DeFi Yield Curve in 2026: Aave, Morpho, Sky, and Pendle

The DeFi Yield Curve in 2026: How to Compare Aave, Morpho, Sky, and Pendle Rates

Short answer: DeFi does not have one yield curve comparable to the U.S. Treasury curve. It has several overlapping rate surfaces: variable borrowing rates driven by pool utilization, governance-set savings rates, fixed-maturity prices for tokenized yield, and incentive yields paid in volatile tokens. A useful curve must hold the asset, maturity, chain, collateral risk, and rate definition constant. Otherwise, the chart compares unlike claims and turns risk premiums into apparent bargains.

This guide builds a practical framework from Aave, Morpho, Sky, Pendle, and U.S. Treasury documentation available through July 11, 2026. Its examples are illustrative, not live quotes or return forecasts.

What a yield curve actually measures

A conventional yield curve compares the annualized yields of debt claims with different maturities but broadly similar credit quality and currency exposure. The U.S. Treasury publishes a par yield curve using indicative bid-side prices for recently auctioned Treasury securities. Its methodology bootstraps forward rates and applies monotone-convex interpolation.

That definition imposes discipline. A three-month bill and a ten-year note differ mainly in maturity and duration, while both are obligations of the U.S. government denominated in dollars. Their yields can therefore reveal expectations about policy, inflation, growth, and term compensation.

Most DeFi dashboards relax several of those controls at once. They may place a variable USDC lending rate beside a fixed sUSDe principal-token yield, a governance subsidy, an ETH staking return, and a token reward. Those positions differ in issuer, collateral, liquidity, maturity, chain, contract stack, redemption path, and unit of account. Connecting them with a line does not create a coherent curve.

The accurate term is often a yield map or rate surface. A curve becomes defensible only after defining the comparison set.

The four rate families

These rates can influence one another through capital movement. They are not interchangeable claims.

Why “the onchain Fed funds rate” is misleading

Sky governance can change a savings rate to influence demand for its stablecoin and the allocation of capital within its ecosystem. That resembles a policy lever in a narrow functional sense. It does not make Sky a central bank or its savings rate the federal funds rate.

The federal funds market concerns unsecured overnight balances at the Federal Reserve between eligible institutions, and monetary policy operates through a legal and institutional framework. A stablecoin savings rate is a return promised by a protocol arrangement whose capacity depends on collateral, revenues, governance, smart contracts, and access rules.

Three distinctions matter:

Sky’s rate is still useful as an onchain opportunity-cost signal. When a large pool of stablecoin capital can earn a higher savings rate, competing lending markets may need more borrower demand or incentives to retain supply. The connection is arbitrage and portfolio substitution, not monetary equivalence.

A rate decomposition that exposes the risk

For a stablecoin position, the displayed annual yield can be decomposed as:

Displayed yield = base cash rate + credit premium + liquidity premium + protocol premium + chain premium + incentive subsidy - fees

This is an analytical identity, not a protocol formula. Each component asks a different question.

Base cash rate

Use a duration-matched sovereign or secured cash benchmark when it is relevant to the asset. A U.S. Treasury rate is not directly investable by every onchain user, and a stablecoin is not a Treasury security. The benchmark merely shows what return is available before accepting the additional stack.

Credit or collateral premium

Who ultimately owes the money? A borrower against volatile collateral, a stablecoin issuer, a bank, a fund, or a protocol treasury may stand behind the return. Overcollateralization reduces some loss scenarios but does not eliminate gap risk, oracle failure, liquidation congestion, or bad debt.

Liquidity premium

Can the position be exited at the displayed value? A lending pool may show accrued assets while lacking enough immediately available cash for all suppliers to withdraw. A fixed-maturity token may have a known redemption claim but a wide secondary-market spread.

Protocol and chain premium

Contract bugs, admin controls, oracle design, bridges, sequencers, governance, and upgrade processes add risk. A deployment on another chain is not the same exposure merely because the ticker is unchanged.

Incentive subsidy

Reward APR is commonly calculated from current token emissions and price. It can disappear by vote, dilute as deposits grow, or fall when the reward token declines. It should be separated from borrower-paid native yield.

Fees

Vault performance fees, management fees, swap costs, gas, bridging, slippage, and exit spreads reduce the return a user can realize. Morpho documentation, for example, distinguishes native yield, rewards, and fees for vault products.

Aave: a variable funding curve, not a maturity curve

Aave pools supplied assets and permits overcollateralized borrowing. Its documentation says supply rates depend on borrow utilization and governance parameters. Utilization is the share of available supply currently borrowed.

When utilization is low, idle liquidity is abundant and borrower demand is weak. Rates tend to be lower. As utilization rises, borrow rates increase. Above an optimal utilization point, many kinked models increase rates more sharply to encourage repayment and attract supply.

The lender does not receive the full borrower rate. A simplified relationship is:

Supply APR = borrow APR x utilization x (1 - reserve factor)

The exact implementation and compounding conventions must be checked for the reserve and deployment.

Worked Aave-style utilization example

Assume a stablecoin reserve has:

Utilization is 80 / 100 = 80%.

The simplified supplier APR is:

8% x 80% x 90% = 5.76%

Now suppose withdrawals reduce total supplied liquidity to 90 million while borrowing remains 80 million. Utilization rises to 88.9%. The borrow-rate model may move to a steeper portion of its curve, so the displayed supply rate can jump. That higher number is not free income. It is partly compensation for scarce withdrawal liquidity and stronger borrower demand.

This creates an important interpretation rule: a sudden variable-rate spike can be a stress signal rather than an improved long-term opportunity.

Morpho: many markets, not one protocol rate

Morpho markets isolate a loan asset, collateral asset, oracle, loan-to-value parameter, and interest-rate model. Official documentation says each market uses an immutable governance-approved IRM, with the AdaptiveCurveIRM designed to steer utilization toward a 90% target.

When utilization remains above target, the curve adjusts upward over time. When it remains below target, it adjusts downward. The mechanism seeks a market-clearing rate rather than relying only on a fixed kink.

This architecture means there is no single “Morpho USDC rate.” Two USDC markets can have different collateral, oracle, liquidity, utilization, and loss risk. Vaults can allocate across markets, and curators influence exposure within configured limits.

Morpho’s own materials warn that a temporary instantaneous rate in one market may not predict the rate users experience. Reallocation from connected vaults can supply liquidity and reduce the spike. Historical averages, allocation weights, withdrawal liquidity, and the curator’s mandate matter more than one screenshot.

How to compare Morpho vault yield

Break the displayed number into:

A weighted average is only the start. A 7% vault concentrated in one thin collateral market can be less attractive than a 5% vault diversified across deeper markets. The yield difference may be a concentration premium.

Pendle: where a maturity curve becomes possible

Pendle splits a yield-bearing position into Principal Token, or PT, and Yield Token, or YT. Pendle documentation states that PT and YT together represent the underlying position. PT is redeemable for the principal value at maturity, while YT receives the underlying yield until maturity under the instrument’s terms.

Because PT has an expiry and expected redemption value, its discount can imply a fixed annualized yield. This is closer to a zero-coupon instrument than an open-ended lending deposit.

Worked PT yield example

Assume one PT is expected to redeem for 1.00 unit of its underlying asset in 180 days. It trades at 0.97 units today.

Using annual compounding:

Implied annual yield = (1 / PT price)^(365 / days to maturity) - 1

The result is:

(1 / 0.97)^(365 / 180) - 1 = approximately 6.36%

That is not automatically a 6.36% dollar return. It is a yield in units of the underlying redemption asset, assuming the PT redeems as expected and is held to maturity. If that asset depegs, the protocol fails, redemption is delayed, or the user exits through a thin market, realized performance can differ sharply.

Constructing a Pendle maturity curve

To build a meaningful curve:

If three comparable PTs expire in 30, 90, and 180 days, their implied yields form a basic term structure. Mixing a 30-day USDe PT with a 180-day liquid-staking PT does not.

Underlying APY and implied APY are different

Pendle describes underlying APY as a recent average of yield generated by the underlying asset. Implied APY reflects market pricing of PT and YT. The two can diverge.

Suppose the underlying recently earned 8%, while the fixed PT yield implies 6%. A YT buyer may be expressing the view that future underlying yield will exceed the market-implied rate. A PT buyer accepts the fixed discount accretion and gives up the uncertain future yield.

The 2-point gap is not a guaranteed arbitrage. The recent underlying APY can fall, transaction costs can consume the spread, and the instruments retain protocol and asset risk.

This distinction also explains why a yield chart must say whether it displays trailing realized yield, current instantaneous rate, fixed yield to maturity, or forecast yield. “APY” alone is not enough.

Sky savings: a policy rate with a balance sheet

Sky’s documentation explains that governance votes on annualized rate changes and that smart contracts accrue savings and stability fees through protocol rate mechanisms. The savings rate can influence demand for the stablecoin and interact with revenues from collateral and lending arrangements.

For analysis, ask:

A governance-set 7% rate and a borrower-paid 7% lending rate may display the same number but describe different cash-flow engines. The former depends on protocol policy and resources. The latter depends on borrower demand, utilization, collateral, and liquidations.

Treasury rates are a benchmark, not a copy-and-paste floor

The official Treasury par curve is based on Treasury security prices. A DeFi stablecoin position adds at least token, issuer, contract, custody, chain, and redemption layers. It should not trade as though those layers were costless.

Define a simple spread:

DeFi spread = net comparable DeFi yield - duration-matched Treasury yield

Then ask whether the spread compensates for:

The Treasury comparison is most useful for dollar-linked assets. Comparing an ETH-denominated staking yield directly with a dollar Treasury yield mixes asset-price exposure with income.

How to build a defensible DeFi yield map

Step 1: choose the unit of account

Build separate maps for USDC, USDS, USDT, ETH, BTC-linked assets, and other claims. Two stablecoins targeting one dollar can still carry different issuer and redemption risk, so preserve the token identity.

Step 2: identify the cash-flow source

Classify yield as borrower interest, staking rewards, trading fees, protocol revenue, offchain asset income, token emissions, or a mixture. If no payer or economic source can be identified, treat the headline yield as a subsidy until proven otherwise.

Step 3: classify rate behavior

Mark each rate as fixed to maturity, variable by utilization, governance-set, trailing realized, projected, or incentive-based.

Step 4: normalize time and fees

Use one timestamp, day-count convention, and APR/APY method. Subtract explicit fees and estimate entry and exit costs. Do not add APR and APY mechanically.

Step 5: map the dependency stack

Record the stablecoin issuer, collateral, oracle, protocol contracts, vault curator, chain, bridge, custody arrangement, and redemption process.

Step 6: record executable liquidity

Displayed total value locked does not prove exit capacity. Capture the amount withdrawable without waiting, the PT market depth within a defined price impact, and any redemption delay.

Step 7: retain history

An instantaneous snapshot is weak evidence. Store daily or hourly observations, methodology changes, governance actions, reward campaigns, utilization, and market depth. A 30-day realized range tells more than a promotional peak rate.

Reading the shape of the curve

Upward-sloping fixed curve

Longer maturities yield more than shorter maturities. Possible explanations include duration uncertainty, expected higher future underlying rates, thinner long-dated liquidity, or greater contract exposure. It does not necessarily signal economic growth.

Inverted fixed curve

Short maturities yield more than long maturities. This can reflect expected rate cuts, near-term liquidity demand, an expiring incentive, concern about short-term redemption, or simply poor pricing in a thin pool.

Humped curve

One middle maturity yields unusually more. Check market depth, instrument terms, points, caps, and expiry-specific events before inferring a macro view.

Variable-rate spike

In Aave or Morpho, a spike often reflects high utilization or a liquidity withdrawal. Determine whether borrowers are paying the rate, whether new supply can enter, and whether the position is withdrawable.

Stablecoin-specific divergence

If one dollar-linked asset yields materially more than another in otherwise similar markets, the gap may price borrow demand, incentive spending, depeg concern, redemption friction, or collateral preference. “Higher demand” is only one hypothesis.

A risk-adjusted yield scorecard

Score each factor from 0 to 2, where 0 is weak or unknown, 1 is mixed, and 2 is strong. The score does not predict loss; it forces comparable questions.

16-20: comparatively strong structure, still not risk-free. 10-15: material risks require position limits. 0-9: the headline rate is not decision-ready.

Do not average away a fatal dependency. A position with excellent yield history but no credible redemption path should fail review regardless of total score.

Worked comparison: three 6% yields

Imagine three positions each displaying 6%:

Position A: variable stablecoin lending

The rate comes from borrowers, utilization is 75%, and withdrawal liquidity is deep. The rate can fall tomorrow, but the cash-flow source is observable.

Position B: fixed PT held to maturity

The implied yield is 6% in units of a stablecoin wrapper. The rate is fixed if held, but the holder accepts underlying protocol, stablecoin, maturity, and redemption risk.

Position C: vault showing 3% native yield plus 3% rewards

Half the return comes from a volatile reward token. Deposit growth or token-price decline can reduce realized reward APR, and fees may apply.

The identical headline number masks different exposures. A rate-sensitive user may prefer fixed PT; a liquidity-sensitive user may prefer the variable pool; a reward-token skeptic may discount Position C heavily. There is no universal ranking without the user’s liability horizon and risk limits.

Common analytical mistakes

Treating APY as guaranteed

Variable APY changes. Reward APY depends on token price and emissions. Fixed PT yield depends on holding and redemption assumptions.

Ignoring the denominator

A 5% ETH yield produces more ETH, not a stable dollar return. ETH price can dominate the income.

Comparing snapshots from different times

Utilization moves quickly. Use synchronized observations or label timestamps.

Counting incentives as organic demand

Separate borrower-paid or asset-generated yield from emissions. Subsidies can be strategically useful, but they are not evidence of durable credit demand.

Using TVL as liquidity

TVL includes deployed or locked assets. Measure immediate withdrawal capacity and executable market depth.

Ignoring recursive leverage

Looping a supplied asset into repeated borrow-and-resupply positions can inflate gross TVL and yield while increasing liquidation sensitivity. Trace the final collateral and debt.

Assuming transparency equals safety

Onchain positions may be visible, but contract behavior, legal claims, private keys, offchain reserves, and governance decisions still require interpretation.

Monitoring checklist

Before entering:

While holding:

FAQ

Does DeFi have a risk-free rate?

No universally accepted onchain asset is free of issuer, protocol, chain, custody, or redemption risk. Treasury rates can serve as external dollar benchmarks, while stablecoin savings rates can serve as ecosystem opportunity-cost signals. Neither makes a DeFi position risk-free.

Is the Sky Savings Rate the DeFi federal funds rate?

No. It is a governance-controlled rate within a stablecoin system. It can influence competing rates through capital movement, but the issuer, legal claim, funding source, and monetary role differ from federal funds.

Why does high utilization raise lending rates?

High utilization means more supplied liquidity is borrowed and less is available for withdrawal. Rate models raise borrowing costs to encourage repayment and attract suppliers.

Is a fixed Pendle PT yield guaranteed?

The discount accretion can be fixed in terms of the underlying redemption asset when held to maturity, but contract, underlying, depeg, redemption, and operational risks remain. Selling early introduces market-price and liquidity risk.

Why can a Morpho market rate change so quickly?

Each market has its own utilization and interest-rate model. Transactions and vault reallocations can move utilization, so an instantaneous rate may be temporary. Historical averages and allocation flows provide better context.

What does an inverted DeFi curve predict?

Not one thing. It can price expected rate changes, near-term liquidity stress, incentives, redemption concern, or market illiquidity. Diagnose the cash-flow and market mechanics before applying a macro interpretation.

Should DeFi yield always exceed Treasury yield?

There is no mechanical rule, but a lower-quality or less-liquid claim normally needs adequate compensation. A temporarily lower DeFi rate can occur because of incentives elsewhere, excess supply, user utility, or market segmentation.

Source note and limits

Protocol mechanics in this guide come from official Aave, Morpho, Pendle, and Sky documentation. The Treasury benchmark methodology comes from the U.S. Department of the Treasury. Protocol parameters and rates can change, so this page deliberately avoids presenting a July 2026 snapshot as a durable forecast.

The utilization, PT, and three-position examples are CryptosEyes calculations. They simplify taxes, gas, slippage, compounding, bad debt, and individual eligibility. They are tools for comparing mechanisms, not recommendations.

What to read next

Read the DeFi yield-farming risk guide for a strategy-level checklist, then use the stablecoin reserve checklist to inspect the asset underneath a quoted rate. For a more complex stack, the liquid-restaking economics analysis traces how additional yield also adds slashing, liquidity, and dependency risk.

The practical takeaway: compare cash-flow engines and exit paths before comparing percentages. A yield curve is useful only when the claims along it are genuinely comparable.

About the Editorial Team

CryptosEyes publishes independent, source-grounded market-structure research. Protocol documentation is treated as a description of mechanics, not proof that a return is safe or sustainable.