Crystal Foresight said the stablecoin market shrank after setting a record near $320 billion in May 2026, with total supply falling to $306.5 billion by July 14. The report puts the 90-day decline at $11.5 billion, or 3.6%, marking the first quarterly contraction in nearly three years.
The report was written by Hannah Curtis, head of product at Crystal Foresight. ChainCatcher published a full translated version credited to @lufeieth. The central point is straightforward: supply did not just move between wallets or chains. These tokens were redeemed and burned, meaning dollars left on-chain circulation and moved back into bank dollars.
A real contraction, concentrated in a short list of tokens
Crystal Foresight says the market’s pullback was real, but it was not broad-based. A handful of stablecoins accounted for almost all of the decline, and the reasons were different in each case.
The report lists five main decliners over the 90-day period:
- USDC: down $5.8 billion
- USDe: down $2 billion
- USDS: down $2 billion
- USDT: down $1.4 billion
- PYUSD: down $1.2 billion
A smaller group of tokens expanded over the same stretch:
- USDG: up $829 million
- USD1: up $338 million
- DAI: up $251 million
- RLUSD: up $67 million
According to the report, the added supply from the gainers amounted to only about one-fifth of what the declining stablecoins lost. It also notes that two gold-backed tokens fell as well, but treats them separately because their drivers came from the gold market rather than dollar demand.
From a record high to the first quarterly contraction since late 2023
The report says stablecoin supply kept rising through most of 2026 and reached an all-time high of $320.4 billion in mid-May. Over the 90 days ending July 14, supply moved from roughly $318 billion to $306.5 billion, a decline of $11.5 billion. That left the market almost $14 billion below the May peak.
Crystal Foresight describes this as the first quarterly contraction since late 2023. It adds that June alone produced the largest one-month dollar-denominated decline since the Terra collapse in 2022.
The report stresses that redeemed stablecoins are burned. On that basis, the drop points to actual capital leaving the on-chain system rather than a depeg or a bookkeeping shift inside the crypto market.
Its chart note says the data covers the total market cap of all tracked stablecoin assets in 2026, with DefiLlama cited as the data source.
Different stablecoins do different jobs on-chain
One of the report’s key arguments is that a stablecoin should not be treated as just another interchangeable on-chain dollar. The same nominal unit can serve very different functions, and those functions shape how supply responds to market conditions.
Using Crystal’s "transfer fingerprint" analysis, the report breaks out the roles this way:
- USDC and USDS function mainly as collateral stablecoins. It says 58% of USDC transaction volume and 80% of USDS transaction volume is tied to capital moving into and out of lending markets.
- USDe is categorized mainly as a yield stablecoin, with close to 40% of its activity occurring in yield protocols.
- USDT is used primarily for payments and trading. About half of its activity is standard transfers, roughly one-quarter is exchange flow, and its collateral share is relatively low.
That usage split helps explain why the same quarter produced very different outcomes across major tokens. The report’s chart note says the graphic shows the share of each stablecoin’s last-seven-day transfer volume by activity type, based on Crystal Intelligence transfer fingerprint data.
Gold-backed tokens belong to a different market
Crystal Foresight separates PAXG and XAUt from the rest of the analysis. Those tokens track spot gold, not demand for dollar redemptions.
The report says spot gold fell by about one-quarter from its record above $5,590 per ounce in January 2026. It attributes that pullback to a stronger dollar and weaker market expectations for Federal Reserve rate cuts.
On that basis, the roughly $900 million combined decline in PAXG and XAUt is described as a gold-market move, not evidence about stablecoin adoption. The report says the two should be treated separately when looking at dollar-pegged stablecoins.
USDC posted the largest decline, and the holder data masked it
USDC accounted for the biggest portion of the contraction, down $5.8 billion over 90 days.
Crystal Foresight ties that drop to USDC’s position inside DeFi. It says only about one-tenth of USDC transaction volume comes from organic use cases such as payments, standard transfers, and settlement. At the same time, 58% of USDC volume is linked to collateral flows and about one-fifth is tied to DEX liquidity.
That makes USDC working capital for DeFi, in the report’s framing. When DeFi cools, USDC supply tends to shrink with it. The report also says USDC’s total transfer volume fell 46.5% week over week.
At first glance, holder data did not seem to show a large retreat. The top 500 addresses reduced their USDC balances by only $1.5 billion, even though total USDC supply fell by $5.8 billion.
The report says one factor distorted that picture: a new Hyperliquid USDC treasury took in $4.9 billion. That address, deployed by Coinbase, received a record transfer of about $4 billion executed by Circle in June when USDC became Hyperliquid’s native stablecoin.
Crystal Foresight says this was a relocation of USDC rather than fresh demand. Once that shift is excluded, the decline becomes much clearer:
- USDC balances on exchanges fell by $7.1 billion
- USDC in Ethena reserves fell by $2 billion as USDe was unstaked and redeemed
- Balances held by smaller holders fell by another $4.3 billion
The report’s conclusion is that one large treasury consolidation obscured a broader redemption trend.
ChainCatcher’s translation note adds that the body text and the waterfall chart do not fully match in the USDC holder breakdown. The report’s own disclaimer says this sort of on-chain attribution is directional and cannot be fully reconciled to total supply.
USDe, USDS, USDT, and PYUSD each fell for different reasons
USDe: down $2 billion as yield growth stalled
The report says USDe fell by $2 billion, a 34% drop over 90 days. The reason given is stalled yield growth.
Through the spring, perpetual futures funding rates moved from negative territory toward roughly flat. That pushed sUSDe yields down into the mid-single-digit range, and about $1.5 billion left the sUSDe staking product.
Crystal Foresight places this in a broader capital rotation out of crypto-native yield products and into tokenized Treasury products such as BUIDL and USDY.
USDS: down $2 billion after a lower Savings Rate
USDS fell by $2 billion, or 23%, over the same period. The report says Sky cut the Savings Rate from 6.5% to about 3.6%, leading capital to exit staked sUSDS.
It also says USDS and DAI share the same Sky reserve base and should be read together. Their combined supply fell from about $13.2 billion to $11.5 billion, and part of that move simply reflected depositors leaving the yield-wrapped product and returning to ordinary DAI.
USDT: down $1.4 billion, but little changed in aggregate
USDT’s decline was much smaller in percentage terms, down 0.7% or $1.4 billion. Crystal Foresight says the move was shaped more by strategic choice than by yields.
The report says Tether chose to keep USDT outside the MiCA and GENIUS Act frameworks instead of adjusting USDT itself to fit those regimes. Demand constrained by those rules would be directed to separate products such as USAT.
It says the modest outflow mainly came from European trading platforms being required to delist USDT. That is framed as a deliberate supply constraint, not as a contraction driven by operating stress.
PYUSD: down $1.2 billion as incentive-sensitive capital left
PYUSD declined by $1.2 billion, or 31%, according to the report. Crystal Foresight says its funding base in DeFi lending markets was highly sensitive to changes in incentives, which gave capital a clear reason to leave.
The report points to two off-chain developments. First, a proposal from the Office of the Comptroller of the Currency in February created uncertainty around the compliance status of issuer-linked yield programs. Second, PayPal reorganized PYUSD at the end of April into a business unit focused on payments first.
Where growth showed up: subsidies, distribution, and infrastructure
USDG: up $829 million, with growth tied to subsidies
USDG was the fastest-growing stablecoin in the report, up $829 million over 90 days, or 40%. But Crystal Foresight says that expansion was driven mainly by subsidized capital.
It says the Global Dollar Network revenue-sharing model funded a lending program of about 7% on Robinhood’s new chain and also seeded initial liquidity into Solana lending pools. Most of the new USDG supply went into those channels.
The report says the durability of that growth depends on how long the incentives remain in place.
USD1, RLUSD, and DAI expanded through new channels
Other gainers grew through distribution and infrastructure rather than yield subsidies alone, according to the report.
- USD1 launched natively on the Tempo payments network.
- RLUSD made new institutional progress, including a cross-bank tokenized Treasury settlement with JPMorgan and Mastercard on the XRP Ledger.
- DAI acted as the mirror side of USDS inside the Sky system.
What Crystal Foresight says matters for the next quarter
The report says no single explanation can account for all stablecoin moves this quarter, and it treats that as the main takeaway.
It highlights two patterns to watch next.
First, much of the decline can be traced back to lower yields from Ethena and Sky rather than a collapse in confidence. The report says that mechanism can work in reverse: if on-chain yields rise again, or if interest rates fall, some of the same supply could come back at a pace close to the earlier outflow.
Second, USDC has high exposure to the DeFi collateral cycle. In practical terms, the report says USDC is likely to keep moving more with DeFi sentiment than with broad stablecoin-sector sentiment alone.
Crystal Foresight lists three things to monitor:
- whether Ethena’s funding-rate trade can return to positive yield
- whether Sky adjusts and raises its Savings Rate again
- how the OCC clarifies its stance on issuer-linked yield programs, which the report says is especially important for PYUSD
Its final analytical point is that yield fingerprints and collateral fingerprints matter more than the top-line stablecoin supply figure. If the market turns, the earliest signals are likely to appear there first.
FAQ: no depeg, and some of the contraction may reverse
Crystal Foresight included an FAQ section to address several points directly.
Does the decline mean one of these stablecoins depegged?
No, according to the report. It says the stablecoins discussed were destroyed through normal redemption flows, not through depegging. A decline in supply means dollars left the chain and returned to bank dollars. Depeg analysis should focus on price, not supply, and the report says none of the stablecoins covered here depegged.
Why did gold-backed stablecoins and dollar stablecoins fall at the same time?
The report says the reasons were different. PAXG and XAUt follow spot gold prices, not dollar redemption demand, so their decline was tied to a broader correction in the gold market rather than to forces specific to stablecoins.
What is a transfer fingerprint?
Crystal Foresight defines it as a classification of the actual purpose behind each transfer. The categories include:
- collateral flows in lending markets
- DEX liquidity
- yield product deposits
- exchange flows
- ordinary payments
The point of the metric, the report says, is to show what a stablecoin is being used for rather than only how much it is being transferred.
Can the contraction reverse?
The report says part of it can. A large share of the decline came from yield programs being cut rather than from demand disappearing permanently. If Ethena’s funding-rate trade turns positive again, or if Sky raises the Savings Rate, some of the supply could return.
Why should USDS and DAI be analyzed together?
Because they use the same Sky reserve base, the report says. When depositors leave staked sUSDS and move into ordinary DAI, the data shows one stablecoin falling and the other rising at the same time. Looking at either token alone would miss that internal migration.
Sources, notes, and disclaimer
The report says stablecoin supply and transfer fingerprint data came from Crystal Intelligence on-chain data as of July 14, 2026, and were cross-checked against DefiLlama. Differences across token datasets were kept within about 1%.
Its gold-price background came from public market reporting, including the record high in January 2026 and the pullback that followed. Off-chain drivers were drawn from public reports and announcements, including CoinDesk, Cointelegraph, crypto.news, Circle’s record transfer of about $4 billion on June 12, 2026, and disclosures from stablecoin issuers and protocols. Its discussion of USDT, MiCA, the GENIUS Act, and USAT relied on public reporting and information disclosed by Tether.
The disclaimer says the analysis is for informational purposes only and does not constitute financial or investment advice. It also says on-chain supply and transfer fingerprint data are directional, reflecting information available as of July 14, 2026. Entity and function classifications were inferred from balances and transfer activity at the largest holder addresses for each stablecoin. Those classifications are meant to show where supply is concentrated and do not amount to a full reconciliation of total supply. The report also says off-chain drivers are possible reasons, not confirmed causal findings. Data sources listed are Crystal Intelligence, Dune, and DefiLlama.
ChainCatcher’s translation note adds that the original report contains a few small inconsistencies across the text, tables, and charts, and the translation preserved the original numbers in each section:
- USDS is shown down $2 billion in the table, but about $1.9 billion in the chart.
- PYUSD is shown down $1.2 billion in the table, but about $1.3 billion in the chart.
- In the USDC holder breakdown, the body text figures for exchanges, Ethena, and smaller holders do not fully match the waterfall chart.
- Total stablecoin market size is listed as $306.5 billion in the text, while the chart endpoint is marked at about $306.2 billion.
The translation note says the report’s disclaimer already acknowledges that on-chain attribution is directional and cannot be perfectly reconciled to total supply.
ChainCatcher’s final note on the report’s USDC takeaway
At the end of the translated article, ChainCatcher summarized what it sees as the report’s core takeaway for USDC: the token has become a central on-chain dollar liquidity asset, but its current demand structure still depends heavily on DeFi.
In that framing, USDC grows quickly as collateral, margin, and liquidity when DeFi expands, and it contracts quickly when DeFi cools. For Circle, the note says, the key question for earnings stability and valuation ceilings is whether USDC can move beyond being DeFi working capital and become a foundational dollar asset in payments, corporate treasury, real-world asset settlement, and institutional finance.