Liquidation Bots: Hyperliquid Perps vs Aave on Ethereum Compared

Running a liquidation bot in 2025 means choosing your arena carefully. The gap between liquidation bots on Hyperliquid perps vs Aave on Ethereum is not just a gas-cost footnote — the two venues differ in liquidation mechanics, margin of safety thresholds, competition density, capital requirements, and the infrastructure stack needed to stay profitable. This piece walks through both systems with the specificity you need to make the call.

How Liquidation Mechanics Differ at the Protocol Level

Aave's liquidation model is reactive and oracle-driven. A position becomes eligible when its health factor drops below 1.0, where health factor = (collateral value × liquidation threshold) / total debt. When that threshold is crossed, any external address can call liquidationCall(), repay up to 50% of the outstanding debt, and receive the equivalent collateral plus a liquidation bonus — typically 5–15% depending on the asset.

The oracle dependency is the key design choice. Aave v3 uses Chainlink price feeds with a heartbeat of roughly 1 hour (or 0.5% deviation triggers), which means there is a structural lag between market price and protocol-recognized price. Liquidators that watch CEX prices in real time and model the oracle state can identify eligible positions before the oracle ticks — that latency edge matters on volatile assets.

Hyperliquid's liquidation engine is endogenous: the exchange itself enforces margin and triggers liquidations without an external oracle call cycle. Positions are marked to a median price derived from Hyperliquid's own order book, updated continuously. When a cross-margin account's equity falls below the maintenance margin, the liquidation engine — running as part of the validator set — begins closing positions. The protocol does not expose an open liquidationCall() to arbitrary searchers the same way Aave does.

This is the fundamental difference. On Aave, you are a liquidator competing with other searchers to be the first to call the contract. On Hyperliquid, you are not liquidating in the same sense at all — you are trading against the liquidation engine's backstop mechanism, positioning to absorb the other side of forced closes or benefit from the Insurance Fund taking losses.

Capital Requirements and Collateral Structure

On Aave, a liquidation requires real upfront capital: you repay a portion of the borrower's debt, so you need the debt token in your wallet (or a flash loan). Flash loans are the standard approach for capital-efficient liquidation, borrowing the repayment amount, calling liquidationCall(), receiving the collateral, selling it, and repaying in a single atomic transaction. The practical floor for profitability after Flashbots MEV-share fees and gas is a position with at least ~$5–10k in eligible debt — smaller than that and the gas cost eats the bonus even with a flash loan.

On Hyperliquid, the relevant capital is your margin to open a perp position. There is no liquidationCall() to fund. What you are doing is entering a directional trade or limit order that benefits from the forced close of a large underwater account. The capital requirement scales with the position size you want to take, subject to normal perp margin rules. $500–$2k can be meaningful depending on the asset and leverage structure, where equivalent Aave exposure would need ten times more.

Gas Costs vs Execution Fees

This is where Ethereum's structural overhead is most painful. A successful Aave liquidation on mainnet costs roughly 250,000–400,000 gas units. At 15 gwei base fee, that is ~$3–6 in execution cost at current ETH prices — before any priority fee or Flashbots bundle tip. On a $100k position with a 5% bonus, the math works easily. On a $20k position, the margin erodes fast. On a $5k position, you need very low base fees or you are working for nothing.

Hyperliquid charges 0.025% taker / 0.002% maker fees for most perp pairs. There is no gas cost in the traditional sense — it is a flat fee schedule on notional. For a $10k position that is $2.50 in taker fees. The comparison in per-trade friction heavily favors Hyperliquid, though the liquidation bonus structure is different.

Competition Density and MEV Infrastructure

Aave liquidations on Ethereum mainnet are among the most competitive MEV opportunities that still exist. The top liquidation bots submit private Flashbots bundles, operate dedicated nodes for fast mempool access, and have years of optimization behind them. The realistic new entrant path is either finding tail assets with lower searcher coverage or operating on Aave deployments on L2s — Base and Arbitrum still have thinner competition than mainnet.

Hyperliquid's relevant competition is informed directional traders and market makers. Because there is no direct liquidation call, what you are competing against is other traders with good position monitoring who also want to trade against or alongside forced closes. This is a softer barrier to entry for a well-equipped shop, but it also means the edge is harder to quantify: you need a view on price, not just a speed race.

Our Hyperliquid Liquidation Hunter service targets exactly this — monitoring open interest, per-account margin levels, and liquidation price proximity to position ahead of forced closes with tight latency.

Infrastructure Stack Comparison

An Aave bot needs:

• A dedicated Ethereum node or fast RPC with mempool access
• Flashbots relay integration for private bundle submission
• Smart contract to execute flash loan + liquidation atomically
• An oracle simulation layer to predict health factor before the Chainlink feed ticks
• Decent capital buffer or flash loan logic for large positions

A Hyperliquid liquidation-oriented bot needs:

• Websocket connection to Hyperliquid's L1 data feed (sub-100ms is achievable)
• Real-time margin and liquidation-price tracking across open accounts
• Order management connected to the native API (EIP-712 signed orders)
• Position sizing and risk management calibrated to expected forced-close impact

The Hyperliquid stack is meaningfully simpler to operate at low latency. There is no mempool game, no gas auction, no flash loan contract to maintain. The trade-off is that the alpha is softer — you are trading a market effect, not atomically capturing a bonus.

Where to Deploy in 2025

If you have existing MEV infrastructure, Ethereum node access, and Flashbots experience, Aave liquidations on Base or Arbitrum are the more defensible play — lower competition than mainnet, lower gas, and the bonus structure is clear. Expect to deploy $50–200k in working capital or flash loan access to cover the tail position sizes worth chasing.

If you are starting fresh or want a faster iteration cycle, Hyperliquid is the better environment. Lower infrastructure cost, no gas auctions, a simpler execution model, and the perp market is deep enough that the forced-close events on major assets produce real tradeable moves. The edge is in monitoring quality and execution latency, not searcher infrastructure arms races.

The /services catalog covers both angles — EVM MEV bots for Aave-style liquidations and dedicated Hyperliquid bots for perp-native strategies. Both verticals have different risk profiles and different infrastructure moats; neither is obviously dominant.

---

If you want a liquidation bot built to production standard — with proper margin monitoring, latency-tuned execution, and risk controls from day one — get in touch and we'll scope it out.