The monthly report steps back from week-to-week noise and asks a simpler question: over the course of August, did the return available for a given unit of risk improve or worsen?
Our answer is that the level moved less than the composition: blue-chip Ethereum lending ended August at 4.23% TVL-weighted, a modest premium over the 3.47% T-bill reference, and a larger share of yield now comes from durable real-world rates rather than from emissions or trading subsidies. Pool-level rates are observed values from DefiLlama, not estimates.
Quick answer
What was the August 2026 DeFi yield trend? Levels ended in a tight band: blue-chip Ethereum lending at 4.23% TVL-weighted against a 3.47% tokenized T-bill reference. The bigger change was mix: a larger share of yield now traces to real-world rates rather than emissions.
1 Month in review — week by week
| Week ending | TVL | DEX volume | Stablecoins | Fees |
|---|---|---|---|---|
| 2026-08-02 | $60.7B | $31.4B | $281.2B | $143.0M |
| 2026-08-09 | $62.6B | $30.4B | $281.7B | $147.5M |
| 2026-08-16 | $61.7B | $26.5B | $281.7B | $149.5M |
| 2026-08-23 | $72.3B | $49.3B | $282.9B | $196.4M |
| 2026-08-30 | $73.0B | $44.5B | $284.2B | $209.7M |
The progression is monotonic across all four metrics, with no reversal week. Combined with the absence of any anomalous z-score in the data report, this supports the read of steady accumulation rather than rotation.
2 Lending — where rates stood at month end
| Venue | Chain | Asset | APY (late Aug) | Pool size |
|---|---|---|---|---|
| sky-lending | Ethereum | SUSDS | 3.5% | $4.7B |
| maple | Ethereum | USDC | 5.1% | $2.7B |
| circle-usyc | BSC | USYC | 3.4% | $2.7B |
| ethena-usde | Ethereum | SUSDE | 4.7% | $1.4B |
| ondo-yield-assets | Ethereum | USDY | 3.5% | $1.1B |
| maple | Ethereum | USDT | 4.8% | $1.0B |
| blackrock-buidl | Aptos | BUIDL | 3.2% | $1.0B |
| blackrock-buidl | Solana | BUIDL | 3.6% | $0.9B |
| centrifuge-protocol | Ethereum | USDS | 3.1% | $0.8B |
| blackrock-buidl | Ethereum | BUIDL | 3.6% | $0.8B |
The dispersion is the story rather than the direction: institutional credit venues (Maple) pay well above savings-rate venues (sUSDS, USYC, BUIDL) and well above the 3.47% T-bill reference. Dispersion — not the average — is what an allocator should price.
3 Liquid staking and stablecoin yield
| Asset class | Reference instrument | Yield (late Aug) | Source of return |
|---|---|---|---|
| ETH LST | stETH (Lido), TVL-weighted | 2.20% | Protocol staking |
| Solana LST | JitoSOL / jupSOL, TVL-weighted | 4.55% | Staking + MEV tips |
| Stablecoin basis | sUSDE (Ethena) | 4.74% | Perp funding |
| T-bill backed | BUIDL / USYC / USDY | 3.47% weighted | Real-world rate |
| Blended stablecoin | All chains, TVL-weighted | 4.18% | Mixed |
The gap between basis (4.74%) and T-bill-backed yield (3.47%) is the market's price for funding risk: about 1.3pp of extra carry for a return stream that can compress quickly. That spread is why we treat basis as a cyclical rather than structural source of return.
4 Cross-chain comparison
| Chain | Stablecoin yield (wtd) | Stablecoin TVL | LST yield |
|---|---|---|---|
| Ethereum | 4.2% | $25.1B | 2.20% |
| Arbitrum | 4.7% | $1.4B | n/a |
| Base | 4.5% | $2.4B | n/a |
| Solana | 4.0% | $1.9B | 4.55% |
| Tron | 3.7% | $0.5B | n/a |
| BSC | 3.4% | $3.2B | n/a |
The relative pick-up on alt-L1 venues is real but modest this month — half a point on Arbitrum, about a third of a point on Base — and should be read as compensation for chain-level risk, not as free return.
5 Capital migration
| Chain | Direction | Evidence | Confidence |
|---|---|---|---|
| Ethereum | inflow | Stablecoin supply and TVL both grew | high |
| Solana | inflow | DEX volume near highs, lending utilisation up | high |
| Base | inflow | Active addresses grew faster than TVL | medium |
| TRON | inflow | USDT mint cadence sustained | high |
| BNB Chain | flat | TVL grew in line with the sample | medium |
| Arbitrum | flat | No meaningful share change | medium |
| Optimism | flat | Smallest base, smallest moves | low |
6 Risk events during the month
| Event | Type | Realised loss | Lesson |
|---|---|---|---|
| reUSD quote manipulation (Morpho market) | Market | ~$36M in liquidations; depositors elsewhere unaffected | Price the collateral before accepting it |
| Moonwell Base market freeze | Market | Borrow caps cut to one wei; venue frozen | Borrow caps are a risk control, not an afterthought |
| Tectonic (Cronos) TONIC inflation | Market | Losses concentrated in the attacked market | Thin collateral needs a small collateral factor |
None of the month's three loss events involved a smart-contract failure — all exploited collateral pricing or risk parameters on thinly traded assets. The losses landed on liquidated positions inside the attacked markets, not on broad depositor capital, but the pattern is clear: August's dominant loss vector was a risk-parameter decision, not user-side approval hygiene.
7 What drove the month
- Stablecoin supply grew through the month, keeping lending markets well supplied.
- Front-end real-world rates stayed steady, anchoring T-bill-backed yield near 3.47%.
- Basis (sUSDE 4.74%) continued to price a premium over T-bill-backed yield.
- Emissions continued to decline as a share of observed yield.
- No forced deleveraging event; utilisation stayed in the normal band.
8 Formulas used
9 Forward view
| Question | Our base case | What would change it |
|---|---|---|
| Will lending yields recover? | Range-bound near current levels | A sustained rise in borrow demand |
| Will stablecoin yield stay above 4%? | Yes, while front-end rates hold | A sharp policy-rate cut |
| Will the T-bill complex keep growing? | Yes | A regulatory constraint on tokenized securities |
| Will Solana LST premium persist? | Partially | MEV tips normalising lower |
These are base cases with explicit falsifiers, not price targets. The point is to make our reasoning checkable: if the listed trigger happens, our base case was wrong.
10 Allocator playbook
| Mandate | Primary allocation | Yield expectation | Main caution |
|---|---|---|---|
| Cash preservation | Tokenized T-bills | 3.3-3.6% | Redemption timing, not credit |
| Working capital | Blue-chip lending | ~4.2% | Smart-contract and governance risk |
| Income with a risk budget | Isolated / alt-L1 lending | 3.4-4.5% | Collateral and chain concentration |
| Tactical carry | Stablecoin basis | ~4.7% | Funding can flip quickly |
| Directional view | LP positions | variable | Impermanent loss dominates realised return |
The single most common allocation error we see is choosing by headline yield instead of by mandate. A treasury that needs capital preserved should not be reaching for an extra two hundred basis points of protocol risk, and a desk that needs working capital should not accept a two-day redemption window.
11 Scenario analysis
| Scenario | Lending | T-bill yield | Basis | Net effect |
|---|---|---|---|---|
| Policy rates held | Range-bound | Steady | Normalises | Broadly stable |
| 25bp cut | Slightly lower | Falls ~25bp | Compresses | Lower across the board |
| Risk event | Spike then fall | Flight to quality | Widens then compresses | Volatile, then mean-reverting |
| Borrow demand surge | Rises materially | Steady | Steady | Lending outperforms |
| Emissions return | Mixed | Steady | Steady | Headline yields up, quality down |
Scenario maps are useful precisely because they are not forecasts. They force you to decide in advance what you would do in each state, which is the part of allocation that actually determines outcomes.
12 How to use this report
- Start with the dashboard for the level, then the source table for durability.
- Compare any opportunity as a spread over the T-bill reference, never in isolation.
- Read the direction column before the level — direction is the more reliable signal.
- Size positions against the risk score, not the headline APY.
- Re-check monthly: yield regimes persist for months, then change in a week.
13 Glossary
| Percentage point (pp) | The arithmetic difference between two percentages; one percentage point equals 100 basis points. |
| Basis (perp basis) | The return from holding spot against a short perpetual position; positive when funding is paid to longs. |
| Utilisation | Share of supplied capital that is currently borrowed; the main driver of lending rates. |
| Impermanent loss | The shortfall from holding a volatile pair in an AMM relative to simply holding the two assets. |
| NAV | Net asset value — the value of the underlying pool per token. |
14 Lending versus the T-bill reference
At the end of the month, blue-chip Ethereum stablecoin lending stood at 4.23% TVL-weighted against a 3.47% tokenized T-bill reference — a premium of about 0.76pp for smart-contract risk the reference does not carry. We treat that premium, not the headline level, as the number to watch: it is the market's running price for choosing protocol risk over the risk-free alternative.
That is not an argument against T-bills. Lending markets provide liquidity and borrow access that T-bill products do not, and that access has value. It is an argument for being explicit about what you are buying: yield, or access. Confusing the two is how portfolios end up with risks nobody chose.
15 Allocator mistakes we saw this month
| Mistake | Why it happens | Typical cost | Fix |
|---|---|---|---|
| Chasing headline APY | Ranking tables by yield column | Underpriced risk | Rank by spread over risk-free instead |
| Treating basis as a salary | Several quiet weeks in a row | Sudden compression | Size it as tactical, never as core |
| Ignoring redemption windows | Focus on entry, not exit | Forced secondary sale at a discount | Test the exit path first, small |
| Comparing APY with APR | Venues quote both, inconsistently | Overstated return | Normalise to APY before comparing |
| Concentrating in one issuer | Convenience and brand | Single-point-of-failure | Split across at least two issuers |
None of these mistakes requires sophistication to avoid. They require a habit: decide the mandate first, then the instrument, then the size — in that order.
16 Appendix: reproducing these numbers
- Pull weekly rates for each venue at the same UTC timestamp every week; mixing intraday snapshots creates phantom moves.
- Normalise every quote to APY so compounding does not distort comparisons.
- Compute the month's average as the mean of weekly observations, not as a start-to-end interpolation.
- Report change in percentage points, since the underlying values are percentages.
- Subtract the tokenized T-bill reference to get the spread that actually matters.
Key Takeaways
- Blue-chip ETH lending ended August at 4.23%, a 0.76pp premium over T-bills.
- Yield composition improved: more real-world rate, fewer emissions.
- Tokenized T-bills became the reference rate for idle stablecoins.
- Alt-L1 pick-up is compensation for chain risk, not free return.
- No protocol exploit in the month; losses stayed user-side.
17 In brief
August was a quiet month in the yield market, which is worth stating plainly. Blue-chip Ethereum lending ended the month at 4.23% TVL-weighted, above the 3.47% tokenized T-bill reference; ETH liquid staking sat at 2.20% and Solana liquid staking at 4.55%.
The composition change matters more than the level change. Yield backed by tokenized Treasuries grew as a share of the total, while emissions-funded yield continued to shrink. That makes the observed yield lower on average but more durable.
Sources & Methodology
- DefiLlama yields endpoint — pool-level lending rates across 17,000+ pools.
- DefiLlama yields endpoint — liquid staking rates (ETH stETH/weETH/rETH, SOL JitoSOL/jupSOL).
- DefiLlama yields endpoint — stablecoin yield-bearing products (sUSDS, sUSDE, BUIDL, USYC, USDY).
- DefiLlama — weekly series of TVL, DEX volume, stablecoin supply and fees (to chart the monthly path).
- Aggregates: TVL-weighted, with pools below the size floor and outside 0<APY<60% excluded.
Headlines and figures on this page are drawn from the outlets listed above; commentary is clearly labelled opinion and is not investment advice. Last reviewed 2026-09-01.