Original title:What drove the stablecoin supply down
Original author: Hannah Curtis, Product Lead at Crystal Foresight
Original compilation: @lufeieth
Translate the full text of the report, chart descriptions, data tables, frequently asked questions, source acknowledgments, and disclaimer.
What is driving the decline in stablecoin supply?
The stablecoin market reached a record of nearly $320 billion in May. By July 14, the market size had declined to $306.5 billion, a reduction of $11.5 billion over 90 days—a 3.6% drop and the first quarterly contraction in nearly three years.
This decline is real but highly concentrated. Only a few stablecoins accounted for nearly all of the supply contraction, and each experienced a different reason for its decline. Below, we’ll explain which stablecoins changed and why.
Key Conclusion
There has been a real and concentrated contraction in supply. The net supply of stablecoins decreased by $11.5 billion. These funds have been redeemed, not merely transferred between different wallets or networks. Several stablecoins accounted for nearly the entire decline.
The reasons for the decline in different stablecoins vary. The supply of USDe and USDS decreased due to lower yields. The supply of USDC decreased due to reduced demand for DeFi collateral. USDT remained largely unchanged, with its movements primarily driven by strategic decisions rather than yield factors.
Different stablecoins are interconnected through underlying funding pipelines. The unwinding and redemption of USDe alone reduced USDC in Ethena's reserves by approximately $2 billion.
The decline in gold-backed tokens stems from entirely different reasons. As the spot price of gold corrected, the total market capitalization of PAXG and XAUt decreased by approximately $900 million. This is unrelated to changes in the stablecoin market itself.
Part of the growth is driven by subsidized purchases rather than organic demand. The fastest-growing USDG has expanded through a yield-sharing reward program. Once incentives stop, this portion of the supply may also leave.
I. After a record, there is a genuine decrease in supply
Throughout most of 2026, the supply of stablecoins continued to rise. In mid-May, the total size of stablecoins reached a historic high of $320.4 billion.
Over the 90 days ending July 14, the stablecoin supply decreased by $11.5 billion, or 3.6%, from approximately $318 billion to $306.5 billion, and is now nearly $14 billion below its May peak.
This is the first quarterly supply contraction since the end of 2023. The dollar-denominated decline in June alone was the largest single-month drop since the Terra collapse in 2022.
When stablecoins are redeemed, they are burned; therefore, this decline represents actual funds flowing back from the blockchain to bank dollars, not a depegging of the stablecoin.
The reason this deserves further breakdown is that the decline was extremely uneven. When examining the market by individual cryptocurrencies, it becomes clear that the supply contraction was concentrated on a very short list, and each stablecoin had its own distinct reason for the decline.
Chart note: The stablecoin market size rose to a record high in May, then declined by $11.5 billion over the following 90 days. Data covers the total market capitalization of all tracked stablecoins for 2026.
Data source: DefiLlama.
Two: Different stablecoins serve different functions
Although these stablecoins are all表面上 "one dollar on the chain," their actual uses vary significantly.
Crystal's "transfer fingerprint" analysis can identify these differences:
USDC and USDS are primarily collateralized stablecoins. Each has 58% and 80% of its trading volume respectively tied to capital inflows and outflows in lending markets.
USDe is primarily a yield-bearing stablecoin, with nearly 40% of activity occurring in yield protocols.
USDT is primarily used for payments and trading. About half of it is used for regular transfers, about a quarter for exchange fund flows, and the proportion used as collateral is relatively low.
The scenarios in which a stablecoin primarily exists determine the factors that drive changes in its supply. This also explains why the same quarter can have entirely different effects on different stablecoins.
Chart description: One dollar with identical functionality performs different tasks on-chain. The chart shows the proportion of total transaction volumes over the past 7 days for each stablecoin, categorized by activity type.
Data source: Crystal Intelligence transfer fingerprint data.
III. Which stablecoins have changed?
Five stablecoins accounted for nearly all of the supply contraction:
USDC decreased by $5.8 billion
USDe decreased by $2 billion
USDS decreased by $2 billion
USDT decreased by $1.4 billion
PYUSD decreased by $1.2 billion
The other two gold-backed tokens also declined, and the reasons will be explained separately below.
Meanwhile, a small portion of stablecoins have grown:
USDG increased by $829 million
USD1 increased by $338 million
DAI increased by $251 million
RLUSD increased by $67 million
The increase in supply of these growth-oriented stablecoins amounts to only about one-fifth of the decrease in supply of the contractionary stablecoins.
The table below lists complete data, including two gold-backed tokens that appear in the chart but were not individually mentioned in the text above.
Chart description: Supply decrease is highly concentrated. The chart shows the 90-day supply change for each cryptocurrency, in billions of US dollars. USDS data is from DefiLlama; all other data is from Crystal on-chain data.
Four: Gold-backed tokens belong to a completely different market.
PAXG and XAUt track the spot price of gold, and their fluctuations reflect demand for gold, not demand for the US dollar.
Gold spot prices have dropped by about a quarter compared to the historical record of over $5,590 per ounce in January 2026. The price decline has been driven by a stronger U.S. dollar and reduced market expectations for Federal Reserve rate cuts.
Therefore, the declines in PAXG and XAUt reflect changes in the gold market more than they explain adoption trends for stablecoins. When analyzing USD-pegged stablecoins in this report, both should be considered separately.
Five: USDC: The Decline Beneath the Surface
The scale of supply contraction contributed by USDC is very large.
The transfer fingerprint of USDC explains why it is so vulnerable to DeFi cooling. Only about one-tenth of USDC’s trading volume comes from organic use cases such as payments, regular transfers, and settlements.
Meanwhile:
- 58% of USDC trading volume is related to collateral liquidity.
- Approximately one-fifth is related to DEX liquidity.
USDC serves as working capital in the DeFi ecosystem; therefore, when DeFi activity slows, the supply of USDC contracts accordingly.
The total volume of USDC transfers decreased by 46.5% week-over-week.
From the holder data, the decline appears relatively modest. The USDC holdings of the top 500 addresses decreased by only $1.5 billion, while the total USDC supply dropped by $5.8 billion.
However, this data was significantly distorted by a single factor: Hyperliquid’s new USDC treasury received $4.9 billion in inflows.
This address, deployed by Coinbase, received a record transfer of approximately $4 billion from Circle in June, when USDC became Hyperliquid’s native stablecoin.
This change is part of the USDC location migration and cannot be considered a new requirement.
After this migration is removed, the downward trend of USDC will become very clear:
- The exchange's USDC balance decreased by $7.1 billion.
- USDC in Ethena's reserves decreased by $2 billion following the unwinding and redemption of USDe.
- The balances of smaller holders decreased by a further $4.3 billion.
A large-scale fund aggregation masked a broader redemption trend.
USDC holder change chart
The chart shows the contribution of different holder categories to changes in the USDC supply over 90 days:
Chart note: The decline has been masked but not eliminated. The chart breaks down the distribution of USDC holders over 90 days, showing a $5.8 billion reduction, in billions of dollars.
Data source: Crystal Intelligence.
Six: Brief Analysis of Other Stablecoins
USDe: Decreased by $2 billion, down 34%
The yield growth of USDe has stalled.
Throughout the spring, the perpetual futures funding rate gradually moved from negative toward zero, and the yield on sUSDe declined to the low single digits. Approximately $1.5 billion in funds exited the sUSDe staking product.
This is also part of a broader capital migration, with funds moving from crypto-native yield products to tokenized U.S. Treasury products such as BUIDL and USDY.
USDS: Decreased by $2 billion, down 23%
Sky reduced the Savings Rate from 6.5% to approximately 3.6%, resulting in the unstaking of funds locked in sUSDS.
USDS and DAI share the same Sky reserve infrastructure, so they should be viewed together.
The combined supply of the two stablecoins decreased from approximately $13.2 billion to $11.5 billion. Part of this change reflects depositors exiting yield-encapsulating products and returning to regular DAI.
USDT: Decreased by $1.4 billion, down 0.7%
USDT has experienced minimal overall change, with its fluctuations primarily driven by strategic decisions rather than yield.
Tether has chosen to keep USDT outside the scope of MiCA and the GENIUS Act, rather than modifying USDT itself to comply with these regulatory frameworks. Regulatory-driven demand will be directed toward independent products such as USAT.
This minor capital outflow was primarily due to European trading platforms being required to delist USDT.
Therefore, the supply of USDT is actively limited, which differs from contraction caused by operational difficulties.
PYUSD: Decrease of $1.2 billion, down 31%
PYUSD's funding base in the DeFi lending market is highly sensitive to changes in incentives, and there is ample reason for this capital to exit.
In February, a proposal by the OCC created uncertainty regarding the compliance of issuer-related revenue programs. Meanwhile, at the end of April, PayPal reorganized PYUSD into a business unit prioritizing payments.
USDG: Increased by $829 million, up 40%
USDG is the fastest-growing stablecoin, but its growth primarily relies on financial subsidies.
The Global Dollar Network's revenue-sharing model funded a new ~7% lending program on Robinhood's chain and provided initial liquidity to Solana's lending pools. The majority of the new USDG supply entered these use cases.
Whether this growth can be sustained depends on how long the incentive program can be maintained.
USD1, RLUSD, and DAI are growing through new channels.
Other growth-oriented stablecoins primarily rely on distribution channels and infrastructure expansion.
USD1 is natively live on the Tempo payment network.
RLUSD has achieved new institutional-grade milestones, including a cross-bank tokenized U.S. Treasury settlement on the XRP Ledger with JPMorgan and Mastercard.
DAI forms a mirrored relationship with USDS within the Sky ecosystem.
Seven, Comprehensive Conclusion
No single conclusion can explain all the changes in stablecoins this quarter, and that is precisely the most important finding of this report.
When assessing the next quarter's trend, two patterns are worth noting.
First, the majority of the supply decrease can be traced to Ethena and Sky lowering their yields, rather than a loss of market confidence.
This mechanism can operate in reverse. If on-chain yields rise or interest rates fall, the same amount of stablecoin supply may re-enter at a pace similar to the previous exit.
Second, USDC has high exposure to the DeFi collateral cycle. This means that in the future, USDC will continue to be more influenced by DeFi market sentiment rather than solely by overall stablecoin industry sentiment.
Next, pay attention to:
- Can Ethena's funding rate trade be restored to positive returns?
- Will Sky readjust and increase the Savings Rate?
- How OCC will clarify its regulatory stance on revenue-related programs for issuers is particularly important for PYUSD.
Yield fingerprints and collateral fingerprints are more important than the headline number of total stablecoin supply. The earliest signals of a market turning point will appear in these metrics.
Eight, Frequently Asked Questions
Does this decline indicate that a certain stablecoin has become de-pegged?
No.
The stablecoins discussed in this report were all burned through normal redemption processes, with no depegging events occurring.
A reduction in supply means funds are flowing back from the blockchain to bank dollars. To determine if a stablecoin is depegged, observe its price, not its supply. None of the stablecoins covered in this report have experienced a depeg.
Why do gold-backed stablecoins and USD stablecoins both decline simultaneously?
The reasons for the decline in both are different.
PAXG and XAUt track the spot price of gold, not dollar redemption demand. Their supply scales down with overall corrections in the gold market and have limited relation to factors specific to the stablecoin market.
What is a "transfer fingerprint"?
“Transfer fingerprint” is Crystal Foresight’s classification of the actual purpose of each transfer.
Categories include:
- Collateral liquidity in the lending market
- DEX liquidity
- Deposit for yield products
- Exchange fund flows
- Regular payment
This metric reveals the use cases of a stablecoin, not just its transfer volume.
Can this supply contraction be reversed?
Some contractions may be reversible.
A large portion of the decline stems from reduced yield programs, not from a permanent loss of demand.
If Ethena's funding rate trading turns positive again, or if Sky increases its Savings Rate, some portion of the stablecoin supply may return.
Why are USDS and DAI discussed together?
USDS and DAI use the same Sky reserve infrastructure.
When depositors exit staked sUSDS and convert to regular DAI, the data will show a decrease in one stablecoin and an increase in another simultaneously.
Observing just one stablecoin in isolation overlooks the internal fund transfers between them.
Nine: Sources and References
Stablecoin supply and transfer fingerprint data are sourced from Crystal Intelligence’s on-chain data as of July 14, 2026, and cross-verified with DefiLlama data.
Differences between data for each cryptocurrency are kept within approximately 1%.
Gold price context is based on public market reports, including a record high in January 2026 followed by a price correction.
Off-chain drivers come from publicly reported sources, including:
- CoinDesk
- Cointelegraph
- crypto.news
- Circle's record $4 billion transfer implemented on June 12, 2026
- Announcement from the stablecoin issuer and protocol
The analysis of USDT, MiCA, the GENIUS Act, and USAT is based on public reports and information disclosed by Tether.
Ten, Disclaimer
This analysis is for informational purposes only and does not constitute financial or investment advice.
On-chain supply and transfer fingerprint data are directional indicators reflecting information available as of July 14, 2026.
The classification of entities and functions is derived based on the balances and transfer activities of the largest holder addresses for each stablecoin. These classifications are intended to illustrate where supply is concentrated and do not constitute a complete audit or reconciliation of total supply.
Off-chain drivers represent possible causes, not confirmed causal relationships.
Data sources: Crystal Intelligence, Dune, and DefiLlama.
Note: Data口径 differences in the original report
There are minor discrepancies between the original report's text, tables, and charts; the above translations preserve the original figures for each section respectively:
- USDS shows a decrease of $2 billion in the table and approximately $1.9 billion in the chart.
- PYUSD shows a decrease of $1.2 billion in the table and approximately $1.3 billion in the chart.
- The numbers listed in the body for USDC holders' breakdown—exchanges, Ethena, and small holders—do not exactly match the numbers in the waterfall chart.
- The total market size of stablecoins is $306.5 billion in the main text and approximately $306.2 billion at the end of the chart.
The report also states in the disclaimer that on-chain attribution represents directional estimates and cannot be fully adjusted to total supply.
For USDC, the key takeaway from this report can be summarized as:
The most important insight from this report is that, although USDC has become a core asset for on-chain USD liquidity, current demand remains heavily dependent on DeFi.
As DeFi expands, USDC grows rapidly as collateral, margin, and a liquid asset; when DeFi cools, this working capital also contracts quickly.
For Circle, the key to determining future profitability and valuation ceilings lies in elevating USDC from a DeFi working capital asset to the foundational dollar asset in payments, corporate treasury, RWA settlement, and institutional financial systems.