The RWA Composability Framework

TL;DR

The headline number for real-world assets on public blockchains is about $27 billion in tokenized AUM. The more useful number is the roughly $2.7 billion that is actively deposited into DeFi lending markets — posted as collateral, supplied into vaults, or used in leveraged yield strategies. That is the share of tokenization that is actually doing something inside the programmable financial system.

The two leaderboards are almost inverted. Treasuries are 48.5% of tokenized AUM but roughly 2% of DeFi deposits. Credit is 17% of AUM but roughly 80% of deposits. The thing that dominates the AUM chart is not the thing that dominates the composability chart. If you want to know where tokenization is actually plugging into DeFi, the second chart is the one to read.

Why? Because composability is a yield-spread phenomenon, not an AUM phenomenon. When collateral earns more than the cost to borrow against it, structured leverage loops are worth building. When it doesn't, the asset sits in a wallet. Credit tokens yielding 6% over a 3% borrow rate create positive carry. Treasuries at 3.5% don't. Permissionless distribution amplifies the gap: assets that anyone can mint and deposit find adoption without partnerships; permissioned products wait.

The divergence

The easiest way to internalize what's happening is to put the two leaderboards next to each other.

Source: Dune, RWAs Onchain — Tokenized vs. Composable, April 2026. Bar widths scale within each column; proportions are directional. Note: the “DeFi deposits” column measures RWA tokens supplied into lending protocols (collateral and vault supply only), excluding borrowed amounts and stablecoin lending liquidity.

The number people usually cite is from the left column. The number that tells you whether tokenization has plugged into DeFi is from the right column. In a year, the right column went from near zero to $2.7B — that's the growth rate that matters.

Why credit wins composability (the yield-spread math)

Credit dominates DeFi deposits because the economics permit structured leverage. Everything else sits downstream of this one fact.

In April 2026, Maple's syrupUSDC yields roughly 6%. Borrowing stablecoins against it costs around 3%. That 3-point spread is enough for professional curators like Gauntlet and Steakhouse to build explicit looping strategies: post the RWA, borrow against it, buy more, repeat. The result is that a single RWA class — credit — shows up across every major lending venue: $957M on Morpho, $929M across Aave's broader markets, $476M on Kamino.

T-Bills yielding 3.5% against a 3% borrow rate create no meaningful spread. That is why Treasuries dominate tokenization and don't dominate composability. The asset is perfectly real and perfectly regulated — it just doesn't have the yield to support a leverage loop. A holder would rather redeem, roll, or hold than lock it into a lending market.

A live case study: USCC → USTB on Aave Horizon

Aave Horizon — the permissioned, institution-focused RWA market on Aave — offers a real-time example of this math in action.

When Horizon launched in August 2025, Superstate's USCC (Crypto Carry Fund) was delivering roughly 15% APY from its crypto-futures basis trade. That yield made it 93% of all RWA collateral on the venue. T-Bill products like USTB were listed but ignored.

By April 2026, USCC's yield had compressed to roughly 4% as basis spreads narrowed across the futures curve. The result:

• USCC collateral share fell from 93% → about 67%.
• USTB jumped from under $1M to $45.6M in 30 days (a ~570% increase) as the yield gap closed.
• Total collateral on Horizon diversified as the highest-yield asset lost its structural edge.

The assets that dominated the first wave of institutional RWA composability (high-yield credit and basis-trade products) may not dominate the next. The collateral mix is an exact function of the yield curve on the underlying products. Watch what happens to spreads; the collateral mix follows.

Source: Dune RWA composability analysis, April 2026; live data at dune.com/entropy_advisors/aave-horizon-rwa.

Permissionless as a distribution channel

Credit wins the yield-spread math, but yield spread alone is not enough. Most tokenized RWAs are permissioned, and permissioned products are slow to integrate. The asset that breaks through is the one designed to be composable on contact.

Maple's syrup tokens are the clearest case. syrupUSDC and syrupUSDT are permissionless ERC-20 tokens — a hybrid between stablecoins and credit RWAs. They're pegged 1:1 to USDC/USDT but earn yield from institutional credit pools. Anyone can mint them, trade them, or deposit them into any lending protocol. No KYC, no whitelist, no bizdev partnership required.

The result is the kind of organic adoption distribution teams dream of:

• 98% of syrupUSDT on Plasma is deployed on Aave.
• 99% of syrupUSDC on Base is deployed on Aave.
• Gauntlet built leveraged vaults around Syrup on Morpho without coordinating with Maple.
• syrupUSDC reached $161M on Kamino on Solana — three chains, one design choice.

Each integration adds utility; utility attracts capital; capital justifies more integration. This is the flywheel that took syrup from a single institutional credit vehicle to about $929M of Aave collateral plus $161M on Kamino, across multiple chains, in months. Permissionless composability is itself a distribution channel.

Four venues, ~$2.7B concentrated

The composable end of the RWA market is unusually concentrated. Five venues across Ethereum, Solana, and several L2s account for nearly all of it.

Two observations. First, Morpho and Aave combined are ~70% of all DeFi composable RWA. Concentration risk runs both ways — a governance or solvency event at either venue is a systemic event for RWA composability, not just a protocol issue. Second, Pendle is not in this list because Pendle is not a lending market — but Pendle is still important here, because Pendle PTs route into Morpho and act as a fixed-rate layer on top of the lending stack. PT-reUSD alone accounts for about $58M of Morpho's RWA deposits. Pendle is to the RWA stack what the swap curve is to a TradFi yield product: the layer where duration and rate get priced.

The emerging category: reinsurance

Most of the composable RWA market is credit, with Treasuries a distant nothing. The category that doesn't fit either bucket is reinsurance — and by deposit rate, it is the most composable asset class of any.

Source: Dune RWA composability analysis, April 2026. Deposit-rate column derived from the AUM and deposit figures; treat as directional at this scale.

Re Protocol's reUSD appears across multiple venues — roughly $96M on Morpho (including $50M in Pendle PT-reUSD) and $94M on Fluid — while OnRe's ONyc accounts for $71M on Kamino. Together, reinsurance represents about 1.2% of tokenized AUM and roughly 10% of DeFi deposits. The deposit rate is an outlier for a reason: reinsurance tokens were designed to be composable from day one, the underlying return profile has low correlation to crypto, and curators have moved quickly to build leverage loops around them.

This is the single most interesting near-term structural story in RWAs. It is small in absolute terms, but it is the first new category in years that launched composable-first instead of being a tokenized version of a TradFi product retrofitted for DeFi.

Three takeaways

1. Growth rates matter more than current size

$2.7B in RWA deposits is a rounding error next to DeFi's total market. But that $2.7B barely existed a year ago. The growth rate is the leading indicator; the absolute size is the lagging one. The deposit figure — not the AUM headline — is the number worth tracking quarter over quarter.

2. What gets tokenized is not what gets used

Treasuries are 48.5% of AUM, 2% of deposits. Credit is 17% of AUM, 80% of deposits. The gap comes from yield-spread math: 6% credit vs 3% borrow rate creates positive carry; 3.5% Treasuries don't. As rate environments shift, the mix will rotate — as we saw live on Aave Horizon with USCC → USTB when basis spreads compressed. Track yield spreads, not AUM.

3. Permissionless access drives distribution

Maple's syrup tokens reached over $1 billion across Aave and Kamino on four chains because they were designed to be composable without a partnership path. Assets that plug in easily find adoption. Assets that require whitelisting are catching up — via deRWA wrappers, LayerZero distribution, and institutional integrations — but more slowly. When evaluating a new RWA product, the most predictive question is not "is it high-yield?" or "is it regulated?" but "can any lending protocol integrate it without asking for permission?"

How TokenIntel uses this framework

This framework lives underneath our coverage of several specific assets and protocols:

• Aave is scored partly on RWA footprint across mainnet + Horizon, with the USCC → USTB rotation as a live example of how institutional collateral actually behaves.
• Morpho is scored partly on its position as the #1 DeFi RWA venue and on the curator ecosystem (Gauntlet, Steakhouse) built on top of it.
• Pendle is scored partly on its emerging role as the yield-curve / fixed-rate layer for RWAs, with PT-reUSD as the leading example.
• Solana's “TradFi / RWA onboarding” pillar in our 5-pillar scorecard is calibrated against Kamino's $587M RWA book and xStocks' live-but-small equity presence.
• Maker / Sky context is shaped by the fact that Treasuries, historically the backbone of USDS-adjacent collateral, are not composability leaders in DeFi — they are stable, not looped.

When we publish a research note or signal that touches tokenized real-world assets, the underlying question we're asking is not “is it tokenized?” but “is the yield spread real, is the design permissionless, and is the capital actually composable?” That is the lens this page is for.