Key points

  • Manual and Unified Account users can borrow USDC or USDT after supplying HYPE or BTC as collateral.
  • Hyperliquid sets a 65% loan-to-value ratio for HYPE and 50% for BTC, while interest rates vary with utilization.
  • The feature adds user-directed credit to HyperCore but exposes borrowers to price, interest, liquidity-cap and liquidation risks.

Hyperliquid has introduced manual borrowing, allowing eligible users to borrow USDC and USDT against HYPE and bitcoin supplied as collateral. The feature went live on September 18 and extends the platform's HyperCore market infrastructure beyond automatically managed portfolio-margin debt. It gives users in Manual or Standard accounts and Unified Accounts direct control over when they take and repay a loan.

Two collateral assets and two borrow assets

Hyperliquid's documentation lists HYPE and BTC as the supported collateral. HYPE carries a 65% loan-to-value ratio, while BTC carries a 50% ratio. A user's borrowing capacity is calculated from the amount supplied, the platform's oracle price and the applicable ratio. USDC and USDT can be borrowed or supplied to earn interest, but stablecoins supplied for interest do not increase manual borrowing capacity. Supplied HYPE and BTC do not earn interest.

Related reporting: Bybit launches 24/7 options tied to SpaceX and Nvidia perpetuals

Interest changes with utilization

Borrowed USDC and USDT accrue interest continuously, with the index updated hourly. Rates depend on how much of the available liquidity is in use, so the cost can change after a loan is opened. Hyperliquid says the borrowing rate is higher than the supply rate because interest paid by borrowers is distributed across a larger pool of suppliers. The protocol also retains 10% of borrowed interest as a buffer for future liquidations.

Manual accounts get the new control

The manual action is available to Manual or Standard and Unified Account users. Portfolio-margin accounts use automated borrowing and do not expose the same manual control. That distinction matters because the launch is not a separate peer-to-peer loan marketplace: it is a user-facing borrowing option built on the same underlying HyperCore infrastructure that supports portfolio margin. An independent report published Friday corroborated the launch and the supported asset set.

Liquidation thresholds exceed initial LTVs

Initial borrowing capacity is more conservative than the threshold used for partial liquidation. Hyperliquid lists an 82.5% partial-liquidation threshold for HYPE collateral and 75% for BTC. A collateral-price decline, another borrow, a collateral withdrawal or accumulated interest can push an account closer to that threshold. The interface's health factor measures LTV-weighted collateral value against outstanding debt, but Hyperliquid cautions that a reading below 100% does not itself define the liquidation trigger.

Capacity remains subject to liquidity and caps

Collateral alone does not guarantee that a requested loan can be filled. The documentation says borrowing is also constrained by available liquidity as well as account-level and global caps. Those limits can bind even when an account appears to have sufficient collateral. Users therefore have to consider both their own balance-sheet position and the market's supply conditions rather than treating the displayed collateral value as an unconditional credit line.

A broader credit layer for HyperCore

The launch broadens Hyperliquid's product stack by turning collateral already held on the platform into user-directed stablecoin liquidity. It may reduce the need to sell HYPE or BTC when users want quote assets, but borrowing introduces leverage and refinancing exposure. Because collateral values and utilization-based rates can move independently, the cost and safety margin of a position may deteriorate at the same time. The most important measures to watch over time are therefore borrow utilization, available liquidity, rate changes, collateral coverage and actual liquidations—not short-term token-price reactions to the announcement.

Sources

AI-generated editorial image; not a photograph of the reported event. Prepared with AI assistance and source verification.