Not all dollar stablecoins carry the same risk

Crystal Foresight tracks the $299.4B stablecoin market in real time, covering 99% of all stablecoin assets. This analysis profiles the nine largest dollar-pegged stablecoins using on-chain transfer data, showing how each one is actually used, who holds the supply, and where the risks are – so that compliance teams, institutional participants, and investigators can see what market cap alone does not show.

Key takeaways

• Only one stablecoin is mainly used for direct person-to-person transfers. That is USDT, at 41.9%. Every other $1B+ stablecoin is dominated by exchange trading, lending, or automated protocol activity. In the 30 days ending May 7, Tether froze $515M across 371 addresses – showing both active compliance enforcement and the scale of illicit activity that regulators are watching.

• USDC’s activity is concentrated in one trading platform. Aerodrome Finance, a decentralized exchange on the Base blockchain, generated $1.1T in 30-day transfers involving USDC. That figure is inflated by how automated trading pools work (each trade triggers multiple transfers), but it shows how dependent USDC is on a single piece of infrastructure.

• 86% of USD1 supply sits with one exchange group. Binance and its custody arm Ceffu hold nearly all of USD1. Since the initial large mint, market cap has dropped 17.6% in 90 days. The supply appears to be circulating within Binance rather than spreading to other platforms or users.

• USDe shrank by 37.6% in one quarter. This is not a crisis – it is how the asset works. USDe pays holders a return based on trading market conditions. When those returns dropped, capital left. The concern is that 80% of supply sits in the issuer’s own vaults, which means the next downturn could trigger a faster exit.

• USDG is growing faster than any asset in this set (+134.5% in 90 days)but most of that growth went to one exchange. OKX holds 27% of supply and received $1.2B from the issuer over 90 days. Fast growth concentrated in a single partner is a pattern worth watching.

The snapshot

Every stablecoin transfer leaves a trail. Crystal Foresight classifies those transfers by what they are actually part of – a trade on an exchange, a deposit into a lending platform, a direct person-to-person send, or an automated loop designed to amplify returns. This classification is the volume fingerprint. It shows how an asset is functionally being used, not just how much money moved.

The table below shows each asset’s size, recent growth, the type of entity generating the most activity, and the dominant category of transfer activity.

One important caveat: decentralized exchange volumes appear larger than they are. Automated trading pools generate multiple back-and-forth transfers for every single trade a user makes. A high DEX share signals deep liquidity, not a matching number of individual transactions.

What the data shows, asset by asset

Tether (USDT) – $189.7B, +3.2% over 90 days

USDT is the only stablecoin in this group that is mainly used for direct transfers between people. Its 41.9% peer-to-peer share sets it apart. Nearly half of all USDT transfers happen on Tron, where transaction fees are close to zero – making USDT the practical choice for everyday payments and remittances in regions where traditional banking is expensive or hard to access.

Binance is the largest single source of USDT volume at $114.1B over 30 days. Tether froze $515M across 371 addresses in the same period, with 98% of those freezes on Tron. This shows active compliance enforcement, but it also quantifies the scale of illicit activity flowing through the network.

What to watch: If the peer-to-peer share drops, USDT is losing relevance in the everyday-payment markets where it matters most. The larger risk is regulatory: any enforcement action against Tether’s reserve structure could cause rapid supply contraction in countries where millions of people rely on USDT as their primary way to hold dollars.

USD Coin (USDC) – $77.1B, +5.2% over 90 days

Half of USDC’s transfer activity (49.9%) is decentralized exchange trading. Another 36.8% is lending. The single largest source of USDC volume is Aerodrome Finance, a trading platform on the Base blockchain, which processed $1.1T in 30-day transfers. That number reflects how automated trading pools work (each trade generates multiple transfers), not $1.1T in individual user transactions – but it shows how central one platform is to USDC’s activity.

USDC grew 73% in 2025, outpacing USDT for the second year running. In May 2026, the CLARITY Act preserved stablecoin reward programs in the US, and Circle received MiCA approval from France’s financial regulator for EU-wide operations.

What to watch: USDC’s volume tracks closely with activity on Base and Aerodrome. If either experiences a disruption, USDC’s volume and usability would be directly affected. This is infrastructure-specific concentration risk.

Sky Dollar (USDS) – $11.0B, +14.5% over 90 days & DAI – $4.6B, +5.1% over 90 days

These two assets are phases of the same project. MakerDAO rebranded to Sky in 2024 and introduced USDS as the successor to DAI. The idea: you mint USDS, deposit it into the Sky Savings contract to earn a return (competitive with short-term US Treasury rates), and that return is funded by Sky lending the pooled deposits out through its own lending market.

The data confirms this closed-loop design. USDS has a 71.3% lending fingerprint, and the Sky Savings contract holds roughly 51% of total USDS supply. DAI’s activity is dominated by governance operations (52.2%) – the protocol converting DAI holders to USDS. Nearly all counterparties for both assets are internal to the Sky system: its own lending market, its own governance contracts, its own savings product.

What to watch: The ratio of USDS to DAI market cap tracks how fast the migration is moving. The bigger concern is dependency. A combined $15.6B in supply relies on a single protocol’s health. If Sky’s lending market or savings product encounters a problem, both assets are affected simultaneously.

USDe – $4.0B, -37.6% over 90 days

USDe is not a payment instrument. It is designed to generate returns for holders. Ethena, the issuer, maintains a trading position (buying crypto while simultaneously betting against it through derivatives) that earns income from market conditions. That income is passed to holders who deposit their USDe into staking vaults.

The data makes this explicit. The dominant transfer type is yield strategies (42.6%), the highest of any $1B+ stablecoin. Ethena’s own vaults hold approximately 80% of total supply. The 37.6% market cap decline over 90 days happened because the trading conditions that generate returns weakened. When returns fall, holders withdraw. The mechanism is working as intended – but 80% concentration in the issuer’s own vaults means any future rate decline could produce a faster and larger exit.

What to watch: The share of USDe deposited in staking vaults. A high staking rate means most holders are there for the returns and will leave when returns drop. The risk scenario is a sharp reversal in the trading conditions that fund the yield, triggering a wave of withdrawals.

USD1 – $4.4B, -17.6% over 90 days

USD1 was created by World Liberty Financial and launched through a single large deal: the $2B MGX/Binance investment settled in USD1 in May 2025. The on-chain data shows what happened after that initial event.

Binance and its custody partner Ceffu hold approximately 86% of all USD1 supply. The two largest outflows from the issuer over 90 days both went to Binance addresses ($467M and $45M). Lending activity is minimal. Usage on decentralized exchanges is thin. A small presence on the Solana blockchain is emerging through Manifest, a trading platform, but it remains early-stage.

Market cap has dropped 17.6% since the initial mint. The supply appears to be circulating within the Binance system rather than spreading to new platforms or user groups.

What to watch: Whether supply moves beyond Binance. Right now, 86% of a $4.4B asset depends on one exchange group. That is a concentration risk regardless of the exchange’s size or reputation. Any disruption to Binance’s operations would directly affect USD1’s usability and liquidity.

Global Dollar (USDG) – $3.4B, +134.5% over 90 days

USDG has a different business model from most stablecoins. Paxos, the issuer, shares nearly all of the interest earned on the dollar reserves with partner platforms – OKX, Kraken, Robinhood, Worldpay, and Anchorage Digital. Those partners then offer users a share of that interest through Earn programs, giving every partner a financial reason to encourage users to hold USDG.

The result is the fastest growth in this group: +134.5% over 90 days. But OKX holds 27% of supply and received $1.2B from Paxos over 90 days. Pendle Finance, a platform for trading future yield, holds 5% ($177M) – a sign that USDG is starting to appear in more complex financial products.

What to watch: Whether growth spreads beyond OKX. Fast expansion driven by a single partner carries the same concentration risk seen in USD1. The test is whether Kraken, Robinhood, and newer partners start holding meaningful shares of supply.

Ripple USD (RLUSD) – $1.6B, +2.1% over 90 days

RLUSD operates two different use cases. Direct person-to-person transfers account for 34.0% of activity – the second-highest peer-to-peer share after USDT. At the same time, 21.0% of activity consists of borrow-and-redeposit loops, where users repeatedly borrow and redeposit to amplify returns. That is the highest share of this type of activity for any $1B+ stablecoin.

The distribution chain is visible: Ripple mints RLUSD to itself, then distributes to B2C2 Group ($351M over 90 days), a professional trading firm named as an official market maker at launch. Aave, a lending platform, holds 10% of total supply.

What to watch: The balance between peer-to-peer and lending activity. Growth in peer-to-peer signals real institutional payment volume. If the lending share drops, that demand channel weakens.

PayPal USD (PYUSD) – $3.5B, -11.6% over 90 days

PYUSD has the most evenly spread activity of any asset in this group. No single transfer type accounts for more than 21.7%. The Spark vault (a partnership with MakerDAO/Sky) holds $767M, targeting $1B in deposits.

The most active growth is on the Solana blockchain.. Manifest, a Solana trading platform, is the second-largest source of PYUSD volume at $3.4B over 30 days. A notable holder is USD.AI, a lending service that finances AI data center hardware using PYUSD as its settlement currency – the first project built on PayPal’s new PYUSDx framework, which allows other companies to create specialized tokens backed by PYUSD.

What to watch: Whether the Spark vault reaches its $1B target. The even distribution could mean broad adoption – or an asset that has not yet found a dominant use case. If no channel pulls ahead, the risk is scattered usage without lasting demand.

What it means for risk assessment

Market cap tells you how big a stablecoin is. It does not tell you how it is used, who holds the supply, or what would cause it to shrink. The on-chain transfer data answers those questions.

Three patterns stand out across these nine assets. First, concentration risk: USD1 has 86% of supply with one exchange group, USDe has 80% in issuer-controlled vaults, and USDG’s fastest growth flows primarily to one partner. When most of an asset’s supply sits with a small number of holders, any single decision by those holders can move the market.

Second, infrastructure dependency: USDC’s volume tracks one trading platform on one blockchain. USDS and DAI depend entirely on one protocol. These are not diversified instruments – they are deeply tied to specific pieces of infrastructure.

Third, the gap between appearance and function: USDe looks like a stablecoin but behaves like a yield product. USD1 looks like a $4.4B asset but circulates primarily within one exchange. PYUSD looks evenly adopted but may simply lack a clear identity. The transfer data reveals what each asset actually is, beneath the market cap figure.

The stablecoin market is not consolidating to a single winner. It is splitting into distinct categories – and the on-chain data is the clearest way to see which category each asset belongs to and what risks come with it.