Are you an LLM? Read llms.txt for a summary of the docs, or llms-full.txt for the full context.
Skip to content

SparkLend

SparkLend is Spark's permissionless money market for supplying assets, borrowing against collateral, and managing lending positions. It has always been operated conservatively relative to peers, with a narrow collateral set, multi-oracle pricing, strict rate limits, and first-loss capital.

Guides

Use these guides for step-by-step SparkLend workflows:

Risk Architecture

SparkLend's risk pillars exist in isolation. They are designed to stack, so that a failure in any single component (oracle, issuer, liquidator, market liquidity) does not cascade into bad debt.

Restricted Collateral Set

SparkLend lists a deliberately short set of assets. ETH e-mode is limited to wstETH and rETH. BTC e-mode is being fully removed: the deprecation has been publicly noticed on the Sky forum and is tentatively targeted for the June 4 spell, with forced liquidation of remaining positions on June 8. Exposure in the affected buckets is already small (one material borrower at ~$1.6M plus a handful of smaller positions), and the removal is being done through a publicly telegraphed timeline rather than an immediate parameter flip.

Supply and Borrow Caps

Each listed asset has per-market supply and borrow caps configured by Sky Governance. Supply caps bound total exposure to any single asset, mitigating infinite-mint and oracle-manipulation risk on listed collateral. Borrow caps modulate how much of an asset can be drawn out, capping insolvency risk under stress. Caps integrate with the Cap Automator so they scale with on-chain market depth rather than waiting on each governance cycle.

Minimal Rehypothecation

Collateral supplied to SparkLend reserves stays in the reserve. It is not redeployed into external strategies.

Rate Limits

Every cross-module flow into and out of SparkLend is rate-limited at the smart-contract level: deposits, withdrawals, cross-chain bridging, and PSM swaps each have configured rate limits. On top of that, Spark's allocation system enforces per-market debt ceilings and inventory min/max bands. A single depositor or a single adverse event cannot drain the protocol in a single block, and rate limits cap the maximum capital at risk on any single route per unit of time.

Three-Oracle Median

Pricing is aggregated using a three-oracle median that pulls data from RedStone, Chainlink, and Chronicle. When all three return valid, non-stale values, the median is used. When two are valid, their average is used. A single-feed fallback also exists. The outcome is that a compromised or faulty single oracle does not drive SparkLend prices.

Killswitch Oracle for Pegged Assets

For collateral that trades against a hardcoded or exchange-rate price (wstETH, rETH, weETH, cbBTC, WBTC, LBTC), a peg-ratio oracle continuously compares the asset's market price to its underlying. When the deviation breaches a per-asset threshold, the killswitch halts new borrowing on SparkLend, preventing users from posting distressed collateral at a stale "face value" price and walking out with healthy debt.

Programmatic Liquidity Injection

SparkLend's liquidity buffer is not static. The Spark Liquidity Layer automatically moves USDS, USDC, and USDT into and out of SparkLend based on target borrow rates, utilization, and available inventory at other venues. If SparkLend utilization runs high, the SLL tops up idle liquidity to allow withdrawals and liquidations to clear. If a venue elsewhere offers better risk-adjusted yield, idle capital rotates there. This is what it means for Spark to be the largest depositor in its own market: liquidity responds to demand rather than being rationed by utilization alone.

USDC via USDS Routing

A periphery integration can offer USDC delivery from a USDS-denominated position by sourcing USDS and swapping it 1:1 through Sky's mainnet LitePSM wrapper. In that pattern, the debt or source asset remains USDS while the wallet receives USDC. Verify the active SparkLend frontend and periphery contracts before relying on this as a specific transaction path. See the LitePSM developer documentation for the underlying Sky contract mechanics.

Variable Liquidation Close Factor

If a position's Health Factor drops below 1 but stays above 0.95, up to 50% of the position's debt can be covered in a liquidation call. Below 0.95, up to 100% can be covered. The collateral transferred depends on the debt covered, oracle prices, and the applicable liquidation bonus. See Liquidations for the full mechanics.

Granular Borrowing Power Control

The borrowing power of any asset can be reduced down to 0% without impacting existing borrowers. This lets governance phase out a collateral type or tighten exposure on a live market without forced liquidations as the first response. Liquidating existing positions remains available where the situation requires it.

Where to Verify

ItemAddress or source
Live market parametersSpark App
Supply and borrow caps mechanismSupply & Borrow Caps
Oracle configurationOracles
LiquidationsLiquidations
Cap Automator auditsSparkLend Cap Automator
SparkLend auditsSparkLend Audits
Core contract addressesSpark Address Registry
Source codeSpark GitHub

Atlas Sources

The following Sky Atlas sections govern the controls described above:

SparkLend specific:
  • A.6.1.1.1.3.2.1 — SparkLend (parameters and operational processes)
  • A.6.1.1.1.3.2.1.1.1 — SparkLend Risk Parameters Definitions (LTV, Liquidation Threshold, Liquidation Bonus, Reserve Factor, Supply Cap, Borrow Cap, Interest Rate Model, Reserve State, E-Mode, Isolation Mode, Siloed Borrowing, Flash Loans)
  • A.6.1.1.1.3.2.1.1.2 — SparkLend Risk Parameters Current Configuration (per-asset values)
  • A.6.1.1.1.3.2.1.1.3 — SparkLend Cap Automators
  • A.6.1.1.1.2.6.1.1.2.1.1 — SparkLend Mainnet Instances Directory
  • A.6.1.1.1.2.6.1.3.1.1 — SparkLend Instance Configuration Documents
Cross-cutting:

FAQ

For step-by-step questions on supplying, borrowing, repaying, and liquidations, see the SparkLend Guides — in particular the Supplying Assets FAQ and Borrowing Assets FAQ.

On which networks is SparkLend supported?

SparkLend is currently supported on Ethereum and Gnosis Chain.

What is the cost of interacting with SparkLend?

Interacting with the protocol requires onchain transactions, so users pay the standard transaction fees of the network they are on. Fees depend on network conditions and transaction complexity.

What is the difference between transparent and variable rate?

Transparent rates act as a fixed rate in the short-term, but can change based on Sky Governance. The variable rate is set by supply and demand within SparkLend and changes continuously.

What is the health factor?

The health factor is the ratio between your collateralization ratio (the value of your collateral vs the value of your debt) and the liquidation loan-to-value. A higher health factor means your position is further from liquidation. If the health factor falls below 1, the position becomes liquidatable. See Liquidations for how to manage and avoid this.