Sky Agents are operating like an onchain credit network. The part I find most interesting is how @SkyMoney distributes its balance sheet across specialized agents @grovedotfinance has drawn around $2.75B, @sparkfinance ~$2.74B, and @obexincubator ~$403M. Each agent has its own mandate, deploying USDS across markets like DeFi lending, tokenized Treasuries, RWA, and private credit. Governance sets the debt ceilings, base rates, and risk limits. The core protocol provides the balance sheet. The structure starts to look a lot like a financial institution with multiple specialized credit desks, except most of the rules, exposures, and capital flows are visible onchain. USDS is gradually becoming the funding layer for a distributed credit network. Sky could become a place where onchain capital is originated, allocated, and recycled across multiple credit markets.

Looking at the growth of the FalconX Credit Facility on @paretocredit, one thing stands out to me. The facility did not grow once and then stop. Its size kept expanding from roughly $40M → $60M → $100M → $125M → $170M. For private credit, I think this matters more than just looking at the headline APY. Lenders usually keep adding capital when they’re comfortable with the borrower, the underwriting, and the risk structure behind the deal. The setup here is pretty straightforward: + @FalconXGlobal creates real borrowing demand through its prime brokerage business + @M11Credit handles underwriting and risk monitoring + Pareto brings the facility, settlement, and capital flows onchain LPs get exposure through the tokenized credit vault There are also SPVs, covenants, collateral monitoring, and first-loss capital from FalconX to give lenders more protection. So the model is actually pretty simple. + TradFi still does what it’s good at: underwriting credit. + Onchain rails make the facility easier to track, settle, and scale. Going from $40M to $170M is a good example of what that can look like in practice.

I think @paretocredit’s $3B in credit extended is more interesting when you look at it as proof of infrastructure, rather than just a volume milestone. Every completed lending cycle adds another onchain track record: borrowers draw capital, pay interest, repay the loan, and potentially come back for the next cycle. > For private credit, that is the part that matters. The longer the repayment history, the more data lenders have to assess risk. At the same time, Pareto builds up infrastructure that has already processed billions of dollars in real credit flows. > That creates the foundation for Pareto to turn the same stack into Pareto Studio. Instead of every borrower or credit manager rebuilding the legal structure, settlement, compliance, and accounting layer from scratch, they can launch a new facility on top of infrastructure that has already been battle-tested. From my view, the $3B milestone is bigger than the volume itself. It shows onchain credit infrastructure can now be standardized, reused, and scaled.

Capital efficiency is still one of the biggest unlocks left for DeFi. @3janexyz takes a pretty interesting approach to pushing it further with a new primitive called Levered Callable Capital (LCC), designed around a simple question: Why should 100% of committed capital need to be prefunded when only a fraction of it may actually be required at any given time? If a lender commits $100m but only $30m is currently drawn, having the remaining $70m sitting idle creates enormous cash drag. LCC introduces a different model. Instead of depositing the full amount upfront, capital providers post a fraction of their commitment as yield-bearing margin, while committing to provide the full capital if called. To make this easier to understand, here’s a simple example: With a 7.5% margin ratio, $75k of posted margin can back $1m of callable capital while the full $1m standby commitment remains productive in your custody (until it’s actually called). And here's where the economics get pretty interesting: At launch terms, LCC is designed to generate ~20% APY on the posted margin, coming from: 1️⃣ the margin asset's underlying base yield (+3.5% APY) 2️⃣ a standby fee for keeping capital available (+16.7% via 125bps on standby notional) Note: If the capital is actually called, the staker also receives a separate funding bonus. So you're effectively getting paid to keep a credible balance sheet commitment available, without having to pre-fund 100% of it from day one. If the facility actually needs liquidity, 3Jane issues a capital call and the staker funds the required USDC within a defined window. That funded capital then becomes USD3 exposure, subject to a 35-day cooldown. Failure to meet the call would result in margin being slashed and a backstop auction is used to source replacement capital. So 3Jane gets something incredibly valuable: Execution certainty without requiring 1:1 prefunding. And capital providers get compensated for providing that certainty without leaving the entirety of their committed capital sitting idle. In many ways you're effectively selling a liquidity option, where you: 🔸 keep uncalled capital productive 🔸 earn ~20% APY on posted margin for standing ready 🔸 receive additional compensation if called 🔸 deploy the full capital only when it is actually required This essentially mirrors to how TradFi has relied heavily on callable commitments, revolvers and unfunded capital commitments forever. What's interesting is bringing that structure onchain through a programmable system that potentially creates an entirely new level of capital efficiency for onchain credit. DeFi doesn't necessarily need more capital, it needs better ways to use the capital already here. IMO, we'll increasingly see onchain credit evolve away from purely fully-funded liquidity toward more credible, programmable balance-sheet commitments. LCC is a pretty interesting step in that direction. Feel free to check it out for more details 👇

Looking at the growth of the FalconX Credit Facility on @paretocredit, one thing stands out to me. The facility did not grow once and then stop. Its size kept expanding from roughly $40M → $60M → $100M → $125M → $170M. For private credit, I think this matters more than just looking at the headline APY. Lenders usually keep adding capital when they’re comfortable with the borrower, the underwriting, and the risk structure behind the deal. The setup here is pretty straightforward: + @FalconXGlobal creates real borrowing demand through its prime brokerage business + @M11Credit handles underwriting and risk monitoring + Pareto brings the facility, settlement, and capital flows onchain LPs get exposure through the tokenized credit vault There are also SPVs, covenants, collateral monitoring, and first-loss capital from FalconX to give lenders more protection. So the model is actually pretty simple. + TradFi still does what it’s good at: underwriting credit. + Onchain rails make the facility easier to track, settle, and scale. Going from $40M to $170M is a good example of what that can look like in practice.

Sky Agents are operating like an onchain credit network. The part I find most interesting is how @SkyMoney distributes its balance sheet across specialized agents @grovedotfinance has drawn around $2.75B, @sparkfinance ~$2.74B, and @obexincubator ~$403M. Each agent has its own mandate, deploying USDS across markets like DeFi lending, tokenized Treasuries, RWA, and private credit. Governance sets the debt ceilings, base rates, and risk limits. The core protocol provides the balance sheet. The structure starts to look a lot like a financial institution with multiple specialized credit desks, except most of the rules, exposures, and capital flows are visible onchain. USDS is gradually becoming the funding layer for a distributed credit network. Sky could become a place where onchain capital is originated, allocated, and recycled across multiple credit markets.

$3,000,000,000 in credit extended through Pareto. $3B of financing provided to institutional borrowers, backed by lenders earning real yield onchain. The credit layer keeps compounding.

Tokenized stocks face the problem of liquidity fragmentation. Take $TSLA stocks for example When it onboarded onchain, you can have: + Custodial stock tokens + Linked securities + Security-based swaps + Tokenized funds + Stock perps They all give exposure to TSLA, but they are different financial instruments, so liquidity cannot simply be merged into one pool. And even within the same structure, @Ondo, @xStocksFi, or @RobinhoodCrypto can issue their own versions. Same underlying, but different market makers, venues, and liquidity The advantage is that more platforms also bring new users and new demand into US equities. But more access also creates more liquidity islands. At some point, tokenized stocks may need their own liquidity orchestration layer. It would route liquidity and handle settlement, compliance, redemption, and different legal structures. Who becomes the liquidity layer for tokenized stocks?

A useful tool for tracking onchain neobanks and card payments. @Glasscade_xyz standardizes onchain data to show how the market is growing, how market share is shifting, and how competition is evolving. A few data points stand out: - Top 5 market share: 96.1% → 89.3% - Effective competitors: 1.3 → 3.1 - RedotPay: 87.1% → 52.8% - KAST: +11.3pp - Etherfi Cash: +9.9pp Across 21 tracked programs, 14 are expanding, and 5 are contracting. Market share seems to be moving away from one dominant player and spreading across more emerging programs.

A useful tool for tracking onchain neobanks and card payments. @Glasscade_xyz standardizes onchain data to show how the market is growing, how market share is shifting, and how competition is evolving. A few data points stand out: - Top 5 market share: 96.1% → 89.3% - Effective competitors: 1.3 → 3.1 - RedotPay: 87.1% → 52.8% - KAST: +11.3pp - Etherfi Cash: +9.9pp Across 21 tracked programs, 14 are expanding, and 5 are contracting. Market share seems to be moving away from one dominant player and spreading across more emerging programs.

Tokenized stocks face the problem of liquidity fragmentation. Take $TSLA stocks for example When it onboarded onchain, you can have: + Custodial stock tokens + Linked securities + Security-based swaps + Tokenized funds + Stock perps They all give exposure to TSLA, but they are different financial instruments, so liquidity cannot simply be merged into one pool. And even within the same structure, @Ondo, @xStocksFi, or @RobinhoodCrypto can issue their own versions. Same underlying, but different market makers, venues, and liquidity The advantage is that more platforms also bring new users and new demand into US equities. But more access also creates more liquidity islands. At some point, tokenized stocks may need their own liquidity orchestration layer. It would route liquidity and handle settlement, compliance, redemption, and different legal structures. Who becomes the liquidity layer for tokenized stocks?

Looking deeper into active wallet behavior, measured by Arrakis, AA-FalconX stands out for having the most persistent collateral usage in the group. Peak debt reached 29% of acquired notional, with 25% still outstanding today. That means roughly 86% of peak debt remains open, a strong signal of how durable leverage around AA-FalconX has been. Within the same dataset: + mF-ONE peaked at 42%, but only 8% remains outstanding. + syrupUSDC and syrupUSDT only reached around 3% and 1% at peak. Part of the reason may come from the infrastructure @paretocredit x @FalconXGlobal has built around the asset. + AA-FalconX has a @Morpho with up to 77% LTV, allowing LPs to use their credit position as collateral to borrow or deploy additional strategies. + The leverage also looks more strategy-driven than speculative, with players like @gauntlet_xyz and @3f_xyz improving capital efficiency around the asset. In other words, yield from the underlying credit is only the first layer. For a credit asset to become truly productive onchain, the ecosystem around it needs to create additional utility for holders through collateral, leverage, and composability.

"For a credit asset to become truly productive onchain, the ecosystem around it needs to create additional utility for holders through collateral, leverage, and composability" - fully agree with @MahoneDeFi

Looking deeper into active wallet behavior, measured by Arrakis, AA-FalconX stands out for having the most persistent collateral usage in the group. Peak debt reached 29% of acquired notional, with 25% still outstanding today. That means roughly 86% of peak debt remains open, a strong signal of how durable leverage around AA-FalconX has been. Within the same dataset: + mF-ONE peaked at 42%, but only 8% remains outstanding. + syrupUSDC and syrupUSDT only reached around 3% and 1% at peak. Part of the reason may come from the infrastructure @paretocredit x @FalconXGlobal has built around the asset. + AA-FalconX has a @Morpho with up to 77% LTV, allowing LPs to use their credit position as collateral to borrow or deploy additional strategies. + The leverage also looks more strategy-driven than speculative, with players like @gauntlet_xyz and @3f_xyz improving capital efficiency around the asset. In other words, yield from the underlying credit is only the first layer. For a credit asset to become truly productive onchain, the ecosystem around it needs to create additional utility for holders through collateral, leverage, and composability.

> @aave V4 revenue is starting to enter an acceleration phase. In August, V4 hit a new ATH at around $53K in monthly revenue, with Ethereum Core still acting as the main engine at $42K, or nearly 80% of total revenue. At the asset level, $USDG is currently the largest revenue contributor at $21K, ahead of WETH at roughly $12K. That is a pretty clear signal that stablecoin credit is becoming one of V4’s first meaningful economic engines. EtherFi Cash and @avax still contribute very little revenue today, but both matter for a different reason. @ether_fi is migrating its credit stack to V4, while Avalanche extends V4 into a new chain and a different borrower base. If utilization picks up, both can become additional revenue surfaces instead of V4 relying mainly on Ethereum Core. > Another notable data point is user growth. Most of the depositor growth in August came from the EtherFi Cash instance, which reached nearly 46K depositors, while total V4 depositors stood at around 49.5K based on the chart breakdown. This suggests V4 is starting to scale in a different way from V3: not only by bringing users directly into Aave, but also by sitting behind products like EtherFi Cash and letting the distribution layer onboard users for the protocol. Revenue is still small, but both sides of the flywheel are starting to move together: more credit markets → more users → more borrow demand → more revenue.

Aave V4’s Ethereum Core Hub is starting to show what shared liquidity can look like in practice. • $375M in deposits • $152M in active loans • 40% weighted utilisation • $2.36M annualised fees • $510K annualised protocol revenue The key point isn’t just the deposit base. At 40% utilisation, borrowing demand is already meaningful while the hub still has substantial available liquidity to absorb more activity. As utilisation rises, fee generation can scale without needing deposits to grow at the same pace. The current 30D run-rate also implies roughly a 22% protocol take from gross fees, a useful signal that V4 liquidity is beginning to translate into revenue, not just TVL. Source: @Token_Logic

$1B+ in RWA-related deposits now sit across @aave V3, V4, and Horizon. That capital already spans multiple asset classes, lending structures, and forms of onchain collateral. 🧵,

Aave V4’s Ethereum Core Hub is starting to show what shared liquidity can look like in practice. • $375M in deposits • $152M in active loans • 40% weighted utilisation • $2.36M annualised fees • $510K annualised protocol revenue The key point isn’t just the deposit base. At 40% utilisation, borrowing demand is already meaningful while the hub still has substantial available liquidity to absorb more activity. As utilisation rises, fee generation can scale without needing deposits to grow at the same pace. The current 30D run-rate also implies roughly a 22% protocol take from gross fees, a useful signal that V4 liquidity is beginning to translate into revenue, not just TVL. Source: @Token_Logic

The CEX front, DeFi back model is starting to look like a repeatable playbook. Coinbase has already shown this pretty clearly with Morpho on Base: users stay inside Coinbase, while the lending engine and most of the infrastructure run onchain behind the scenes. Now Bitget is following a similar structure for BTC yield. Users deposit BTC on Bitget → BTC gets wrapped into $bgBTC → Gauntlet runs the strategy on Aera → Morpho provides the lending markets → @MorphNetwork handles settlement. The interesting part for me is where @redstone_defi sits in this stack: + RedStone provides the price feeds @Morpho uses to value collateral and trigger liquidations. + When liquidation happens, ATOM, one of RedStone’s core products, continues the flow by running the auction in under 300ms and recapturing OEV So RedStone covers more of the liquidation stack: price discovery → liquidation trigger → OEV recapture It doesn’t directly create yield. It helps the same collateral operate more efficiently while reducing value leakage. That’s why capital efficiency is becoming a bigger part of the oracle thesis.