In an Era of Declining Returns, Where Does Sky's Outperformance Come From?

Null

Written by: Vadym

Translated by: Luffy, Foresight News

Today, the widespread decline in on-chain yields is an undisputed fact, and related discussions have long been pervasive. Lending rates are converging with Federal Reserve interest rates, with “safe” deposit rates averaging only about 3%, lower than U.S. Treasuries, SOFR, and other products. Notably, at the time of writing, the yields on USDC and USDT on Aave are around 2%, Steakhouse and Gauntlet Prime offer approximately 3.2% returns (more competitive on Base chain), and sUSDe yields are at 3.5%.

Among over $20 billion in vault assets on Ethereum and Layer 2 networks, 58% of stablecoin TVL has an annualized yield below 3%, and 35% are between 3% and 5%. This makes Sky’s USDS savings product, offering 3.75%-4%, a highly attractive safe haven. But the question remains: with such large TVL, how can Sky offer interest rates above the market average?

Data source: Galaxy Research

Of course, directly comparing Sky’s savings rates with Aave or Morpho’s lending rates isn’t entirely accurate. The latter’s yields are driven entirely by borrowing demand and capital utilization, whereas Sky’s yield mechanism differs:

• Set by governance (stability fee / SSR);
• Accepts any collateral approved by Sky for lending, including off-chain assets via financing (though this mode has its risks, with lower flexibility and higher costs within its ecosystem).

Nevertheless, on the supply side, all projects are in competition. Sky acts similarly to a central bank, providing unsecured credit lines to various entities and participating in most yield sources. Therefore, exploring the essence of its yield sources, the proportion of on-chain versus real-world assets, and the correlation and scalability of various yields is highly valuable.

As Sky co-founder Rune states: “Currently, stablecoins worth over $300 billion generate no yield”… but how much of this enormous potential demand can be absorbed?

On-Chain Yield Panorama

Before delving into Sky’s balance sheet and financials, let’s briefly review how on-chain yields are generated and the overall yield scale expected by 2025.

Simplified On-Chain Yield Sources

Lending Interest

Like traditional finance, the money market is core to DeFi, accounting for over 60% of total DeFi TVL. By 2025, total income from lending (including collateralized, uncollateralized, and low-collateral loans) is estimated at about $1.76 billion. What motivates users to borrow? There are three main lending models: Aave v3’s unified liquidity pools, Morpho’s risk-managed isolated pools, and Sky’s central bank-like model within its ecosystem. Let’s analyze each.

Top 15 borrowers on Ethereum Aave v3

Top 15 borrowers on Morpho

The core scenarios driving borrowing demand are mainly three:

1. Circular Lending. Borrowing related assets and cycling to maximize interest-bearing assets’ returns. On Ethereum Aave v3, about 39% of borrowing is used to amplify ETH staking rewards, with Fluid, Etherfi, and Lido as main players, sometimes exceeding 45%. sUSDe circular borrowing accounts for about 11.6%. On Morpho, at least 27% of borrowing is directly used for protocol-internal USD cycling strategies to earn yields on assets like sUSDS, syrupUSDC, sUSDe; about 5% for ETH-related strategies.

2. Spread Trading and Leverage. About 45% of Aave borrowing demand falls into this category: users collateralize non-stable assets like BTC, ETH, borrow stablecoins, then deploy funds elsewhere to earn spreads, improving capital efficiency. For example, the largest borrower on Morpho (6.5%) borrows USDC and deposits into Sky savings. Another scenario is spot leverage long positions (top borrowers on Aave like 0x54d25, xed0c6, 0x28a55, 0x741aa use this), or borrowing stablecoins while maintaining collateral exposure for other uses. Currently, about 40% of interest income on Morpho comes from borrowing against cbBTC (excluding top borrowers), much of which is from Coinbase users.

3. Other Secondary Demands. Including borrowing non-stable assets against stablecoins for shorting, or investment strategies based on other trading pairs.

It’s clear that roughly half of borrowing is essentially for leveraging to amplify other yield sources. This raises two core questions: where do these interest payments come from? After existing strategies exhaust, what alternatives remain?

On-Chain Staking Rewards

Native asset staking rewards mainly come from two sources: network issuance rewards and MEV profits (priority fees + bribes).

Data sources: Dune, Helius, Blockworks

Ethereum staking rewards mainly derive from network inflation, with a maximum daily issuance of about 2,700 ETH, totaling around 1 million ETH in 2025; Solana’s annual issuance is about 24-25 million SOL. These yields are stable but subject to significant principal price volatility.

Staking rewards include 5%-20% (Solana 5%-30%) from on-chain activities (priority fees and MEV), which has been declining since Ethereum’s merge. The reasons: MEV profits are roughly split between arbitrage and front-running; the latter has decreased significantly due to OFA, solvers, and anti-MEV tools (currently about 90% of transactions are private). The arbitrage portion was historically monopolized by non-neutral searchers and block builders, but now heavily depends on market conditions and competition.

Funding Rates

Ethena pioneered the on-chain perpetual contract funding fee yield model, generating about $240 million in fees in 2025 (90% from funding fees), with total revenue around $300 million. Its unique value lies in:

• Introducing new, composable yield sources into DeFi via tokenized mechanisms;
• Easy to capture;
• Persistent, even with market volatility (2021 yield 16%, 2022 0.6%, 2023 about 9%, 2024 about 13%).

This makes it an ideal underlying for fixed-rate and interest rate swap derivatives, with related products already on Pendle.

While scalability remains to be proven, the open interest of BTC perpetual contracts is between $35-65 billion, ETH perpetual contracts around $20-40 billion in 2025. Early 2026, total open interest hits about $75 billion, with Ethena accounting for roughly 1.8%.

Trading Fees

Token swaps are among the most fundamental activities on blockchain, providing infrastructure and earning fees—a stable profit path. In 2025, AMM liquidity providers will earn about $4.2 billion in trading fees, with 62% from Uniswap, Meteora, and Raydium.

However, capturing these yields via structured products is challenging:

• LPs often face losses: pools are vulnerable to malicious order flows, especially concentrated LPs, usually only accessible to professional traders; Uniswap LP management products are not widely adopted.
• Limited composability: LP positions as collateral are rarely used.
• Yield concentration: much trading fee volume comes from niche assets, hard to fit into typical yield products. In 2025, about 25% of Ethereum’s trading volume is ETH-stablecoin pairs (currently about 60%), with 41% from USD pairs; on Solana, 50% of volume is SOL-USD, 30% Meme coins, and only 5% stablecoins (current ratio roughly 62:12:17).

The AMM vs. PMM landscape: scalability of AMMs remains debated. Although most front-end volume is routed via solver networks, only about 11% is via professional market makers (PMMs), the rest still AMMs. In 2025, Solana’s dedicated AMM share is 30%, up from 60%, dominating SOL-USD trading with peak market share at 86%.

Nevertheless, some trading strategies are managed via vaults on Gauntlet, with Sky holding some AMM LP positions.

As DeFi perpetual contracts grow, their market-making vaults (HLP, LLP, etc.) can be seen as alternatives to Uniswap LP management. The yield scale is moderate (~$130 million), with JLP contributing an additional ~$670 million. Risk managers are closely watching this space; risk features include: “HLP, LLP, Giga vaults, OLP etc. have max drawdowns of 5%-9%, with some months of floating losses; daily returns are mostly small profits, some breakeven, occasional large losses; risk-adjusted returns are good, but distribution is uneven and path-dependent.”

Risk Transfer Yields

Handling risk involves three methods: retention, mitigation, and transfer. Most participants currently retain risk, using advanced mitigation tools. Transferring volatility or protocol risks (technical, economic, governance) is still limited, with minimal premiums—an underdeveloped blue ocean.

Decentralized options are not new; DeFi has seen many attempts, including options AMMs, perpetual options, options vaults (DOV), but most have not stood the test of time. Some teams continue building in this space, and more tokenized volatility yield products may emerge.

Market opportunities show fierce competition between CeFi options and perpetuals. By 2025, CeFi options open interest is around $30-50 billion; current on-chain options open interest is about $1.8 billion (mainly Derive). Options premiums and funding fees are both persistent, but bundling and capturing are complex. As risk managers’ asset allocations become more refined and competition intensifies, vaults may incorporate options to maintain advantages.

Insurance premiums remain small, mostly from established players like Nexus Mutual. In 2025, the platform earns over $5.5 million selling insurance, with risk concentrated in protocols like Fasanara, Infinifi, Dialectic. In the low-yield environment, on-chain risk pricing is complex, and the insurance sector remains early-stage. As yield sources diversify and protocols mature, demand may grow. Indirect mechanisms like Aave’s Umbrella insurance and fee-based reserves can mitigate single-protocol insolvency risks but are less suited for structured products.

Real-World Assets (RWA)

Since this article focuses on on-chain yield sources, RWA is briefly estimated for market reference. According to rwa.xyz, RWA total value grows from about $5.6 billion in 2025 to $27 billion now, with U.S. Treasuries making up the largest share (~41%), private credit about 25%. Applying average yields to these assets and including income from commodities and real estate, estimated annual RWA yields are roughly $600 million to $900 million.

In summary, total on-chain yield in 2025 is about $8 billion, but the distribution is uneven. Some yields, like AMM fees and volatility profits, are hard to capture reliably, and safe yields are more limited. However, ongoing DeFi innovation and native crypto opportunities suggest that more on-chain fee and yield capture mechanisms will emerge.

Returning to Sky

Focusing on Sky’s 2025 performance and current setup, we analyze how it integrates various yield sources into a coherent, stable strategy.

Sky 2025 – Early 2026 Financial Briefing, data sources: Dune, Sky Forum

Simplified asset overview and expected yield structure, from Sky official

Currently, about 55% of Sky’s revenue comes from its own direct lending (PSM, crypto vaults, traditional RWA, SKY token-backed loans), with the remaining 45% from unsecured credit lines to related ecosystems.

PSM (Pegged Stable Module) remains the largest income source, contributing about 44%. In 2025, roughly 27% of PSM’s total income derives from USDC deposits at Coinbase.

Crypto vaults generate moderate yields, mainly from ETH and WSETH collateral.

Excluding PSM, direct exposure to traditional RWA continues to decline, mainly through new ecosystem partner Grove Finance. Its accumulated income includes 47% from Janus Henderson JAAA (cash-like instruments and mortgage-backed notes), 16.7% from Janus Henderson JTRSY (US short-term Treasuries), and 15.2% from BlackRock BUIDL-I (cash-like instruments).

Obex is a new credit line partner, with about $600 million fully deployed in syrupUSD, aiming to become a core platform for structured yield products.

Spark’s income and asset share have steadily increased since April 2025, surpassing the d3m lending model and becoming a key allocation channel, contributing 20%-65% of revenue. Spark functions as both asset manager and risk manager:

• Operating its own lending protocol Sparklend (a fork of Aave v3);
• Within its risk framework, flexibly deploying yield opportunities above borrowing costs (SSR+0.3%).

Where does Spark’s yield come from?

Data sources: Dune, Spark official

Sparklend’s lending is Spark’s largest liquidity asset, contributing 37% of gross revenue in February 2026 (excluding distribution rewards), consistent with Q3-Q4 levels (33.4%, 37.8%). Why do users borrow on Sparklend?

Top 15 borrowers on Sparklend

Since only WSETH is accepted as collateral, core use is classic circular strategies, accounting for about 50% of total borrowing, mainly generating reserve income, with participants like Ipor, Threehouse, Mellow, Summerfi.

Most other borrowing is in stablecoins. USDT’s interest rate model is independent of SSR, based on utilization, which may explain why it’s the second-largest borrowing asset (~25%). Why do users borrow other stablecoins at above-market rates? The reasons are unclear. For example, top borrower addresses on Spark are linked to 7 Siblings, accounting for at least 15% of total borrowing (with idle ETH holdings). Recent activities include leveraged long ETH and SKY positions, and SKY staking—likely motivated by ecosystem fund reuse, yield sharing, and SPK mining. The second-largest borrower continuously buys and stakes SKY via TWAP.

Over half of the income comes from other yield strategies, dynamically adjusted based on current market conditions. For example, in Q3-Q4, much income came from Morpho, Maple, Ethena; now daily income mainly from Maple (10.6%), Anchorage (11.2%), and PayPal deposits (30%).

How does Maple generate yield? About 28% of funds are lent to institutions at 6%-9% annualized, collateralized by blue-chip assets like BTC, XRP, SOL; the rest are deployed into yield assets, with one-third flowing back into USDS savings, and the rest allocated to Aave, PYUSD, Superstate, and syrupUSD pools.

What about Anchorage? It provides BTC loans to institutions at below 6.5% interest.

Morpho and Ethena? Their yield logic has been detailed earlier.

Finally, a brief overview of Sky’s financials: total revenue in 2025 is $338 million, paying $194 million to savings users. The current USDS savings pool of $6.7 billion pays a 3.75% rate, with daily payouts of about $688,000, and daily income around $1.17 million. Based on current revenue structure, roughly 50% of Spark’s income is from on-chain sources, and 70% of Sky’s total income is from off-chain sources. What does this mean for the market?

Regardless of origin, sustainable on-chain yield sources have emerged, less dependent on market activity and leverage demand, with weaker correlation to traditional on-chain yields, providing a stable capital refuge for individuals and protocols. For Sky, diversified and flexible allocations enable quick adjustments when on-chain risk appetite rises. Moreover, as USDS becomes the largest real-time yield stablecoin, the behavior of leading borrowers suggests Sky indirectly supports other protocols in maintaining interest rates.

Some argue traditional finance is swallowing DeFi; indeed, many tokenized RWAs face permission constraints, limiting user participation. But even if yields originate off-chain, they ultimately flow back on-chain for distribution. This trend is crucial, bringing liquidity and diversification to DeFi, allowing permissionless protocols to benefit. Furthermore, it may drive the development and scaling of next-generation yield derivatives like fixed-rate products, interest rate swaps, risk layering, and structured products. Is a true DeFi renaissance imminent?