The dollar system is not collapsing; its plumbing is migrating — and in the most important development this scan surfaces, the US is now deploying trade-policy enforcement to slow even that migration. The replacement bid for the marginal foreign official-sector Treasury flow is not being seized by foreign reserve managers: it is being mandated by US statute.
The Atlantic Council's Dollar Dominance Monitor (February 2026 update, https://www.atlanticcouncil.org/programs/geoeconomics-center/dollar-dominance-monitor/) puts USD reserve share at roughly 58%. The IMF's COFER series confirms the structural drift — from ~73% in 2001 to ~56-57% by 2024-25 — absorbed predominantly by gold and a basket of smaller currencies rather than any credible rival; renminbi share sits at ~2.5% after a decade of explicit Chinese internationalisation effort (https://www.imf.org/en/blogs/articles/2025/10/01/dollars-share-of-reserves-held-steady-in-second-quarter-when-adjusted-for-fx-moves). These numbers describe a system in controlled secular drift — exactly what C-03 and C-18 predict.
The Pix episode rewires the political model
The most consequential single datum from this scan is not a reserve-share figure. It is the June 12, 2026 USTR examination of Brazil's Pix instant-payment system (https://www.atlanticcouncil.org/blogs/econographics/as-the-us-targets-brazils-payment-system-europe-should-pay-close-attention/). Almost all dedollarisation commentary treats the US as passive — the dollar will be displaced if foreign reserve managers choose to allocate elsewhere. Pix demolishes that assumption. Pix is a domestic real-time payments rail in a politically adjacent emerging market, not an adversarial network. The US response is trade-policy escalation, not architectural concession and not regulatory cooperation; FedNow remains domestic-only at the Fed side. Successful non-USD payment rails now carry trade-policy consequences — this is active defence of the dollar system, not passive incumbency, and it is the first dated 2026 instance of this posture.
The volume asymmetry that anchors the structural argument
The alternative payment architecture is real and growing; it is not at the scale required to displace dollar-system plumbing within the manuscript's five-to-ten-year window. CIPS handled US$26T of annual business volume in 2024, with 193 direct participants and 1,573 indirect by December 2025 (https://www.atlanticcouncil.org/programs/geoeconomics-center/dollar-dominance-monitor/). That sounds substantial until set against BIS-tracked daily FX trading volume of roughly $7.5 trillion per day, of which ~88% has the dollar on one side. On a daily-equivalent basis, CIPS processes ~$100 billion per day — approximately 1.4% of the dollar FX leg's daily turnover. The mBridge figure is more revealing still: cumulative transaction volume of US$55.49 billion by November 2025 is less than half of one percent of the ~$13 trillion stock of dollar-denominated external debt outside the US requiring ongoing servicing. The Atlantic Council's May 2026 piece (https://www.atlanticcouncil.org/blogs/econographics/the-global-push-for-local-currency-cross-border-payments-is-intensifying/) confirms that more than one hundred countries now operate domestic instant-payment systems, with bilateral cross-border connections multiplying — but at the transactional layer where the dollar's role is most contestable, not at the store-of-value and financial-architecture layer where its reserve role lives. Conflating the two is the error the rotation thesis routinely makes.
The captive bid: stablecoin as mandated Treasury demand
The financial-architecture layer is not merely holding; it is deepening via a mechanism most reserve-currency commentary misses entirely. Circle's transparency dashboard, live as of June 15, 2026 (https://www.circle.com/transparency), confirms the reserve composition mandated under the GENIUS Act — the Circle Reserve Fund holding cash, sub-three-month Treasuries, and overnight repurchase agreements. This is the statutory conversion of stablecoin float into a front-end Treasury bid that scales with stablecoin issuance regardless of fundamental view on US fiscal credibility. Tether's H2 2025 reserves already stood at $141.6 billion in US Treasuries; combined USDT and USDC Treasury exposure is now plausibly in the $200-250 billion range (inferring; explicit Q1 2026 attestation not retrieved within scan budget). The replacement for marginal foreign official-sector flow is not RMB-denominated alternatives — it is a regulatory-mandated demand function with no internal price-discovery brake.
Where capital sits
The value-chain implication runs in two directions. On the accumulate side, the picks-and-shovels layer offers the most durable economics. Visa and Mastercard operate the dominant USD cross-border payment rails, reinforced rather than threatened by the current trajectory. ICE operates US Treasury futures clearing and the US Dollar Index futures complex; CME operates SOFR, Treasury futures, and gold futures — the structural plumbing that prices the system. BlackRock sits at the downstream end as the USD-asset-scale incumbent: it operates IBIT, BUIDL, and the passive-flow machinery that benefits from continued architectural lock-in. Circle Internet Group (CRCL), as the publicly listed stablecoin pure-play, is operationally a regulatory-captive Treasury bid that scales with GENIUS Act implementation; Coinbase (COIN) sits alongside it as USDC distribution infrastructure. Short-duration USD cash equivalents — SGOV, BIL — remain the capital-preservation expression of the C-03 accumulate posture: front-end yield without the duration risk.
On the avoid side, the argument is unambiguous. Long-duration nominal US Treasuries — TLT, EDV — are the direct expression of C-24: duration that carries the fiscal-credibility risk the stablecoin bid substitutes for at the front end. BIS data on EM external sovereign issuance (https://www.bis.org/statistics/secstats.htm) confirms that roughly 70% of EM external sovereign bond issuance remains USD-denominated — the investor who buys the dedollarisation thesis without discriminating by institutional quality recycles dollar-system exposure into less liquid wrappers. EWZ warrants watching rather than avoiding outright: the USTR-Pix examination introduces a capital-flow headwind not previously priced. Gold — GLD, and royalty plays FNV, NEM, AEM — is not a dedollarisation winner but a credibility-arbitrage instrument, the asset central banks accumulate as the dollar system drifts without conceding to a rival.
What would change my mind
A single-quarter IMF COFER release showing USD share below 55% — a drop exceeding two percentage points in one quarter — would partially falsify the gradualism premise at the core of C-03. RMB reserve share crossing 4% in two consecutive COFER releases would push the C-18 falsification criterion materially closer. If combined USDT and USDC Treasury holdings stagnate below $250 billion, the GENIUS captive-bid mechanism I rely on as the structural replacement for foreign official-sector flow weakens materially. And if the USTR-Pix examination is withdrawn or formally settled without escalation, the active-defence read I place on the June 12 filing requires revision toward passive incumbency.
What I'm watching next
Four operational dockets carry the argument forward: the IMF COFER Q2 2026 release, which is the first test of whether the ~58% figure holds; the Q3 2026 Treasury implementing rule on the GENIUS Act stablecoin reserve-composition mandate, which sets the statutory floor on the captive Treasury bid; the formal USTR docket number and any expansion of the Pix examination to UPI or PromptPay; and Circle's next quarterly reserve attestation, which should provide the first explicit Q1 2026 data on the combined USDT+USDC Treasury exposure figure I am currently carrying as an inference.