Circle's Arc is not about Treasuries. It is about Coinbase

Circle opens Arc, its own Layer-1, to the public on 16 September. Eleven institutions secure the chain alongside it: BlackRock, DTCC, Galaxy, Global Payments, ICE, Mastercard, MoneyGram, SBI, Standard Chartered, Sumitomo and Visa. Not a crypto-native cohort, but settlement and card infrastructure taking a position in a public chain. A validator seat is governance and signalling, not flow. The volume is BUIDL once live and DTCC from H2 2027.

The token is where most readings go wrong. ARC has a 10 billion supply, 25% allocated to Circle. Fees convert into ARC, then split between validators, stakers and a burn. The whitepaper is blunt: ARC "does not represent any equity, debt, dividend right, revenue share... or other claim on Circle."

Three P&Ls

Why it matters for institutional players

Whether Arc reduces exposure to Treasuries is the wrong question. It cannot. Reserve composition is fixed by the GENIUS Act and MiCA: short-dated government paper, segregated. Circulation stands at roughly $74.3 billion.

The better question: does Arc grow the float, and improve who captures it?

Every settlement flow Arc attracts has a cash leg, and that leg is USDC. Gas is stablecoin-denominated, a BUIDL subscription settles in USDC, and so does collateral against DTC assets. Institutions clearing on Arc hold working balances, and balances are float.

Where the float sits decides what it is worth. Under the Coinbase agreement, renewed in August, Coinbase receives 100% of the reserve income on USDC held on its own platform, and roughly half on balances elsewhere. A dollar of USDC on Coinbase earns Circle nothing. At Q2 close it held 27% of circulation, up from 21% a year earlier, against Circle's own platform at 17%.

Case study: $5bn of USDC leaves Coinbase

The line that moves is distribution cost, not revenue. On Coinbase's platform the pass-through is 100%. Elsewhere it is about half. Annualised, on a $5bn slice:

The Big Whale's take

Arc is a distribution strategy in the clothes of a technology launch.

State the identity. Change in reserve income is roughly change in float times yield, plus float times change in yield. Q2 is the worked example. Float rose $15.5 billion year on year, worth $160 million a quarter at the old 4.14% rate. The 66 basis point yield fall took $126 million back out. Net, $34 million, exactly what the accounts show: $634 million to $668 million.

The second term is now turning. Fed funds sit in the 3.5% - 3.75% range, the two-year is back to 4.64%, and futures price in at least 25 basis points of tightening, with near 91% odds at the September meeting and a ~77% chance of at least an additional 25 basis points at the December meeting. The compression that squeezed Circle through 2025 and 2026 is reversing.

Good for reserve income, awkward for everything else. Fifty basis points on $74.3 billion is roughly $371.5 million of extra gross reserve income a year. About $100 million of it, the share on Coinbase's platform, passes straight to Coinbase. A hiking cycle does not relieve the distribution problem. It inflates it, while flattering Circle's headline revenue.

On circulation, the real measure of dominance, it holds 24.4% against Tether's 60%, both diluted from 27% and 68% at Q2 close as newer issuers took share of a market that has grown to roughly $305 billion.

Circle's 2.5 billion ARC, against digital assets of $107 million and equity of $3.5 billion, is a warning not a pillar: unlocks are unannounced, the token inflates at 2-3% initially, and any gain is non-cash.

Three things to watch, none the token price: whether Circle's platform share climbs past 17%, whether BUIDL and DTCC volumes migrate, and whether net reserve margin widens or erodes as rates rise, since Coinbase's cut scales with yield.

The conclusion is narrow, and the turning cycle sharpens it. Arc can move balances off Coinbase's platform, where Circle keeps none of the reserve income, onto its own, where it keeps most of it. Same dollar, same Treasury yield, different recipient. What it cannot change is the reserve itself. Circle remains as leveraged to Fed policy as before, only now the leverage runs in its favour, raising the price of leaving the float where it sits.