How Jupiter Lend, Project 0 and Loopscale Are Rethinking Solana Lending

Solana lending isn’t just a fight over TVL between a handful of similarly shaped protocols. Underneath the scale race sits a more interesting split: three genuinely different architectures for how a lending market should work, each making a different bet about what borrowers and lenders actually want.

Jupiter Lend bets on a unified pool wired into Jupiter’s trading infrastructure. Project 0, the protocol formerly known as MarginFi, bets on becoming something closer to a prime broker than a money market. And Loopscale bets that fixed-rate, order-book lending is a better fit for real-world assets and structured credit than the variable-rate pools everyone else runs. None of these three are really trying to win the same game.

Jupiter Lend

Jupiter Lend’s core pitch hasn’t changed since its August 2025 launch: use Jupiter’s existing trading distribution to bootstrap a lending market, then merge lending and trading economics through features like Smart Collateral and Smart Debt, which let eligible deposits and borrows double as trading liquidity. It currently holds roughly $1.05 billion in TVL and around $907 million in active loans, putting it firmly in second place behind Kamino among Solana’s major lending markets.

The model is still a pooled one at heart: deposit into a shared pool, borrow against collateral, rates float with utilization, but the trading integration is what differentiates it from a plain money market. It’s the same basic shape as Kamino or Aave, with an added yield mechanic bolted onto the side.

Project 0 (formerly MarginFi)

MarginFi was, at its 2024 peak, a top-three Solana lender with over $800 million in TVL and one of the most widely integrated money markets on the chain, accepted as collateral logic or liquidity by more than 100 other Solana protocols. Then came a leadership dispute, a year-plus of delayed token plans, and a steady bleed of deposits that had little to do with the underlying lending contracts and everything to do with trust and incentives drying up.

The team’s answer wasn’t to fix MarginFi as a standalone money market. It was to rebuild the brand entirely. In late 2025, MarginFi relaunched as Project 0, a “DeFi-native prime broker” that lets a user borrow against their entire portfolio across multiple venues, Kamino, Drift, Jupiter, and others, from a single unified credit line, rather than posting collateral separately on each one. The underlying marginfi-v2 lending contracts are still there and still working; what changed is the layer built on top and the pitch to users.

The numbers tell an honest story about where that leaves things today. The core lending markets now hold around $27.6 million in TVL, down roughly 96% from the 2024 peak. Most of that decline happened before the rebrand, not because of it. Project 0 isn’t dead, and the prime-broker thesis is a genuinely different product than what MarginFi used to be, but it’s also unproven at scale and running on a fraction of the liquidity it once had.

Loopscale

Loopscale takes the architecture question in a different direction entirely. Instead of a shared pool with algorithmically-set floating rates, it runs an on-chain order book where lenders post custom offers, specific collateral, specific rates, specific durations, and borrowers get matched against those offers directly. That gives both sides a fixed rate and a fixed term, something none of Solana’s pooled lenders can offer.

That design has made Loopscale the default venue on Solana for tokenized real-world assets and structured credit. The protocol currently sits at roughly $95 million in TVL, and by Q1 2026 real-world assets made up close to 40% of its deposits, spanning tokenized government bonds, reinsurance yield, and other instruments that don’t fit neatly into a floating-rate pool. It’s processed more than $1 billion in cumulative borrowing volume since its April 2025 launch, and survived a $5.8 million oracle exploit early on that it resolved through a negotiated white-hat return of 90% of the stolen funds.

The tradeoff is liquidity flexibility. A fixed-term loan can’t just be repaid early the way a pooled loan can; unwinding a position before term means refinancing or selling it back into the order book. For lenders chasing tailored, predictable exposure, that’s the point. For anyone who wants same-day flexibility, it’s a real constraint.

Three Bets, Not One Race

Put the three side by side, and the differences run deeper than TVL. Jupiter Lend is making a distribution bet: use an existing user base and existing trading flow to make a fairly standard pooled lending market more capital-efficient. Project 0 is making an architecture bet on the layer above lending entirely, wagering that unified cross-venue margin is worth more to serious DeFi users than being the biggest single money market. Loopscale is making a matching-model bet that fixed terms and custom offers serve certain kinds of capital, especially real-world assets, better than utilization curves ever could.

That’s a genuinely different lens than looking at these three purely as competitors chasing the same deposits. A tokenized bond issuer probably has no use for Jupiter Lend’s Smart Collateral mechanic. A trader who wants unified margin across five different protocols has no reason to care whether Loopscale’s rates are fixed or floating. They’re adjacent markets more than they are a single contested one.

What Comes Next

The open questions differ for each. For Jupiter Lend, it’s whether the Smart Collateral and Smart Debt mechanics generate durable yield once trading activity normalizes rather than just riding Jupiter’s existing distribution. For Project 0, it’s whether a prime-broker pitch can rebuild deposits that a straightforward money market couldn’t hold onto, especially once the delayed token finally launches and incentives get tested against real usage. For Loopscale, it’s whether real-world asset issuance keeps growing fast enough to justify order-book lending’s flexibility tradeoff, or whether that niche stays a niche.

None of these are really racing toward the same finish line. They’re three different answers to the question of what a lending market is even for.