Bitcoin: Digital gold Ethereum: The world computer Base: The blockchain for global finance Ondo: Institutional financial markets onchain Bittensor: Decentralized AI intelligence marketplace Hyperliquid, Lighter: The onchain trading & financial markets layer Aave: Onchain credit & lending market Curve: The home of stablecoins & stablecoin liquidity layer Convex: Curve liquidity coordination layer Frax: better money w/ $frxUSD did i miss something?

After comparing both architectures together with @royal1dd, I came away with a much stronger appreciation for how much protocol design shapes the entire user experience. What impressed me most about Curve is how coherent the entire architecture feels. The CDP model, debt creation, collateral management and LLAMMA all fit together naturally. Every component reinforces the next one, and that level of consistency is something I genuinely appreciate from an engineering perspective. What stands out most is the way collateral remains at the center of the system. LLAMMA continuously manages collateral as market conditions evolve, allowing positions to adapt throughout the move. I think that creates an elegant balance between capital efficiency and risk management, because the protocol is actively working with the position throughout different market conditions. To me, LLAMMA feels like the logical outcome of the architecture Curve chose for crvUSD. f(x) gives a completely different perspective on protocol design. The part I find most interesting is the focus on leverage itself. Exposure becomes something dynamic that can adjust together with market conditions, creating a completely different way to think about leveraged positions. I think this approach is especially interesting because it changes the role of leverage inside the system. Instead of treating leverage as a fixed parameter, f(x) turns it into something that can actively evolve as the market moves. That creates a very different user experience. The position adapts over time, while the user gets a more flexible way to access leveraged exposure without managing every adjustment manually. The mechanism behind xPOSITIONS shows that there is still a lot of unexplored design space around leverage in DeFi. Looking at both protocols together, the biggest difference comes from what each protocol decided to optimize. Curve invested an incredible amount of engineering into collateral management and built an architecture where every component contributes to that objective. f(x) invested its engineering effort into making leverage itself more dynamic, creating a different experience around leveraged exposure. Both approaches have their own strengths. Curve provides a deeply integrated lending architecture where debt creation, collateral management and risk adjustment work together. f(x) provides a flexible leverage framework where exposure management becomes the core mechanism and opens new possibilities for how leveraged products can be designed. I think the interesting part is that both designs serve different preferences and different ways of interacting with risk. Some users may value a system built around continuous collateral management, while others may prefer a model focused on dynamic leverage and exposure. Both approaches add something valuable to the DeFi landscape, and having different architectures gives users more options depending on what they are trying to achieve. This comparison was never about finding a winner. It was about understanding why two strong engineering teams arrived at two different solutions to the same challenge. be tuned and check out the sum up from Royal if you have something to add, let me know🫡

Another interesting piece being built around the @CurveFinance ecosystem. trUSD/strUSD shows how liquidity, yield and incentives can be combined into new strategies for DeFi users. The Curve + Convex flywheel keeps expanding with more integrations finding their place in the ecosystem. Excited to see how @tori_finance develops from here👀

Part 2 Llamand (Curve) vs xPOSITIONS (f(x) Protocol - @Diphunter18 x Royal - Investigation xPOSITONS / soon fx100 @protocol_fx introduces a different approach to leverage a lending position that prioritizes capital efficiency & still provides structural safety over classical liquidation mechanics. But first things first, how do xPOSITIONS actually work, and how could fx100 behave once it goes live? At its core, f(x) introduces what you could call “internal managed leverage.” Instead of relying on classical lending markets with liquidation thresholds and borrow rates, the system removes debt entirely and replaces it with internally structured exposure that gets actively rebalanced. When opening an xPOSITION, you deposit collateral and mint $fxUSD, f(x)’s own decentralized stablecoin, against it via a flashloan, as well there is a one-time entry fee, but no ongoing borrow APY, which makes the position easier to hold and reason about over time. The core mechanism here is continuous rebalancing. Rather than letting positions drift toward liquidation, f(x) dynamically adjusts exposure based on market movements. If price moves against you, the system reduces leverage to prevent a hard liquidation & if the price moves in your favor, exposure can expand again depending on the internal state of the system. So instead of hard liquidations, you get soft adjustments. Example: say you open exposure around $3. Price trends down into the $1.6–$1.9 range over time. In a traditional setup, you’re either getting liquidated or actively managing the position (adding collateral, deleveraging, etc.). With f(x), none of that is a must, the system continuously reduces your exposure as price declines. The position survives, but with lower exposure. That’s the trade-off, you’re not maximizing upside at all times, but you’re also not exposed to liquidation drama. If the market recovers, you participate, but from a reduced base, since leverage was scaled down on the way down. This makes the system structurally different from anything debt-based. To exit an xPOSITION, you first reduce or fully close the position by repaying the outstanding exposure. Once that’s done, the remaining collateral is released back to you, in other words, you unwind the position first, and only then withdraw the collateral. Looking ahead, fx100 (50x) will likely push a similar system into much higher leverage. There are no full details yet, but the key thing to watch is how aggressive the rebalancing becomes. If the current design holds, it could allow for very high leverage without classic liquidations, just with faster and more noticeable adjustments to your exposure along the way, will update on this fs. Next part, @Diphunter18 will continue with his final statement, on how he evaluates both facilities & maybe even what he prefers or what he suggests depending on the situation the potential user is in.

Part 1, LLAMMA (Curve) vs xPOSITIONS (f(x) Protocol)- Diphunter x @royal1dd - Investigation LLAMMA explains how $crvUSD manages collateral risk. Looking beyond the liquidation mechanism, the CDP model explains why Curve was able to build this system in the first place. Users provide collateral and mint crvUSD against their position. The protocol creates the debt directly and manages the relationship between collateral and borrowed stablecoins. This architecture matters because LLAMMA is built around continuous collateral management. To adjust a position through different price bands, the protocol needs direct insight into the collateral, the debt, and the state of the position. This is where CDPs differ from traditional money markets. In systems like Aave, borrowers access liquidity supplied by other users. Depositors provide assets, borrowers use that liquidity, and interest rates adjust based on supply and demand. With a CDP, the stablecoin is created when users borrow against their collateral, creating a debt position that is managed directly by the protocol. For LLAMMA, this design choice is what makes soft liquidation possible. The system can continuously adjust collateral exposure as prices move instead of relying on one liquidation event. The connection between debt creation and collateral management allows LLAMMA to gradually shift positions through different price bands. Another example, A user deposits $20,000 worth of ETH and mints 10,000 crvUSD against the position. At the beginning, the position is mostly exposed to ETH while the collateral value remains comfortably above the debt. Now imagine ETH drops significantly and the collateral value falls toward the liquidation range. In a traditional liquidation model, the position approaches a specific threshold where collateral is sold and the user loses control over the process. With LLAMMA, the position moves through different price bands. As the market declines, parts of the collateral are gradually converted into crvUSD to protect the system. The user still has an active position, but the composition of the collateral changes as the market moves. If ETH later recovers, LLAMMA can move the position back in the opposite direction, gradually increasing ETH exposure again. The important difference is that risk management happens continuously instead of at a single liquidation point. This is the core idea behind Curve's approach. The CDP model gives the protocol control over the debt position. LLAMMA uses that control to create a liquidation mechanism that adapts to market movements. Every lending model comes with different trade-offs. Money markets optimize around shared liquidity and efficient lending markets. CDP systems optimize around protocol-controlled debt creation and custom risk management. Curve chose this path because crvUSD was designed around a different approach to handling volatility. One of the things I like about DeFi is seeing different protocols look at the same challenge and build completely different architectures around it. Curve chose a CDP model to enable LLAMMA and its approach to continuous collateral management. In Part 2, Royal will continue this comparison by looking at another approach with f(x) and how xPOSITIONS tackles leverage and risk from a different direction.

>More liquidity on mainnet >more liquidity comes to @llamalend and @CurveFinance / @ConvexFinance >more people levering up stable LPs on @ResupplyFi >more fees go to $RSUP stakers >more competition and incentives needed for $reUSD LP depth on Convex > $CVX lockers get heavily incentivized for their votes Ya dig?

Concepts are useful. Seeing them play out with real numbers makes them much easier to understand. Let's walk through a simple $crvUSD position. Alice deposits 10 ETH as collateral when ETH trades at $3,000. Her collateral is worth $30,000. She decides to borrow 15,000 crvUSD against it. Like every borrowing position, Alice pays interest on her debt. The borrow rate is dynamic and changes depending on market conditions and protocol parameters. When the loan is created, LLAMMA does not assign Alice one single liquidation price. Instead, her position is placed across a liquidation range made up of multiple price bands. This range determines where LLAMMA starts adjusting the position if ETH moves against her. As long as ETH stays above the liquidation range, nothing happens. Her collateral remains entirely in ETH. Now the market starts moving. ETH falls to $2,850. Alice's position enters the liquidation range. This is where Curve's liquidation protection becomes active. Like other lending protocols, @llamalend tracks the health of the position. The difference is what happens as that health changes. On traditional lending protocols, reaching the liquidation threshold can trigger a liquidation event. In LlamaLend, LLAMMA gradually rebalances the collateral while the position moves through the liquidation range. Think of the price bands as checkpoints rather than a single trigger point. As ETH moves through these bands, a portion of Alice's ETH is exchanged for crvUSD. If the price continues lower, more of the position is gradually adjusted. There is no single moment where the entire position changes. Everything happens progressively as the market moves. ETH reaches $2,700. More of the position has moved through the liquidation range. The health of the position continues to update based on where the collateral sits within that range. Now imagine buyers return. ETH recovers to $2,950. Because Alice's position has not reached full liquidation, LLAMMA can start moving in the opposite direction. The AMM gradually converts part of the crvUSD back into ETH as the price recovers. The position adapts to changing market conditions instead of only reacting after liquidation has already happened. Entering the liquidation range does not automatically mean losing the position. However, once soft liquidation starts, the recommended approach is usually to improve the position by reducing debt rather than simply adding more collateral. That does not mean losses disappear. If ETH continues falling or stays volatile for a long period, collateral can still be converted at lower prices. Soft liquidation reduces the shock of traditional liquidations. It does not eliminate market risk. For me, that's the real innovation behind LLAMMA. Curve redesigned how a lending position behaves while markets are moving. But managing collateral risk was only one part of building a complete stablecoin system. A stablecoin also needs something else, Maintaining its peg and managing liquidity around it. That's where the next piece of crvUSD's design comes in. PegKeepers.

A good Stablecoins is the one you don’t need to be checking the peg. You don’t even care about the highest yield. You don’t even care on the cleanest design. The big battle is on distribution. More places it’s accepted → more people actually use it → deeper liquidity → stronger trust → even more distribution. That’s the network effect flywheel. Once it starts spinning hard, the lead becomes structural. Almost impossible to reverse. Most new stablecoins still obsess over the collateral, the yield strategy, or the decentralization narrative. Genius Act era is coming. The ones that will matter in 5 years obsess over one thing: getting into every wallet, every protocol, every payment rail, and every real-world flow as fast as possible. Distribution isn’t marketing. Distribution is the product. You cannot vibecode distribution. ⚔️

Convex is moving to onchain voting. One of the best new features is separate delegation for gauge weights and for governance DAO votes. On the Lock CVX page, delegate your Gauge weight votes to VOTIUM.ETH And delegate your DAO votes to Convex Team!

The easiest way to understand LLAMMA is to forget everything you know about traditional liquidations for a minute. Imagine you deposit ETH as collateral and mint $crvUSD. As long as ETH trades comfortably above your liquidation range, nothing happens. Your collateral remains entirely in ETH. Now ETH starts to fall. Like other lending protocols, @llamalend also tracks the health of your position through a health factor. The difference is what happens as that health deteriorates. On traditional lending protocols, a low health factor eventually leads to a liquidation event. In LlamaLend, the position enters a liquidation range where LLAMMA starts managing the collateral gradually instead of waiting for one final trigger. Think of these bands as checkpoints rather than trigger points. As the market moves through each band, LLAMMA gradually rebalances your position. A small portion of your ETH is exchanged for crvUSD. If the price continues to fall, more collateral is gradually converted. There is no single moment where everything suddenly changes. The transition is continuous. That's why Curve calls it a soft liquidation. The interesting part comes next. If ETH recovers before moving through the entire range, LLAMMA starts working in the opposite direction. The AMM gradually converts crvUSD back into ETH. Your position doesn't simply survive. It can partially rebuild itself as the market recovers. This is also why LLAMMA depends on external price oracles. The protocol always needs to know where the market is trading so it can determine which price band the position currently occupies. None of this means losses disappear. If the market remains inside the liquidation range for a long time, or continues falling without recovering, part of the collateral may still be sold at lower prices. Soft liquidations reduce the shock of traditional liquidations. They don't eliminate market risk. That trade-off is exactly what makes LLAMMA so interesting to me. It accepts that volatility cannot be removed. Instead, it changes how the protocol responds to it. In the next post, I'll walk through a complete crvUSD position step by step, so you can see exactly what happens as the price moves through each band.

Excellent overview. As mentioned earlier, 🟦 Substantia Core Fund is gradually moving its Bitcoin and ETH from Aave to LlamaLend @llamalend Moreover, all future ⬜️ The Holding @theholding_ strategies – as well as all new onchain companies built by The Holding at – that use collateral assets like BTC & ETH for borrowing will be built on top of LlamaLend. Why? Because its liquidation-protection mechanics are highly attractive and help reduce risk. And for us, risk reduction is a priority. Every decision is thought through carefully, in detail, and with a long-term view.

Traditional lending systems are built around a simple assumption. At some point, a position reaches a level where action has to be taken. A user deposits collateral and borrows against it. The position remains healthy as long as the collateral value stays above a certain threshold. Once that threshold is crossed, liquidation begins. This model has powered a large part of DeFi lending. Money markets like @aave work by connecting borrowers with liquidity supplied by other users. Users deposit assets into a pool, and borrowers take liquidity from that pool. CDP systems like $crvUSD follow a different structure. Users provide collateral and mint a new stablecoin against their position, creating additional stablecoin supply that can be used throughout the ecosystem. The structures are different, but both systems face the same fundamental question: What happens when markets move quickly? A sharp price drop can push many positions toward liquidation at the same time. The protocol needs to protect itself, liquidators step in, collateral is sold, and the borrower's position changes immediately. The mechanism works exactly as designed. The challenge is the transition itself. A position can move from healthy to liquidated because of a short period of volatility. The market moves continuously, but liquidation usually happens at a specific moment. That gap is where Curve focused its research. Instead of only optimizing the existing liquidation process, Curve redesigned how collateral reacts to changing market conditions. This idea became LLAMMA. Lending-Liquidating AMM Algorithm. The core concept is simple. Collateral does not move directly from "safe" to "liquidated". LLAMMA creates a gradual process where collateral can shift between assets as the market moves. This is known as a soft liquidation. When collateral prices move against the borrower, LLAMMA gradually converts parts of the collateral into the borrowed stablecoin. If the market recovers, the process can reverse and the position can move back toward its original state. The position continues to exist. The system adapts. The key innovation behind this design is the use of price bands. Instead of one liquidation point, collateral is distributed across different price ranges. As the market moves through these ranges, the AMM continuously adjusts the position. Risk becomes something that changes over time instead of something that suddenly switches on. This changes the relationship between borrowers, protocols, and liquidators. The goal is to manage collateral throughout the entire movement of the market. This follows the same design philosophy that has shaped @CurveFinance from the beginning. StableSwap redesigned how stable assets trade. veCRV redesigned how governance power is created. Gauges redesigned how liquidity incentives are allocated. LLAMMA redesigned how collateral risk can be managed. The mechanism behind LLAMMA is what gives crvUSD a fundamentally different approach to collateral management. The next step is looking at a real example. What actually happens when someone opens a crvUSD loan, the collateral price falls, and the position moves through different price bands?