DeFi & Web3Guide

Perpetual DEXs Explained: Hyperliquid, Aster & Lighter

Hyperliquid traded $7.14B of perpetual futures in the 24 hours to 16 November 2026, 61% of five on-chain venues measured, while Binance traded $38.2B. This guide compares how each venue matches and backs trades.

Perpetual DEXs Explained: Hyperliquid, Aster & Lighter
Illustration: CoinPaprika

Introduction

Hyperliquid traded $7.14B of perpetual futures in a 24-hour window on 16 November 2026, about 19% of Binance's USDT futures volume, and held $11.1B of open interest. Perp DEXs have moved from curiosity to core decentralized finance (DeFi) infrastructure ↗, yet one trader withdrawing margin from a 50x Ether position still cost Hyperliquid's community vault about $4M in March 2025. That scale hides three different designs, an on-chain order book, a zero-knowledge rollup and oracle-priced pools. This guide measures five venues from their public APIs, explains who takes the other side of each trade, and sets out the documented failures. Every figure is a dated 24-hour snapshot, because no free source offers a verified 30-day series. It closes with a method for comparing any perp DEX before depositing.

Key Takeaways

  • Hyperliquid carried 61% of the $11.67B that five measured perp DEXs traded in 24 hours, while Binance's futures market alone traded $38.2B.
  • Three designs exist: an on-chain order book (Hyperliquid), off-chain matching proven on Ethereum (Lighter), and oracle pricing against a pool (GMX, Jupiter).
  • Vaults take the other side of trades: HLP holds $151.8M, JLP holds $676.0M, and both absorbed losses in documented incidents.
  • HIP-3 lets a deployer who stakes 500,000 HYPE list markets, and one deployer's dex carries 98% of HIP-3's $2.43B daily volume.
  • Lighter turned over 5.82 times its open interest in a day against 0.65 on Hyperliquid, a gap that volume data alone cannot explain.

What Is a Perpetual DEX, and How Does It Differ From a Centralized Perp?

A perp DEX settles perpetual futures from a self-custodial wallet instead of an exchange account. On Bitcoin, Hyperliquid, the largest on-chain venue measured, traded $2.75B in 24 hours against Binance's $9.46B, yet held $2.36B of open interest to Binance's $6.03B (venue APIs, 2026). Volume trails; positions do not.

A Perpetual With No Expiry

A perpetual future, or perp, is a leveraged contract on an asset's price with no expiry date. Perp DEX, perpetual DEX, decentralized perpetuals and on-chain perps all name the same product class. A trader posts margin as collateral and holds exposure many times larger. A periodic funding payment between longs and shorts keeps the contract price near the spot price. Funding, leverage and liquidation formulas belong to dedicated trading-mechanics coverage; this guide compares the venues. Leverage limits still shape every venue's risk: Hyperliquid caps Bitcoin at 40x and Ether at 25x, and at 40x the maintenance margin is 1.25% of the position, so a 1.25% move against a trader triggers liquidation (Hyperliquid docs, 2026).

Perp DEX Versus Spot DEX

A spot decentralized exchange (DEX) swaps one token for another, and the buyer ends up owning the token. A perp DEX trades a contract on price, so no token changes hands; the position is an entry in a margin account funded with stablecoin collateral, which on Hyperliquid is USDC deposited through its bridge. Spot volume is much larger and concentrated in automated market makers: DefiLlama's Dexs category recorded $213.3B over the 30 days to 15 November 2026 (DefiLlama, 2026). The breakdown of spot DEX market share by chain ↗ covers that side. Hyperliquid runs a spot order book as well, which traded $4.6B in the same 30 days (DefiLlama, 2026).

Perp DEX Versus Centralized Perp

A centralized exchange (CEX) perp is a claim on the exchange: it holds the collateral and matches orders on private servers. On a perp DEX, balances, positions and liquidations sit on a public ledger or are proven against one. Binance remains far larger. Its USDT-margined futures traded $38.2B across 758 symbols in the 24 hours to the snapshot, about 5.3 times Hyperliquid's $7.14B across 350 markets (venue APIs, 2026). The trade is depth for verifiability and self-custody ↗. Neither model removes operator power: a CEX can freeze accounts, and Hyperliquid's validators voted in March 2025 to delist a market and settle open positions.

VenueType24h volumeOpen interestVolume / open interest
Binance USDT-MCentralized$9.46B$6.03B1.57
HyperliquidOn-chain order book$2.75B$2.36B1.16
LighterZK rollup$1.45B$0.12B12.08
AsterPerp DEX, BNB Chain origin$0.42B$0.39B1.08
Jupiter PerpsOracle and pool$34MNot publishedn/a

Data current as of November 2026.

Six stat cards: USD 7.14B Hyperliquid volume, USD 11.1B Hyperliquid open interest, USD 38.2B Binance volume, USD 2.45B Lighter, USD 1.79B Aster, USD 35M GMX

Volume rankings hide the larger difference, which is architecture.

What Are the Three Ways to Build an On-Chain Perpetual Exchange?

Every on-chain perp exchange answers three questions: where orders match, where the price comes from, and who takes the other side. Hyperliquid runs a full order book on its own chain, Lighter proves off-chain matching on Ethereum, and GMX and Jupiter price trades from oracles against a pool (venue docs, 2026).

On-Chain Order Book: Hyperliquid

Hyperliquid runs its own layer-1 blockchain, secured by HyperBFT, a variant of HotStuff consensus. HyperCore, its execution layer, holds margin and matching-engine state on-chain with no off-chain order book (Hyperliquid docs, 2026). Hyperliquid reports median end-to-end order latency of 0.2 seconds and capacity near 200,000 orders per second. A second layer, HyperEVM, hosts general smart contracts beside HyperCore. Liquidations use a mark price that combines external exchange prices with Hyperliquid's own book state, so the venue still leans on centralized price feeds (Hyperliquid docs, 2026). The design's cost is a small validator set that doubles as the bridge's signers, covered in the custody section below.

Off-Chain Matching and On-Chain Proofs: Lighter

Lighter is a zero-knowledge (ZK) rollup on Ethereum. A sequencer orders transactions off-chain, a prover generates proofs that matching and liquidations followed the rules, and Ethereum contracts hold deposits and the canonical state root (Lighter docs, 2026). Lighter's whitepaper argues that proving the matching itself limits an operator's ability to favor some orders over others. Posted data lets users rebuild their own state, and an escape-hatch mode covers an operator failure. Standard accounts pay no maker or taker fee (Lighter API, 2026). Lighter also runs a second deployment on Robinhood Chain, which DefiLlama lists with $112M deposited (DefiLlama, 2026). edgeX, another perp venue, served a web page instead of data on every public endpoint tried, so it is excluded from the comparison.

Oracle Prices Against a Pool: GMX and Jupiter

GMX and Jupiter Perps have no order book. GMX takes prices from Chainlink Data Streams oracles and fills orders against GM and GLV liquidity pools (GMX docs, 2026). Jupiter prices SOL, ETH and wBTC positions with on-chain oracles, and traders borrow from the Jupiter Liquidity Provider (JLP) pool to open leveraged positions (Jupiter docs, 2026). Trades execute at the oracle price with no order-book slippage, and a price-impact fee protects the pool. Jupiter also runs six Beta markets on a separate order-book engine, so one venue now mixes two designs. Jupiter allows up to 250x on its pool markets and GMX up to 100x, both far above Hyperliquid's 40x on Bitcoin.

DesignExampleMatchingPrice sourceCounterpartyMain risk
On-chain order bookHyperliquidOn-chain, by validatorsBook plus exchange-fed oracleTraders and HLPSmall validator set
Proven off-chainLighterOff-chain sequencerOrder bookOther tradersCentralized sequencer
Oracle and poolGMX, JupiterNone; fills at oracle priceOracle feedLiquidity poolTrader profits drain the pool

Data current as of November 2026.

The counterparty column explains why the vaults matter as much as the matching engines.

Who Takes the Other Side of a Trade on a Perp DEX?

On a perp DEX, the counterparty can be a vault. The Hyperliquidity Provider (HLP) vault held $151.8M on Hyperliquid, and Jupiter's JLP pool held $676.0M (venue APIs, 2026). Depositors in both earn fees and absorb the losses a venue cannot close out.

HLP: Market Making and Liquidations

HLP is a community-owned vault that runs market-making strategies, performs liquidations and accrues platform fees (Hyperliquid API, 2026). Depositors share its profit and loss, and its all-time profit stands at $140.9M. One component, the liquidator vault, takes over positions once account equity falls below two-thirds of maintenance margin, and it keeps the maintenance margin as a buffer (Hyperliquid docs, 2026). Hyperliquid itself warned after the March 2025 loss that HLP "is not a risk-free strategy" (Hyperliquid, 2025). DefiLlama counted 99% of Hyperliquid's perp fees ↗ as flowing to its Assistance Fund, which converts them into HYPE; Hyperliquid's documentation says the HYPE is burned, while the fund's address still held 49.0M HYPE on 16 November 2026 (DefiLlama, 2026; Hyperliquid API, 2026).

Pool Vaults: GLV and JLP

GMX splits liquidity into GM pools, one per market, and GLV vaults that spread deposits across several GM pools. Liquidity providers earn 63% of fees on Arbitrum and Avalanche (GMX docs, 2026). Jupiter's JLP pool earns 75% of fees from its SOL, ETH and wBTC markets and holds the assets themselves: 46.9% SOL, 30.3% USDC, 12.9% wBTC, 7.7% ETH and 2.2% JupUSD (Jupiter API, 2026). JLP's value moves with those prices and with trader profit, because winning traders are paid from the pool. A pool depositor sits opposite the average trader.

Both models share one weakness: the vault absorbs whatever the engine cannot close.

How Do Hyperliquid, Lighter, Aster, GMX and Jupiter Compare on 16 November 2026?

Hyperliquid carried 61% of the 24-hour volume of the five perp DEX venues measured here on 16 November 2026: $7.14B of $11.67B. That is a share of this sample, not of the perp DEX market, which excludes venues such as edgeX (venue APIs, 2026).

What the Public APIs Measured

Every figure comes from a keyless public API, read between 09:50 and 09:52 UTC as a rolling 24-hour window. Hyperliquid's total adds the default venue ($4.71B, 179 markets) to 171 builder-deployed markets ($2.43B). Open interest is the open-interest field multiplied by mark price. Lighter's own dashboard may define it differently. GMX volume is executed position size on Arbitrum from GMX's public indexer, and GMX open interest is the average of its long and short sides. Jupiter's figure covers the three pool-backed markets only. Aster's count is 595 trading contracts, 563 of them margined in USDT and 27 in USD1, and Lighter's is 208 active of 243 records (venue APIs, 2026).

What Could Not Be Measured

edgeX's three public endpoints returned the exchange's web page instead of data, so it is absent. Jupiter's endpoint exposes no long open interest, only $8.6M of short positions, and its six Beta markets and GMX's other chains were not queried. No 30-day figure appears here: DefiLlama's derivatives volume endpoints sit behind a paid plan, and the 24-hour snapshots are not extrapolated to a month. Aster's 20x comes from the 5% required margin in its public exchange metadata; its March 2025 announcement advertised up to 1001x in a Simple Mode, which this pull did not test (Chainwire, 2025). Aster's and Lighter's numbers come from the venues' own APIs, which no independent party audited for this article.

VenueMarkets24h volumeOpen interestMax leverageBase feeToken market cap
Hyperliquid350$7.14B$11.05B40x (BTC)0.045% takerHYPE $15.9B
Lighter208$2.45B$0.42B50x (BTC, ETH)0% standard accountLIT $0.69B
Aster595$1.79B$1.15B20x (BTC, ETH)Not measuredASTER $0.91B
Jupiter Perps3 of 9$0.26BNot published250x0.06% plus borrowNot measured
GMX147$0.04B$0.02B per side100x0.04% to 0.06%GMX $0.07B

Data current as of November 2026.

Horizontal bar chart of 24-hour volume: Hyperliquid USD 7.14B, Lighter USD 2.45B, Aster USD 1.79B, Jupiter USD 0.26B, GMX USD 0.04B

Volume is the easiest column to inflate, which makes the deposits behind it worth checking next.

What Backs the Deposits: Bridges, Validators and Rollup Contracts?

Funds on a perp DEX sit in a bridge or a contract, and the trust assumption differs by design. DefiLlama counts $5.81B in Hyperliquid's bridge, $0.55B in Lighter's and $0.59B in Aster's (DefiLlama, 2026). Bridge value measures custody of deposited USDC, not protocol total value locked (TVL).

Hyperliquid's Bridge and Validators

Deposits on Arbitrum are credited once more than two-thirds of validator staking power signs. A withdrawal needs the same two-thirds, then a dispute period during which the bridge can be locked; only cold-wallet signatures from two-thirds of stake-weighted validators unlock it again (Hyperliquid docs, 2026). A withdrawal costs 1 USDC, and Zellic audited the bridge. Hyperliquid's validatorSummaries call listed 35 validators on 16 November 2026, 27 of them flagged active (Hyperliquid API, 2026). A May 2026 Bitwise prospectus describes the active set as the top 24 validators by stake weight (SEC, 2026), so the signer count depends on the definition used. The trust assumption is two-thirds of a stake-weighted set of roughly two dozen signers.

Lighter, Aster and Jupiter Custody

Lighter's deposits sit in Ethereum contracts that update only when a proof verifies, with an escape hatch if the operator stops (Lighter whitepaper, 2025). Jupiter's JLP pool holds its assets on Solana, which the venue's API valued at $676.0M, against $0.57B for DefiLlama's Jupiter Perpetual Exchange listing; the two measure differently (Jupiter API, 2026). Aster lists bridges across BNB Chain, Ethereum and Arbitrum. In October 2025 DefiLlama's founder said Aster provided neither node software nor order-level data that would let outsiders verify its trades (DL News, 2025). No equivalent public verification tool was located for this article.

Horizontal bar chart of DefiLlama deposits: Hyperliquid USD 5.81B, Aster USD 0.59B, Jupiter USD 0.57B, Lighter USD 0.55B, GMX USD 0.19B, edgeX USD 0.06B

A bridge secures deposits; it does not decide which markets exist, and that decision carries its own risk.

How Do New Perp Markets Get Listed, and What Is HIP-3?

Hyperliquid lets outside builders list perp markets under HIP-3, a step toward decentralizing perp listing. A deployer stakes 500,000 HYPE, about $33M at the 16 November 2026 price of $66.94, then defines the oracle, contract specification and leverage limits for its own markets (Hyperliquid docs, 2026).

HIP-3 Rules

The stake must stay in place for at least 183 days after deployment. Each deployer runs one perp dex with independent margining and order books, and the first three assets need no auction; later assets go through a shared Dutch auction. The deployer can halt trading and settle positions at the mark price. It can also take a fee share of 0% to 300%, capped at 100% in growth mode, and a share above 100% raises the protocol fee to match (Hyperliquid docs, 2026). Quote assets that fail permissionless requirements lose their status by validator vote, which disables dexes built on them.

What HIP-3 Added

Hyperliquid's perpDexs call listed ten dexes on 16 November 2026, and four returned live markets: xyz with 122, para with 33, io with 15 and mkts with 1 (Hyperliquid API, 2026). xyz alone traded $2.39B, 98% of the $2.43B HIP-3 volume, with $4.14B of open interest, in tickers such as Apple, gold, Brent crude, the euro and the yen. The other six dexes returned no live markets in the pull. HIP-3 turned Hyperliquid into a venue for equities, commodities and currencies as well as crypto, with the deployer, not the core team, responsible for each oracle.

Listing freedom widens the attack surface that the incident record below documents.

What Went Wrong on Perp DEXs, and What Did It Teach?

Perp DEX incidents cluster around two targets: the vault that backstops liquidations and the keys that control the protocol ↗. None of the six episodes below was a matching-engine failure.

Liquidation Backstop Attacks

On 12 March 2025, a trader opened a 50x Ether long, withdrew margin and was liquidated with a profit of about $1.8M; HLP lost about $4M, and Hyperliquid cut maximum leverage to 40x for Bitcoin and 25x for Ether (CoinDesk, 2025). Hyperliquid said no exploit occurred. On 26 March 2025, a JELLY short liquidated into HLP while the token was pumped, and the vault's loss reached about $13.5M. Validators voted to delist the market and settled it at $0.0095, against roughly $0.50 on DEX-fed oracles (CoinDesk, 2025). Flagged addresses were excluded from compensation, other users were to be made whole by the Hyper Foundation, and HYPE fell as much as 22% (The Block, 2025). In April 2026, PeckShield put HLP's loss from a FARTCOIN self-liquidation at about $1.5M, while another analyst estimated $1.2M (Gate News, 2026).

The October 2025 Cascade

The 10 October 2025 liquidation cascade ↗ erased $19B across crypto in 24 hours, and more than 1,000 Hyperliquid wallets were fully liquidated (CoinDesk, 2025). Auto-deleveraging (ADL), the last-resort step that closes profitable positions to cover losses, ran at least 40 times and closed about $2.1B of positions in 12 minutes across 19,337 wallets (Chitra et al., arXiv, 2026). The authors estimate the production queue used roughly 28 times more ADL than their optimal policy needed. Hyperliquid's open interest then fell nearly 50% while Lighter's and Binance's recovered (Chitra et al., arXiv, 2026). The paper notes that numerous parties attribute the decline to Hyperliquid's ADL policy.

Key Compromise on Drift

On 1 April 2026, attackers drained about $285M from Drift, a Solana perp DEX (Chainalysis, TRM and Drift count $285M drained; DefiLlama records $295M of outstanding losses). According to Drift's post-mortem, Security Council members pre-signed transactions through Solana's durable-nonce feature, which handed over admin control. The council had been set up as a two-of-five multisig with no timelock on 27 March 2026, five days earlier. A fake token was then accepted as collateral, and JLP tokens worth $159.3M were the largest asset taken, out of a loss above half the protocol's deposits (Chainalysis, 2026). No smart-contract bug was involved, and Chainalysis calls a North Korean link likely but formal attribution pending.

Mar 2025
ETH whale liquidation
A 50x long withdrew margin and was liquidated; HLP lost about $4M; leverage caps cut to 40x BTC and 25x ETH.
Mar 2025
JELLY manipulation
A short was pushed onto HLP and the token pumped; loss reached about $13.5M; validators delisted at $0.0095.
Oct 2025
Aster volume dispute
DefiLlama delisted Aster perp volume after it mirrored Binance almost one-to-one; wash trading was not proven.
Oct 2025
October 10 cascade
Auto-deleveraging closed about $2.1B of positions in 12 minutes; Hyperliquid open interest fell nearly 50% afterwards.
Apr 2026
Drift takeover
Pre-signed admin transactions drained about $285M by Chainalysis's count ($295M per DefiLlama) from a Solana perp DEX; no smart-contract bug was involved.
Apr 2026
FARTCOIN self-liquidation
Four wallets built a $15M long and let it liquidate; HLP lost $1.2M to $1.5M by two estimates.

Data current as of November 2026.

Each failure traced to a vault, a key or a price feed, which gives a reader specific things to check.

How Can a Reader Judge a Perp DEX Before Depositing?

Open interest costs margin and funding to hold, which makes it harder to fake than volume. Measured on 16 November 2026, Lighter turned over 5.82 times its open interest in 24 hours, GMX 2.00, Aster 1.55 and Hyperliquid 0.65 (venue APIs, 2026).

Open Interest Against Volume

High turnover has innocent explanations. Zero-fee accounts, market makers and short-term traders all raise volume without raising open interest, and Lighter's standard accounts pay nothing to trade. Binance, the centralized benchmark, turned over 1.57 times on Bitcoin, Hyperliquid 1.16 and Lighter 12.08. The same pattern also fits inflated volume, and the data cannot separate the two. Aster formed when the yield protocol Astherus merged with the BNB Chain perp exchange APX Finance, and YZi Labs, formerly Binance Labs, backs it (Chainwire, 2025). DefiLlama delisted its perp volume on 5 October 2025 after it tracked Binance almost one-to-one on XRPUSDT and ETHUSDT. DefiLlama's founder said wash trading was not proven (The Block, 2025). Cointelegraph quoted him saying DefiLlama had earlier delisted Lighter and other perp venues over what he called blatant wash trading (Cointelegraph, 2025).

Checks Before Depositing

Five checks cover the incident record. A reader should compare volume with open interest, as above. The reader should name the counterparty and the vault's size, since HLP's $151.8M and JLP's $676.0M absorb the losses. The custody mechanism comes next: bridge signers, proofs or a multisig. Thin tokens are the attack surface in the JELLY and FARTCOIN episodes, so the listing rules for the specific market matter. The last check is the leverage cap and fee for that market, which differ across a single venue: Hyperliquid allows 40x on Bitcoin but 10x on HYPE, and charges a 0.045% base taker fee, against 0.06% plus hourly borrowing on Jupiter and 0.04% or 0.06% on GMX.

Bar chart of 24-hour volume divided by open interest: Lighter 5.82, GMX 2.00, Aster 1.55, Hyperliquid 0.65

Conclusion

A perp DEX is a perpetual future whose collateral, positions and liquidations live on a public ledger or are proven against one. The five venues measured here solve the same problem in three ways, and the choice decides who matches orders, who prices them and who absorbs losses. Scale follows liquidity rather than design: Hyperliquid's order book leads, while pool-based GMX trades a fraction of its volume, and centralized exchanges remain several times larger.

The difference between a sound venue and a fragile one is the backstop. Every documented episode, from the March 2025 liquidations to the Drift takeover, hit a vault, a key or a price feed while the matching engine kept running. Those weaknesses stay invisible while volume is high and markets are liquid. They surface within minutes once a thin token or a signer is targeted, too fast for a depositor to leave.

Before depositing, you should read volume next to open interest, find out who the counterparty is and how large that vault is, and identify whether a bridge, a proof or a multisig holds the collateral. Check the leverage cap and the listing rules for the specific market you plan to trade. A venue that answers all four plainly deserves a small test deposit first.

Quick Stats

  • $7.14B — Hyperliquid 24-hour perp volume, core and HIP-3 markets, 16 November 2026
  • $11.05B — Hyperliquid open interest across 350 markets, the largest of five venues measured
  • $151.8M — HLP vault account value, against $676.0M in Jupiter's JLP pool
  • 5.82x — Lighter's daily volume per dollar of open interest, against 0.65x on Hyperliquid
  • $285M — drained from Drift, a Solana perp DEX, on 1 April 2026, by Chainalysis's count (DefiLlama records $295M)

Data current as of November 2026.

FAQ

?Can a trader lose more than the margin posted on a perp DEX?

On Hyperliquid, a position is liquidated once account equity falls below maintenance margin, so the loss is limited to the equity deposited in the account. The remaining risks run the other way. Auto-deleveraging can close a winning position early, as it did across about $2.1B of positions in 12 minutes on 10 October 2025, and a backstop liquidation keeps the maintenance margin instead of returning it (Hyperliquid docs, 2026).

?Does a trader need to hold HYPE to trade on Hyperliquid?

No. Collateral is USDC, deposited through the Arbitrum bridge, so a trader can open positions without owning HYPE. Staking HYPE earns a trading-fee discount, one of several reductions that sit alongside referral rewards and volume tiers. The base perp fee is 0.045% for takers and 0.015% for makers, and both fall with 14-day volume (Hyperliquid docs, 2026).

?Is HLP the same thing as Hyperliquid's Assistance Fund?

No. HLP is a vault that depositors fund; it makes markets, performs liquidations and accrues fees for its depositors. The Assistance Fund is a separate pot that, per DefiLlama, receives 99% of perp fees and converts them into HYPE. Hyperliquid's documentation says that HYPE is burned, while the fund's address held 49.0M HYPE on 16 November 2026 (Hyperliquid API, 2026).

?How does a withdrawal from Hyperliquid reach Arbitrum, and what does it cost?

The balance leaves the Hyperliquid account immediately, and validators sign the withdrawal as a separate transaction. Once two-thirds of the staking power has signed, a transaction on Arbitrum requests the payout, and a dispute period follows during which the bridge can be locked. A flat 1 USDC fee covers the validators' Arbitrum gas, so the trader needs no ETH. The docs reviewed do not state the dispute period's length (Hyperliquid docs, 2026).

?Is holding JLP the same as holding SOL, ETH and wBTC?

No. JLP is a share of a pool holding 46.9% SOL, 30.3% USDC, 12.9% wBTC, 7.7% ETH and 2.2% JupUSD, and the pool pays out when traders profit. It earns 75% of the fees from its markets, which Jupiter's API reported as a 15.48% annualized return, last updated on 14 November 2026. In the Drift takeover, JLP tokens worth $159.3M were the largest asset drained (Jupiter API, 2026).

?Where can a reader check a perp DEX's volume and open interest without trusting a dashboard?

Each venue publishes raw numbers: Hyperliquid's info API (metaAndAssetCtxs), Lighter's orderBookDetails, Aster's ticker and open-interest endpoints, Jupiter's perps API and GMX's markets endpoint. Dividing volume by open interest, and comparing a venue's pattern with Binance's on the same pair, is the test DefiLlama applied to Aster in October 2025 (The Block, 2025).

?How does trading perps on Jupiter differ from swapping through Jupiter's aggregator?

The aggregator routes spot swaps across many exchanges and hands over tokens. Jupiter Perps lets a trader borrow from the JLP pool to open a leveraged long or short on SOL, ETH or wBTC at an oracle price, with a 0.06% base fee, hourly borrow fees and leverage up to 250x. The companion guide to DEX aggregators ↗ covers the swap side (Jupiter docs, 2026).

References / Sources

Platform & Company Data

Official venue APIs, documentation and announcements.

  • Hyperliquid: Info API Market, Vault and Validator Data (api.hyperliquid.xyz, Nov 2026)
  • Hyperliquid Docs: Bridge, HIP-3, Fees, Liquidations (hyperliquid.gitbook.io, Oct 2026)
  • Lighter: Docs and Order Book API (docs.lighter.xyz, Nov 2026)
  • Jupiter: Perps Docs and JLP Pool API (docs.jup.ag, Nov 2026)
  • GMX: Trading, Fees and Liquidity Docs (docs.gmx.io, Oct 2026)
  • Chainwire: Astherus Rebrands to Aster (chainwire.org, Mar 2025)
Market Research

Aggregated market data and news reporting on incidents.

  • DefiLlama: Protocol TVL and Fee Data (api.llama.fi, Nov 2026)
  • CoinPaprika: Token Market Capitalizations (api.coinpaprika.com, Nov 2026)
  • Binance: USDT-Margined Futures 24-Hour Data (fapi.binance.com, Nov 2026)
  • CoinDesk: Hyperliquid ETH Whale and JELLY Reports (coindesk.com, Oct 2025) (two editions cited)
  • The Block: Aster Volume Delisting and JELLY Coverage (theblock.co, Oct 2025)
  • Cointelegraph: Aster Delisting and DeFi Data Integrity (cointelegraph.com, Oct 2025)
Academic & Technical

Research papers, protocol whitepapers, filings and security analyses.

  • Chitra et al.: Autodeleveraging, Impossibilities and Optimization (arxiv.org, 2026)
  • Lighter: Verifiable Order Book Matching Whitepaper (assets.lighter.xyz, 2025)
  • Chainalysis: Drift Protocol Hack Analysis (chainalysis.com, Apr 2026)
  • SEC: Bitwise Hyperliquid Trust Prospectus (sec.gov, May 2026)
  • DL News: Aster Perp Volume Scrutinized (dlnews.com, Oct 2025)
Share this guide
Written by Bartek Hagan

Original CoinPaprika guide.

Report a correction
All guides

Cryptocurrencies are highly volatile and involve significant risk. You may lose part or all of your investment.

All information on Coinpaprika is provided for informational purposes only and does not constitute financial or investment advice. Always conduct your own research (DYOR) and consult a qualified financial advisor before making investment decisions.

Coinpaprika is not liable for any losses resulting from the use of this information.