Solana DeFi runs on four primitive layers: a router that finds prices, a money market that lends against collateral, a perpetuals exchange that prices leverage, and a stablecoin that settles the trade. Each layer has a dominant protocol, and most user activity touches all four within a single session. DeFiLlama lists Solana TVL at roughly $9.4B as of April 2026, with Jupiter, Kamino, Marinade, and Jito holding the largest concentrations. This piece explains what each layer does, names the protocols that matter, and shows how stablecoin liquidity flows through the stack.
​

The Solana DeFi stack is the set of onchain protocols that price, route, and settle financial transactions on the Solana blockchain. Unlike Ethereum, where Uniswap, Aave, and Lido each operate as a self-contained pool, most Solana protocols compose through aggregation. A swap on Jupiter routes through Raydium, Orca, Meteora, and Phoenix in a single transaction. A loan on Kamino borrows liquidity from MarginFi or Drift's spot market. A short on Drift settles against PYUSD or USDC reserves held in the same program.
​

This composability is the structural feature. Solana's monolithic design (single global state, parallel execution via Sealevel) lets one transaction touch a dozen programs without bridging delays. The cost: every protocol shares the same congestion surface. When a memecoin pump saturates the validator schedule, the whole stack slows.
​

The stack matters for stablecoin teams because USDC, PYUSD, USDT, and USDS all live as SPL tokens on Solana, and the protocols that move them (routers, lenders, perps) determine where stablecoin liquidity actually accumulates. The four layers below are the surface a stablecoin transaction touches between deposit and final settlement.
​

Routers aggregate liquidity across the venues that hold it. Jupiter is the dominant aggregator. Per Jupiter's own dashboard, the router clears more than $2B of swap volume on busy days and integrates 30+ DEX programs. The routing engine, called Metis, scans liquidity across constant-product pools (Raydium, Orca), concentrated-liquidity pools (Orca Whirlpools, Meteora DLMM), and central-limit-order-book venues (Phoenix), then splits the trade across the cheapest path.
​

The technical challenge is split-routing without lookahead. Solana's transaction size cap (1232 bytes for legacy, ~1644 with v0 lookup tables) limits how many programs a single instruction can touch. Jupiter solves this by precomputing route graphs offchain and submitting a single versioned transaction with the optimal split. Jupiter's open core shows the routing math, and the v6 API exposes the same algorithm to anyone building on top.
​

Routing matters for stablecoins because the price of USDC ↔ PYUSD ↔ USDT on Solana is a function of pool depth, not a CCTP-style mint-and-burn. 1:1 stablecoin swap mechanics differ on Solana versus Ethereum, and routing depth is one of the reasons. A USDC ↔ PYUSD swap on Solana typically clears in a single Jupiter transaction with sub-1bp spread when pool depth is healthy. The same swap on Ethereum often requires a Uniswap v3 hop or Curve stableswap pool, and the spread depends on which pool the aggregator picks.
​

Other Solana routers exist (DFlow, Hashflow's Solana deployment, Titan), but Jupiter's volume share is roughly 70–85% of aggregator-routed swaps in 2026 per Dune dashboards. Most wallet integrations (Phantom, Backpack, Solflare) use Jupiter as the default swap engine, which reinforces the concentration.
​

Solana lending markets are structurally similar to Aave on Ethereum but with Solana-specific liquidation mechanics. Kamino Finance runs the largest market, with roughly $2.1B in TVL across borrows and concentrated-liquidity vaults per its dashboard. MarginFi and Save (formerly Solend) hold the rest of the lending share. Drift also runs a spot lending market against its perp collateral.
​

Lending pools price assets through Pyth Network oracles. Pyth publishes prices on a sub-second cadence using first-party data from Jane Street, Two Sigma, Jump, and 80+ other publishers. When a position's loan-to-value ratio crosses the liquidation threshold, anyone can call the protocol's liquidate instruction and seize the collateral at a discount. Liquidation is permissionless and competitive.
​

Stablecoin yield on Solana is concentrated in lending markets. USDC supply on Kamino has paid 4–9% APY across 2026, depending on borrow demand, with PYUSD supply paying similar rates after the issuer expanded supply on Solana in late 2025. Borrow rates compress when SOL price is flat (less leverage demand) and spike during volatility events when traders borrow stables to short or borrow SOL/JitoSOL to lever long.
​

The mechanism that distinguishes Kamino from Aave is automated liquidity vaults. Kamino's K-Lend market is paired with concentrated-liquidity vaults that auto-rebalance Orca Whirlpool positions. A user deposits USDC, the protocol routes it into both the lending pool and a USDC/SOL liquidity range, and the vault rebalances when the range moves. This is uncommon on Ethereum, where Aave-style lending and Uniswap-style LP are typically separate user actions.
​

Solana perps are a separate primitive from spot trading. Drift Protocol is the dominant perp DEX with roughly $400M in open interest as of April 2026. Zeta Markets and Jupiter Perps run smaller markets. Mango Markets V4 also operates after rebuilding from the 2022 exploit.
​

Drift uses a hybrid liquidity model. The protocol matches user orders against three liquidity sources in priority order: a just-in-time (JIT) auction where market makers bid, an onchain limit-order book (DLOB), and an automated market maker (AMM) of last resort. This lets large orders fill at maker-quoted prices when MMs are present, then fall back to AMM pricing when they aren't. Drift's documentation details the JIT mechanism and the AMM curve.
​

Perps settle in stablecoins. Drift's collateral is USDC; PnL pays out in USDC. This makes the stablecoin layer the settlement currency for every leveraged position on Solana, and it's why USDC liquidity depth on Solana matters for perp market quality. Funding rates on Drift's BTC-PERP and SOL-PERP markets are usually within 1bp of Binance's funding, which is a tight benchmark for an onchain venue.
​

Jupiter Perps takes a different design. Instead of an order book or hybrid AMM, traders open positions against a pooled liquidity provider (the JLP token). LPs deposit USDC, SOL, ETH, BTC, or USDT into the pool and earn fees from trader losses, taking the other side of all open positions. This concentrates risk on LPs but simplifies the trader experience to a single button.
​

Stablecoins on Solana are SPL tokens (Solana Program Library), not ERC-20s. The key issuers:

Stablecoin liquidity moves between Solana and other chains through three mechanisms. Circle's CCTP burns USDC on the source chain and mints on Solana (or vice versa) using attestations. Wormhole and LayerZero wrap or attest cross-chain transfers for non-CCTP tokens. PYUSD currently moves through LayerZero's OFT standard for cross-chain consistency, while USDT primarily moves through Wormhole.
​

The SPL token model has implications for application design. SPL tokens require an associated token account (ATA) per holder per mint, which means a wallet receiving USDC for the first time pays roughly 0.002 SOL in rent (~$0.30 at recent prices). This is a one-time cost, but it surprises users migrating from EVM. Wallets like Phantom and Backpack now auto-create ATAs and abstract the cost.
​

A representative Solana DeFi session: a user deposits USDC into Kamino, borrows JitoSOL against it, swaps the JitoSOL for SOL on Jupiter (which routes through Raydium and Phoenix), opens a long-SOL perp on Drift collateralized by part of the USDC, and uses the rest to provide liquidity to a USDC/PYUSD pool on Orca. Five protocols, one wallet, one transaction-cluster.
​

This composability is the thing that distinguishes Solana DeFi from EVM DeFi. On Ethereum, the same session would require multiple transactions, gas-fee spikes between them, and bridge hops between L1 and any L2 the user lives on. On Solana, it's atomic.
​

For builders, the implication is that liquidity on one venue tends to attract liquidity on neighbors. Solana DeFi apps covers the broader app surface; this piece focuses on the financial primitives underneath. Within stablecoins specifically, the composition extends to swap aggregators and cross-chain swap infrastructure — Solana sits inside a multi-chain stablecoin graph, and the routing decisions made on Solana feed back into bridge volume.
​

The same composability that makes Solana DeFi efficient also concentrates failure modes. Three risks recur:
​

Oracle dependency. Pyth feeds power lending, perps, and parts of routing. A Pyth outage or staleness window cascades into liquidation pauses and routing errors across the stack. Pyth's documentation describes the publisher quorum and staleness thresholds protocols use to mitigate this. Most lending markets pause liquidations when Pyth flags a price as stale (>10 seconds since last update), which protects users but freezes risk management during the outage window.
​

Validator congestion. When the network saturates (memecoin launches, NFT mints), every protocol on Solana competes for the same blockspace. Priority fees rise, transactions drop. Drift, Kamino, and Jupiter have all built priority-fee logic into their clients to handle this, but the underlying constraint is shared. Solana Compass tracks congestion metrics in real time.
​

SPL composability bugs. Composable programs can call each other in ways the original audits didn't model. The 2022 Mango Markets exploit used a price-manipulation path through a Drift-style AMM that wasn't accounted for in the perp's risk model. This class of bug is structural to permissionless composition, not specific to any one protocol.
​

Stablecoin issuer concentration. Native USDC dominates Solana stablecoin liquidity. Most lending markets, perp DEXs, and routing depth are denominated against USDC. If Circle paused USDC mints on Solana for any reason (compliance, infrastructure), large parts of the stack would experience reduced capacity within hours. PYUSD growth since 2024 has reduced this concentration, but USDC remains the dominant settlement currency.
​

Solana is one of 15 chains Eco supports for stablecoin movement. When a treasury team needs to move USDC, PYUSD, or USDT from Ethereum, Base, or Arbitrum to Solana (or back), Eco Routes selects between CCTP, LayerZero OFT, and Wormhole based on cost, speed, and finality requirements. The Solana DeFi stack assumes stablecoin liquidity is already on chain. Eco handles the part before that. For teams running cross-chain operations, the integration is one API call, and the routing decision is opaque to the user. Cross-chain liquidity protocols covers the broader rail set Eco orchestrates across, and cross-chain stablecoin swap infra shows the venues stablecoin teams compare.
​

By TVL, Jupiter (the router) processes the most volume, while Kamino Finance holds the largest locked TVL at roughly $2.1B as of April 2026 per DeFiLlama. Solana DeFi apps ranks the broader app set with current TVL data.
​

Both are native-issued by Circle and 1:1 redeemable. The technical difference is that Solana USDC is an SPL token; Ethereum USDC is an ERC-20. Circle's CCTP makes the two interchangeable through burn-and-mint. See stablecoin swap mechanics.
​

Solana targets 400ms slot times and sub-second economic finality. In practice, swap-to-confirmation is typically under two seconds. Network congestion can push this higher; priority fees are how protocols compete for blockspace during peak load.
​

No. Solana programs use a different runtime (Sealevel) and account model. Ethereum smart contracts do not deploy on Solana. Cross-chain interaction happens through bridges and message protocols like CCTP, Wormhole, and LayerZero.
​

USDC, with approximately $9.5B in native supply per Circle's April 2026 transparency report. PYUSD has grown rapidly since the Solana migration; USDT remains smaller but present. Digital dollars explained covers issuance mechanics in depth.
​