Stablecoin supply growth: what Coin Metrics can explain

Coin Metrics published a research note on 2026-10-20 asking why USDC and USDT supply grows in some weeks and not in others. The answer is partial: leverage demand matters most, but no model explained more than 25% of the variance. This article looks at what stablecoin supply measures, how the study was built and why the unexplained part is the most useful lesson.

What the Coin Metrics study found

The note, written by research associate Cooper Duschang for the State of the Network series, starts from a simple pattern. Combined USDC and USDT supply climbed past $255B during 2025, but growth has stalled in 2026. This year neither coin has posted month-over-month supply growth above 10%.

The study tested which outside conditions tend to come before supply changes. Perpetual futures funding was the strongest single factor. It explained about 10% of USDC and about 4% of USDT weekly growth variance. Coin Metrics reads this as traders minting stablecoins to post as collateral when they want leverage.

The two coins also reacted differently to market returns. A rising market in the prior week was followed by faster USDT growth, which explained 5.3% of its variance. For USDC the link ran the other way and explained 4.6%. Coin Metrics said USDT grows when traders pile into leveraged risk, while USDC tends to be minted after markets fall, as money moves out of positions and looks for on-chain yield.

What on-chain analysis of stablecoin supply measures

On-chain analysis means reading the public ledger directly instead of relying on company reports. For a stablecoin, supply is the number of tokens in existence on the chains the issuer supports. It rises when the issuer mints new tokens for a customer who deposits dollars, and it falls when tokens are redeemed and burned.

That makes supply a clean count but a blunt signal. It shows that dollars entered or left the token, not who minted, why, or where the tokens went next.

The method: lagged regression and explained variance

Coin Metrics limited the test to USDC and USDT because each exceeds $25B, is spread widely across networks and exchanges, and pays holders no native yield.

The researchers took four years of weekly data and set six possible drivers against supply growth: perp funding, the supply of rival stablecoins, the change in total crypto market cap, the change in the 2-year Treasury yield, Fed net liquidity and the dollar index. Each driver was lagged by one week, so last week's conditions were compared with this week's growth. Lagged factors fit better than same-week ones, explaining 13% of USDC and 16% of USDT weekly growth variance. The authors read this as the factors leading supply rather than following it.

A later test added a carry spread: the Aave v3 average stablecoin supply APY minus the 3-month treasury bill, using data from 2023 onward. It measures how much more a lender earns on-chain than in government bills. Carry and funding correlate at 0.61, so they partly measure the same thing. Carry added nothing for USDT, but for USDC it was significant and lifted explained variance to 15.9%. That fits where each coin is used: USDT is margin for roughly 42% of futures markets, mostly on centralized exchanges, while USDC covers about 10%, concentrated in Hyperliquid, dYdX and Deribit.

How to read this: when a factor explains 10% of the variance, it accounts for that share of the week-to-week swings in growth. The rest comes from outside the model. It is an association across many weeks, not a rule for any single week.

Limits: why the data can mislead

As Duschang put it: "Still, a large share of variance stays unexplained."

• No model explained more than 25% of the variance, so most supply moves have causes the study does not capture.
• Issuer treasury actions, mint and redemption incentives, on and off-ramps and local regulation are left out, and each can move supply on its own.
• Regimes change: including 2022 brings in the FTX, Celsius and Three Arrows Capital collapses, while newer data reflects institutional inflows.
• Single shocks can dominate. After the March 2023 Silicon Valley Bank failure hit part of its reserves, USDC lost its peg and its supply fell from $56.6B in 2022 to a low of $27.9B in February 2024, a move that no funding or yield factor in the study was built to capture.

The practical lesson: a jump in stablecoin supply is a fact about minting, not a verdict on demand for crypto. Read it next to funding and lending data, and keep the large unexplained share in mind. Leverage magnifies losses, and anyone trading with it can lose money.

What to watch next

Stablecoin growth picked up in November 2024 on institutional demand, and USDC regained its earlier highs in March 2025. The model does not claim to say whether 2026's stall lasts. Readers can track whether supply changes line up with the drivers found: funding rates for both coins, and on-chain lending yields for USDC. If supply moves while those stay flat, the cause likely sits in the unexplained share, such as an issuer action.

Sources

• Coin Metrics, 20 Oct 2026

Drafted by AI from the research and on-chain sources listed, then published automatically once checks confirmed the figures against those sources and found no copied text. Not reviewed by a person. Metrics are explained, never turned into price calls.