Fully Reserved, Still Fell to $0.87: What Saved USDC Wasn’t Its Reserves—It Was the US Government

If you work in payments and clearing, corporate treasury, asset management, or a family office, sooner or later you’ll have to pick a stablecoin for someone else’s money—or issue one yourself. This piece isn’t about price, and it isn’t about that whitepaper line, “1:1 fully reserved.” It’s about the chain nobody wants to read behind that promise: whose balance sheet your dollars are actually sitting on.

Start with the data. March 2023: USDC—then the straight-A student of full reserves, with ample backing, frequent disclosure, clean compliance—fell to about $0.87 because roughly $3.3B in cash was stuck at one bank, Silicon Valley Bank. Note: on paper, it wasn’t missing a cent.

A Named Framework: Reserve Counterparty Risk

Most people assume holding a stablecoin means bearing market risk—”will it depeg?” Look again: a 1:1 fully reserved stablecoin is, at its core, a risk-transfer contract. It swaps on-chain price volatility for the default risk of a chain of off-chain counterparties.

Draw out that chain:

User → issuer → settlement / custodian banks → money-market funds → short-term Treasury and repo dealers.

Every hop is a counterparty. You think you’re buying a “digital dollar”; you’re actually buying the probability that every single link holds. If any link breaks, 1:1 fails on the spot—and the full reserves on the books won’t save you a cent.

Reserves aren’t one number, either. They’re assets layered by speed: bank demand deposits move T+0 but earn almost nothing; overnight reverse repo is available next day; short-term Treasuries must be sold into the secondary market; money-market fund shares add yet another counterparty layer. Redemption speed and yield trade off against each other. The real exam for a reserve sheet was never the total—it’s this: when a run hits, how much can you actually move at T+0?

The SVB Autopsy: How Three Days Broke Through “Full Reserves”

Mechanism, not headlines. The skeleton timeline: ~$3.3B of reserve cash parked at SVB → SVB suffers its own run and is closed by regulators → over the weekend, the US banking system doesn’t settle → USDC holders can’t confirm whether, or when, that money comes back → secondary-market panic, discount to ~$0.87 → on Sunday, March 12, 2023, the Treasury, the Fed, and the FDIC jointly invoke the systemic risk exception for SVB, giving depositors access to all their funds from Monday (March 13) → the peg returns to near $1.

The root cause of the depeg wasn’t a hole in the reserves. It was that the certainty of redemption vanished over a banking holiday—

Stablecoins trade 7×24. Banks settle 5×8. That mismatch is structural explosive.

Trouble hits on a Friday, and the secondary market gets 48 straight hours to panic while the stuck cash has no answer until banks open Monday. The market won’t wait for your explanation; in every minute of your silence, it imagines the worst-case ending for you, then votes with its feet. $0.87 didn’t measure Circle’s solvency. It measured the market’s panic discount on one question: does that money come back?

The key pivot: what saved USDC wasn’t Circle’s reserve sheet—it was regulators announcing, on March 12, 2023, that they would fully protect SVB deposits, restoring the certainty of redemption. Pin down the nature of the event: that time, USDC’s solvency was never in question—the reserves were genuinely there. What failed was liquidity, plus single-point concentration in one banking counterparty. Had regulators chosen not to backstop SVB deposits that weekend, a fully reserved straight-A stablecoin could genuinely have collapsed because one bank failed. You thought you were trusting Circle; in that moment, you were betting on whether the Fed and the FDIC would step in.

The Law Sets a Floor, Not a Ceiling

The GENIUS Act (Public Law 119-27, signed July 18, 2025) wrote 1:1 reserves and monthly disclosure into law—the industry’s floor got raised. But anyone doing due diligence needs to know where that line ends:

First, most reserve reports you’ll see are attestations, not audits. An attestation is an accountant looking at a single point-in-time snapshot and saying “I checked; it matched that day”—it covers a photograph, not the whole film. An audit covers an entire period, including internal controls. In due diligence, don’t let one sentence—”we’re audited”—blur the two together.

Second, monthly disclosure means your visibility into the reserves runs a month behind. In the 30 days between two reports, nobody outside knows how far one bank’s share of the cash quietly climbed.

Disclosure frequency is the sampling rate of trust. Monthly sampling means you’re forever judging a 7×24 moving risk exposure with a month-old snapshot. An SVB-type accident happens precisely in the gap between two snapshots.

Rate Cycles: This Business Shorts Its Own Future

Circle Q1 2026 (10-Q): reserve income of $652.5M, the overwhelming bulk of $694.1M total revenue. Tether, same period: net income ≈ $1.04B, Treasury exposure ≈ $141B. The engine of this business is interest, not fees. But the P&L hides an asymmetry almost everyone underestimates—

Reserve income is procyclical profit; distribution cost is countercyclical liability.

When rates fall, reserve yield softens that same quarter. Distribution costs, meanwhile, are written into contracts—Circle paid Coinbase alone $330.6M, and it pays that whether rates sit at 5% or 1%. When the cutting cycle arrives, revenue slides while costs don’t move an inch, and the issuer gets squeezed from both ends. At high rates it looks like a money printer; when rates drop, it shows its true form: a high-fixed-cost operating system—and the players who bought their scale with heavy distribution costs are the first to bleed.

Reserve management is boring. But these three boring things—which bank holds the money, who custodies it, where rates are heading—are exactly where a stablecoin’s real cause of death hides.

The peg is written in the whitepaper; the trust is written on your bank’s balance sheet. Full 1:1 reserves are not a get-out-of-death card. They just move your risk off the price chart and onto someone else’s balance sheet—one you can’t see and can’t control. And the bank never answers for your stablecoin.

When you run due diligence, have you ever asked one question of that stablecoin: how much of the total reserves sits at its single largest bank?

Sources (verified):

• Circle Q1 2026 10-Q (reserve income $652.5M, total revenue $694.1M, $330.6M paid to Coinbase): https://www.sec.gov/Archives/edgar/data/1876042/000187604226000150/crcl-20260331.htm
• Tether Q1 2026 attestation (net income ≈ $1.04B, Treasury exposure ≈ $141B): https://tether.io/news/tether-posts-1-04b-q1-2026-profit-despite-highly-volatile-global-markets-reaches-all-time-highs-8-23b-reserve-buffer-and-maintains-u-s-treasury-heavy-backing/
• GENIUS Act (Public Law 119-27, signed 2025-07-18): https://www.congress.gov/bill/119th-congress/senate-bill/1582/text
• Attestation vs. audit framing: https://www.circle.com/transparency · https://tether.to/en/transparency

—— From The Stablecoin Operating Manual: Distribution, Reserves, and Compliance, Chapter 8