Stablecoin Reserves Are Everything
A stablecoin is a promise: one token equals one dollar. The issuer can only keep that promise if it actually holds one dollar somewhere. That somewhere is the reserve, and stablecoin reserves are everything. Unlike other tokens where supply grows through vesting schedules, stablecoin supply grows only when new reserves come in and shrinks when reserves leave.
In traditional banking, reserves are a fraction of deposits. A bank might hold ten cents for every dollar customers think they own, counting on the fact that not everyone withdraws at once. The assumption is sound for normal times. The system breaks when everyone withdraws at the same time.
Stablecoins do not have that luxury. They circulate on blockchains where redemption is instant, anonymous, and global. No redemption limits exist. No account freezes, no weekend delays. If one million people wake up and want their dollars back at the same time, the stablecoin issuer has minutes to deliver.
How Reserves Get Tested
A stablecoin's reserve sits in accounts at banks, Treasury bills, short-term corporate debt, or other liquid instruments. The reserve is published quarterly in an attestation. Circle reports 100% USD cash or equivalents backing USDC. Tether's reserve report shows roughly 70% in cash and equivalents; the rest in loans to companies, real estate, and gold.
As long as users believe the reserves exist, the token stays at $1.00. But belief is fragile. It rests on three things: the attestation itself, the issuer's history, and the issuer's access to reserves during a crisis.
The third one breaks easily.
In March 2023, regional banks failed across the U.S. Silicon Valley Bank, a major custodian of startup cash, collapsed. Suddenly, users could not tell which banks held their stablecoins or whether those banks would survive. USDC, despite having a genuine 100% reserve, depegged to $0.88. The reserve was there. Access to it was not clear. Fear caused the trade.
When a Run Starts
A bank run in crypto is different from one in the traditional system only in speed. In a traditional bank, a run might take days or weeks. On-chain, it takes hours. One large holder redeems. The reserve shrinks. News spreads that the reserve is being drawn down. Every other large holder wants out at the same time. The issuer processes redemptions in order: the first get dollars, the tenth-thousandth get... nothing, if the reserve runs dry before reaching them.
To stave off a run, an issuer can raise the price of redemption or limit redemptions per holder per day. But a blockchain-based stablecoin cannot freeze accounts without reputational suicide. Users would move to competitors immediately. The issuer's only real tool is speed: redeem quickly enough that users see the reserve is real, and the panic stops.
An algorithmic stablecoin, one with no reserve at all and relying instead on burning and minting mechanics to hold the peg, has no speed defense. It has no reserve to display. The only defense is the algorithm itself. When the market stops believing in the algorithm, the token loses its peg instantly and never recovers. UST, Ampleforth, BASIS, and dozens of others followed this path and failed.
Reserve Ratio and Fully Diluted Value
Not all reserves are created equal. What matters is the ratio of reserve to supply, and what the reserve is actually worth on the open market.
USDC is simple: 100% reserve in cash and Treasuries. If all USDC holders redeem at once, Circle can pay every single one. The system works even in the worst case. Institutional stablecoins like RLUSD follow the same reserve model, building trust through transparency.
Tether is harder to model. The reserve sits in multiple assets with different liquidity profiles. Cash earns no yield. Treasuries earn yield but illiquid. Loans to companies earn high yield but are illiquid and carry default risk. Tether uses this structure to generate revenue; it keeps the token at $1.00 but takes profit from the reserve. When redemptions spike, the reserve can cover them, but only if the illiquid portions liquidate without loss. In a market crash, liquidating a $100 million loan portfolio instantly might produce $98 million in proceeds, not $100 million. The difference comes out of Tether's reserves, not the users' redemptions. The system works as long as Tether can sustain losses.
What happens if Tether cannot?
The Lesson From Other Assets
Fractional-reserve systems fail when two conditions align: panic and insufficient reserves. Lehman Brothers had roughly $600 billion in assets but $619 billion in liabilities when filed for bankruptcy. The math said it was insolvent by $19 billion. The market discovered this instantaneously.
Stablecoins operate the same way. If a stablecoin issuer has $1 billion in supply but only $900 million in reserve, the token is fractional-reserve by definition. No attestation, no smoothness, no promise can change that math. The only question is when the market finds out.
The preventative is radical transparency. Circle publishes its reserve daily. Tether publishes quarterly, which is both more opaque and more sufficient than the traditional banking quarterly reporting; crypto moves faster. Regulators have begun pushing for clearer reserve rules, recognizing that on-chain transparency is the only real guarantee. If either issuer hid the reserve or inflated it, the market would discover the fraud in hours on the blockchain. The incentive to lie is negative: one leaked contradiction and the stablecoin is dead.
This is why banks can operate with 8-10% reserve ratios and stablecoins cannot. Banks have deposit insurance, central bank backstops, and regulatory intervention. Stablecoins have only the truth.
What to Watch
Stablecoin supply climbed from $100 billion at the end of 2021 to over $200 billion today. That growth reflects adoption; it also reflects choice. Users pick which stablecoin to hold based on reserve strength. USDC gained share even after its brief depeg in 2023 because Circle's reserve report is unambiguous. Tether held its share because Tether's reserve, while less transparent, has never failed, even as competitors built around institutional partnerships.
Neither issuer is truly at risk of a run in ordinary circumstances. Both hold real dollars in their reserves. Both have redemption volumes large enough to handle panic without liquidity stress. The real risk is cascade: if one stablecoin depegs, users fleeing the entire sector to regular assets could trigger a flash liquidity crisis that drains reserves faster than they can be replenished. That risk is system-level, not issuer-level.
The smaller stablecoins, the ones with $10 million to $100 million in supply backed by promise rather than attestation, carry real depeg risk. They have smaller reserves, less liquidity, and weaker brand trust. A rumor that the reserve was moved or that the company is in financial trouble can trigger a run in hours. Then the token is gone.
This is why reserve attestation matters more than any promise. Not because issuer executives are dishonest, but because the system does not trust them. On a blockchain, trust is replaced by cryptographic proof. The closer a stablecoin comes to that ideal (daily attestations, real assets, no off-chain magic), the safer it is from depeg.