TL;DR
Kamino is the financial layer of Solana DeFi: K-Lend (borrow and lend), automated liquidity vaults, Multiply (leveraged loops), and in V2 a permissionless market layer plus curator-run vaults. The engineering is serious and the record is clean: roughly $16B in loans issued with no bad debt, including a smooth pass through the October 2025 deleveraging cascade. Audit coverage is deep (10 external audits on the V1 core, plus formal verification) and the Scope oracle stack is well built. The one thing that keeps this at caution rather than solid is admin: the K-Lend program that custodies deposits is upgradeable, and that upgrade authority is a Squads multisig with no confirmed on-chain timelock. That is the classic Solana tradeoff, and it is a real fund-loss surface, so it gates the verdict.
Checklist
Audits and contracts. Broad, real, and public. The V1 core primitive, which V2 builds on and keeps as the foundational layer, has undergone 10 external audits: Sec3, Offside Labs, OtterSec, Ackee, Certora and RX Security, with OpenSec and Certora formal verification, all published in the Kamino-Finance/audits GitHub repo. The offsetting concern is surface area: K-Lend, the Scope oracle, liquidity vaults, farms, Limo, and now the V2 market and vault layers are distinct systems with cross-dependencies.
Admin control. The weak point, and the reason for the verdict. The K-Lend program is upgradeable, and control sits with a Squads multisig (the associated K-Lend multisig is 6hhBGCtmg7tPWUSgp3LG6X2rsmYWAc4tNsA6G4CnfQbM). Kamino's docs also list team multisigs for KLend and KFarms, and the protocol runs multisig control plus an emergency pause. Two things keep this a red rather than a yellow. There is no on-chain timelock confirmed on program upgrades, so a malicious or coerced upgrade could land without a depositor exit window, and the third-party risk rater Exponential reports the multisig has fewer than four signers. KMNO holder governance is only planned for 2026, so today this is few-key, fast admin power over the contracts holding your money. In V2, per-market administration is transferred to a Squads multisig before launch, so the same pattern repeats at the market level.
Oracle. Solid. Prices come through Scope, Kamino's oracle aggregator, which pulls Pyth and Switchboard feeds with confidence-interval and staleness checks rather than trusting a single raw feed. In V2, each permissionless market sets its own oracle configuration, so oracle quality becomes a per-market property. For the core Kamino-run markets this is well handled. For a random permissionless market, checking the feed is on you.
Liquidity and exit. Deep, but utilization-dependent like any lender: if a reserve is fully borrowed you wait for repayments. The real test is stress, and Kamino has passed it. On 10 October 2025, when SOL fell about 14% in a roughly $20B market-wide liquidation cascade, Kamino processed around $20M of its own collateral liquidations cleanly, with no queue failure and no bad debt. Total supply dipped only about 3.5% to roughly $4.5B. The next-gen liquidation engine in V2 is aimed squarely at this surface.
Yield: real vs emission. Real. It is borrow interest plus liquidity and vault fees. KMNO points and token rewards are a separate incentive layer on top, not the base yield, and should not be confused with the organic rate.
Holder concentration. refresh. KMNO circulating supply and top-holder distribution were not confirmed on-chain this run. It matters more going forward than today, because on-chain KMNO governance voting is only planned for 2026. Once live, token concentration becomes a governance-capture vector, and the gap between fully diluted value and market cap points to meaningful future unlocks, so track the emission schedule.
Track record. Strong. Kamino states roughly $16B in cumulative loans issued with zero bad debt, and that record held through both the February and October 2025 liquidation stress events. The protocol publishes monthly risk reports, runs a public risk dashboard with ongoing liquidation stress tests, and uses external risk curators (Gauntlet, Steakhouse Financial, Allez Labs, Rockaway, Sentora) for market parameters. This is professional risk management, and it is the main thing offsetting the admin red.
Worst case
The sharp edge is the upgrade authority. A compromised or coerced multisig ships a malicious K-Lend upgrade and drains reserves before anyone reacts, because no confirmed timelock gives depositors an exit window and the signer set is small. That is the scenario the red flags. Short of a program compromise, the ordinary risks are contained: an oracle-config error or an extreme liquidation cascade socializes bad debt across lenders in that market (K-Lend's stated backstop is proportional loss to users in the affected market), and in V2 a poorly configured permissionless market or an aggressive vault profile concentrates that risk on whoever opted into it. There is no large insurance fund, so position sizing and sticking to the core Kamino-run or reputable-curator markets are your mitigations.
Bottom line
Caution, and a strong one within that band. Kamino is well audited, correctly engineered, deep, and has a genuinely clean bad-debt record through real Solana stress, backed by professional risk curation and public reporting. It does not clear the solid bar for a single structural reason: the program custodying deposits is upgradeable under a small multisig with no confirmed timelock, which is a live fund-loss surface that good risk management does not fully offset. A useful way to frame it is Morpho with one extra risk. Like Morpho, in V2 your market and vault choice drives your market-quality risk (the curator layer). On top of that, unlike Morpho Blue's immutable core, the base program itself is mutable. Usable for the showcase at a sized allocation in the core or reputable-curator markets, not a set-and-forget blue-chip until the upgrade authority is timelocked or frozen and KMNO governance decentralizes (targeted 2026).
Data appendix
- TVL: ~$1.18B (DefiLlama "kamino", Solana-only, live this run). Kamino's own framing runs larger: "market size around $2B" in the V2 proposal and roughly $4.5B gross supply during the Oct 2025 event. TVL (net) and gross market size are different metrics, and the figure is volatile. 90d / ATH: refresh.
- Products: V1 K-Lend (borrow/lend across isolated and cross-collateral markets), automated liquidity vaults, Multiply (leveraged loops), Limo (on-chain limit orders). V2 adds a permissionless Market Layer (any-asset isolated markets, risk oracles, collateral rehypothecation, automation), a Vault Layer (single-asset auto-allocating vaults with curator profiles: Risk-Adjusted, Yield Maximizer, Thematic, plus partner vaults), a next-gen liquidation engine, an on-chain lending orderbook, and spot leverage. Source: docs.kamino.finance + gov.kamino.finance "Introducing Kamino Lend V2".
- Curators (V2): Gauntlet, Steakhouse Financial, Allez Labs, Rockaway and Sentora set market and vault risk parameters. As with Morpho, depositor risk in a curator vault depends on the curator's allocation choices. Source: docs.kamino.finance.
- Audits: 10 external audits on the V1 core: Sec3, Offside Labs, OtterSec, Ackee, Certora, OpenSec and RX Security, with formal verification on K-Lend by Certora and OpenSec. Published in github.com/Kamino-Finance/audits (per-file, per-product) and kamino.com/docs/security/audits.
- Admin / governance: K-Lend program upgradeable; control sits with a Squads multisig (associated K-Lend multisig 6hhBGCtmg7tPWUSgp3LG6X2rsmYWAc4tNsA6G4CnfQbM). Kamino docs list team multisigs (KLend authority A9rQoX1sictAQkyXxaZA8nz674xutHwoqpK2mwLyexCZ, Kamino MS E35i5qn7872eEmBt15e5VGhziUBzCTm43XCSWvDoQNNv, KFarms BbM3mbcLsa3QcYEVx8iovwfKaA1iZ6DK5fEbbtHwS3N8) plus an emergency pause. No on-chain timelock confirmed. Third-party rater Exponential reports fewer than 4 signers and no timelock (exponential.fi/protocols/kamino-lend). Verify yourself: look up the K-Lend program GzFgdRJXmawPhGeBsyRCDLx4jAKPsvbUqoqitzppkzkW on a Solana explorer and read its upgrade authority. KMNO on-chain governance voting planned 2026. Exact signer count and threshold: refresh.
- Oracle: Scope aggregator over Pyth and Switchboard with confidence and staleness checks; per-market oracle config in V2. Source: Kamino docs.
- Track record: roughly $16B cumulative loans issued, zero bad debt. 10 Oct 2025 (SOL down ~14%, roughly $20B market-wide cascade): about $20M of Kamino liquidations processed cleanly, zero bad debt, total supply down ~3.5% to ~$4.5B, borrow down ~2.6% to ~$1.6B. Feb 2025 stress also passed clean. Monthly risk reports plus a public risk dashboard and stress tests (gov.kamino.finance risk category). No protocol exploit found to date.
- KMNO / holder concentration: refresh (circulating supply and top holders not pulled on-chain this run; the fully-diluted-to-market-cap gap implies meaningful future unlocks).
Maintained monthly. Methodology: DeFi Research Instruction v2.