Frgmnt CEO Aurélien Roussel Proposes Reserve Fixes After USDC's 86-Cent Depeg

Roussel's Three-Part Bank-Risk Program

Aurélien Roussel, co-founder and chief executive of the firm Frgmnt, set out a three-part program for managing banking exposure inside stablecoin systems in an interview made public on Saturday (October 17). The measures are concrete: spread an issuer's cash across several trading banks rather than one; segregate reserves so their size and custody can be independently confirmed; and keep multiple liquidity sources and redemption routes available so a tokenholder is never tied to a single payout channel. Each step comes with a stated purpose. Bank diversification is meant to keep one counterparty's failure from freezing reserve access. Segregation and verification are meant to make the reserve pile checkable rather than trust-based. Redundant redemption paths are meant to keep conversions running while one institution is closed or constrained. Roussel framed the program against a shift he sees in what stablecoins are for: the sector's center of gravity, in his reading, has moved from a trading medium inside decentralized finance toward base infrastructure for reaching the dollar, used for transfers, payroll and corporate cash management. Working at that scale requires, he argued, that users send and receive digital dollars without having to learn how a blockchain operates. A smoother interface, he cautioned, changes none of the underlying structure. The token moves on a blockchain, but the dollar value behind it can sit in bank accounts, and converting tokens back into dollars runs through financial institutions. “Blockchain does not eliminate traditional financial risk,” he said, arguing issuers should surface those dependencies and structure them so they can be managed. The list is his proposed framework; the interview does not claim any issuer currently meets all three conditions.

The 86-Cent Weekend of 2023

The case behind the proposals is the failure of Silicon Valley Bank in 2023, which turned Circle's banking relationship into a market event. Circle disclosed at the time that it held $3.3 billion of USDC reserves at SVB, roughly 8% of its total reserves. Over the weekend that followed, with Circle's direct redemption capacity constrained, USDC broke its dollar peg in open trading and fell as low as 86 cents. An analysis the US Federal Reserve published in December 2025 records what came next: redemption requests surged once the reserve-access problem became public. Roussel reads the episode as proof that market confidence depends on more than the size of the reserve stack. What mattered in 2023, in his account, was whether the issuer could reach its bank, and when and through which route holders could actually redeem. His wider market view follows from the same premise. Stablecoins will interlock with existing finance rather than replace it, so even broad business-to-business settlement in digital dollars still needs banking connections and local payment networks, a dependency Circle's plan to take USDC payments into about 40 emerging markets makes concrete. On competition, he judged that Tether's entrenched USDT distribution network is settled enough that regulatory compliance alone will not dislodge it, and he expects the market to fragment by use instead, split across retail payments, corporate treasury, settlement between institutions and savings. Regulation needs no single global rulebook in his view: regional frameworks such as the EU's MiCA can converge on shared technical foundations while local rules govern disclosure, identity checks and redemption, an approach already visible in how the EBA designated USDC the bloc's first significant e-money token from November 12.

What the Checklist Leaves Open

The load-bearing primary record here is the Federal Reserve's December 2025 analysis, and it makes Roussel's checklist concrete: redemption demand reacted to information about reserve access, not to the reserve total itself. The $3.3 billion exposure and the 86-cent low describe one mechanism from two sides, a partial banking failure becoming a market-wide event once holders doubted the payout channel. For USDC holders today the questions his framework poses are countable: how many banking counterparties an issuer keeps, whether reserves are segregated and independently verifiable, and whether more than one redemption route exists. Circle's disclosures cover the distribution side of scaling; our earlier reporting put 58.7% of Circle's H1 2026 revenue toward USDC distribution and transaction costs, alongside $819 million in H1 partner payouts. What no source records is an issuer publishing bank-by-bank reserve detail, so the gap Roussel identifies remains open.