circle Mint France on 1 July 2026 (https://www.circle.com/blog) extends this logic into a MiCA-compliant EU jurisdiction, meaning the USDC captive-bid mechanism now operates under two sovereign regulatory frameworks rather than one. MiCA's reserve-composition regime broadly parallels GENIUS on high-quality-reserve mandates, so the EU-registered float still points at USD Treasuries. CRCL's addressable float expands; the Treasury-bid mechanism compounds across jurisdictions.
What the prior scan missed
The July 24 adjacent sovereign-fiscal scan surfaced the WGC Central Bank Gold Reserves Survey (June 16) — 45% of 76 central-bank respondents intending to raise gold reserves over the next 12 months, 74% expecting a lower USD share within five years, and a four-year rolling average of approximately 1,000 tonnes per year in purchases — which the July 8 dollar-system scan failed to pick up. That 74% intent figure is a forward test of the gradual-COFER-decline criterion and belongs in this file from this scan forward. Separately, two Fed FEDS Notes — "China shock 2.0" (29 May 2026) and "Vietnam's Export Boom to the U.S.: The Role of Chinese Firms" (17 July 2026), both at https://www.federalreserve.gov/econres/notes/feds-notes/ — are the first dated Fed-research surfaces directly testing the Pettis surplus-rerouting mechanism: Chinese firms integrating into Vietnamese export supply chains post-tariffs. Both support the mechanism; neither breaks it.
The rest of the company slate is intact. Visa and Mastercard sit in the cross-border corridor rails layer; CME and ICE in FX-clearing infrastructure; BlackRock as the USD-asset incumbent; COIN alongside CRCL in the stablecoin-issuer layer; SGOV as the short-duration real-value expression; GLD as the reserve-manager-intent expression. TLT remains a structural avoid — the long-duration nominal UST position is the purest expression of the wrong-hedge thesis, unchanged across all three scans.
What would change my mind
A single-quarter IMF COFER release showing USD below 55% — a drop of more than two percentage points in one quarter — would partially falsify the gradualism claim. RMB crossing 4% in two consecutive COFER prints would be a more durable signal. The cleanest falsification is a SOFR-IORB spread above 25 basis points sustained for 60 or more days without Fed-led resolution: that is the criterion I set for eurodollar and repo dysfunction, and the hedge-fund basis-trade doubling makes the plumbing more sensitive to that stress, not less. A stablecoin aggregate market cap stagnating below US$350 billion by end of Q4 2026, or any issuer de-peg event exceeding 5% for more than 24 hours, would break the captive-bid conduit. And hedge-fund Treasury exposure reversing more than 20% in a single quarter would simultaneously impair the fourth conduit and the AI-capex transmission channel running through the same desks.
What I'm watching next
The immediate docket: the long-overdue IMF COFER Q1 and Q2 2026 releases — three consecutive scan-envelope 403 errors on the IMF data page remain unresolved; the Treasury implementing rule on GENIUS reserve composition; the USTR-Brazil Section 301 trajectory following the June 1 determination; any FSOC stablecoin systemic-risk report; the Fed's Fifth Conference summary FEDS Note, typically published three to six months after the conference; and a dedicated basis-trade-microstructure follow-up covering FICC franchise exposure at JPMorgan Chase, Goldman Sachs, and Morgan Stanley, sponsored-repo facility usage, and stress-unwind analogue modelling — now the most analytically urgent new rotation slot this theme has opened.