By spring 2026, Hyperliquid had become the de facto standard in the perp-DEX segment. According to CoinGecko, in April, the platform recorded trading volumes of $190.28 billion (about 3.9% of the total volume), ranking ninth among all perpetual futures platforms, including CEX.
However, the infrastructural issues of on-chain derivatives remain unresolved. Manipulations involving meme coins and the unstable operation of mark prices during periods of high load continue to be vulnerabilities that competitors are addressing with their own architectural solutions.
On May 18, 2026, AFX joined this race—a Layer 1 network designed for trading perpetual futures. Together with the project team, we explore how AFX operates and how its approach differs from Hyperliquid.
Hyperliquid Sets the Standard
Hyperliquid has integrated a full on-chain order book with a user experience familiar to traders. In October 2025, the protocol activated the HIP-3 update, allowing any developer to launch a perpetual futures market based on HyperCore by staking 500,000 HYPE. The market creator sets the contract specifications, oracles, leverage limits, and settlement parameters, with a commission share that can reach 50%.
This mechanism quickly propelled Hyperliquid beyond the crypto-native perimeter. The platform became a 24/7 trading venue for macro risks: during the escalation of the Middle Eastern conflict, the volume for the CL-USDC contract linked to WTI reached $1.1 billion, with an open interest exceeding $274 million.
The segment leader has already undergone several stress tests. In March 2025, Hyperliquid faced manipulations involving the meme token JELLYJELLY: attackers exploited the liquidation mechanism and the asset's low liquidity, resulting in unrealized losses of approximately $12 million for the Hyperliquidity Provider Vault (HLP). The exchange team had to manually halt trading and delist the contract.
A new incident occurred in November of the same year involving the meme coin POPCAT. According to on-chain analysts, the attacker withdrew about $3 million in USDC from OKX, distributed the funds across 19 wallets, and opened a total long position of approximately $30 million. When the trader removed the "wall" of buy orders, their position of $20-30 million was liquidated in seconds. The losses fell to the exchange's liquidity provider (HLP), which lost $4.9 million. Amid suspicions of manipulation, Hyperliquid temporarily suspended deposits and withdrawals.
These incidents did not break the business model but highlighted specific scenarios where Hyperliquid behaved differently than stated in its documentation. They also served as a reference point for the next generation of projects.
L1 for Order Books
AFX positions itself as a Sovereign Trading Layer—a dedicated infrastructure for professional derivatives trading. The team abandoned the universal blockchain model and built its own stack: a custom execution layer, the Mysticeti consensus based on DAG BFT, and a modular structure using ABCI/Cosmos SDK. Execution and consensus are separated, allowing order flow to be processed independently from other network processes.
Hyperliquid also operates on its own L1 and builds infrastructure around HyperCore—an on-chain engine for spot and perpetual markets with an order book and order matching based on price-time priority. Through HyperEVM, developers can connect to HyperCore's liquidity and create applications within the ecosystem.
AFX took a different approach: instead of an ecosystem-wide horizontal focus, it concentrated on the vertical trading stack—blockchain validators, mempool, DAG consensus, ABCI communication layer, virtual machine, account module, bridge, and trading engine are packaged into a single system where the order matching logic is physically separated from consensus and cannot be delayed by it.
The exchange added two more elements aimed at professional flow:
- Zero gas fees eliminate friction when placing and canceling orders, meaning traders' decisions are no longer dependent on transaction costs.
- Protection against MEV is built through a dedicated mempool and an execution architecture that separates the order of operations from the regular blockchain flow.
The dedicated trading layer gives AFX more control over mempool behavior, transaction flow, and execution logic than DEXs launched on overloaded general-purpose blockchains.
At launch, four contracts were opened: Bitcoin and Ethereum with leverage up to 40x, as well as gold (XAU) and oil (CL) with leverage up to 25x. At the time of publication, the exchange had added Solana, XRP, silver (XAG), and HYPE. Margin is in USDC.
Deposits are made through the Arbitrum network, with connections available via MetaMask, Rabby, Coinbase Wallet, or the WalletConnect protocol. The initial asset set is indicative: AFX is entering the same niche where Hyperliquid is already recording macro asset and RWA records.
Not Just Speed
Hyperliquid has set a high bar: orders are executed in an average of 0.2 seconds, and even in rare cases of significant delays, around 0.9 seconds. AFX claims about 0.12 seconds for processing a typical request, over 50,000 operations per second, with the potential to increase this figure to 100,000–200,000.
However, comparing just the numbers does not tell the whole story. During calm periods, almost any major exchange operates quickly. The real test comes during sharp market movements when positions are rapidly closed, holding fees change drastically, and trading bots simultaneously send a massive number of orders.
“In such moments, it’s not just the maximum speed of the system that matters, but also how stably it operates under load: how quickly it processes and cancels orders, how it distributes the order queue, and how predictably it executes trades. The AFX trading system is designed to prioritize stability and predictability during peak load periods,” AFX representatives note.
Mark Price from Three Sources
Mark price is the calculated price of a contract used by the exchange to determine margin, PnL, liquidations, and conditional order triggers. In illiquid markets, a single large order can sharply shift the last transaction price. If the mark price is too reliant on the local order book, it opens the door to manipulations and forced liquidations.
AFX calculates mark price as the median of three independent components:
- An external oracle price for the underlying asset smoothed over a 2.5-minute moving average;
- The mid price from AFX's own order book;
- The weighted median mid price of the same contract on centralized exchanges Binance, OKX, Bybit, Gate, and MEXC, with weights of 3:2:2:1:1 respectively.
If one of the sources provides anomalous data due to order book manipulation, oracle failure, or a stalled feed from an external exchange, the median simply ignores them and takes the average of the remaining values. To significantly shift the mark price, an attacker would need to simultaneously influence at least two independent sources. The calculation is updated once per second.
In practice, this reduces the risk of false liquidations during short-term dislocations in the local order book. A stop-loss triggered by mark price will not activate on a single price spike in AFX's order book if the broader market remains stable.
Orders triggered by the last price, on the other hand, react specifically to local movements. Users can choose the type of trigger, but the architecture of the mark price makes classic attacks through manipulation of a single order book significantly more challenging.
Four Phases of Liquidation and the Role of LP Vault
Cascading liquidations are one of the main stress tests for on-chain exchanges. AFX employs a four-stage protection system that gradually tightens restrictions as the value of assets in a trader's account approaches the minimum required level of position collateral.
The first stage involves canceling open orders that reserve part of the funds. The system first attempts to free up capital without forcibly closing the position itself.
If funds are still insufficient and the value of assets in the account drops below the maintenance margin level, the second stage is triggered—forced liquidation of the position on the open market. If there are still funds remaining after the liquidation, they are returned to the trader.
The third stage activates if standard liquidation fails to stabilize the account—such as during a sharp market movement when the position quickly incurs significant losses. In this case, the LP Vault (or ALP) takes on the open position and the trader's remaining funds, becoming the counterparty to this transaction and assuming the risk.
The fourth stage is the automatic reduction mechanism for profitable opposing positions (auto-deleveraging, ADL). This is used only as a last resort: if accepting the losing position causes the LP Vault's own funds to go negative, the system begins to reduce profitable opposing positions of other participants to cover the remaining loss.
Source: Medium.In this model, the LP Vault serves as the main protective buffer of the system. The pool maintains liquidity in the order book, accepts liquidated positions, and takes on some risk during sharp market movements. ADL remains a backup mechanism in case even the LP Vault cannot fully cover losses.
Professional Orders and Cross-Margining
In cross-margin mode, AFX allows unrealized profits to be used as additional collateral. If an open position becomes profitable, the available margin automatically increases—there's no need to close the trade and withdraw profits for this. This makes capital management more flexible: traders can increase positions, open hedges, or rebalance portfolios without unnecessary actions.
Hyperliquid is also developing its margin system: its documentation describes margin tiers that determine maximum leverage and maintenance margin requirements.
“The difference lies in the emphasis. Hyperliquid has built one of the strongest ecosystems for on-chain perpetuals. AFX enters the market with capital efficiency as a key architectural principle, targeting traders who manage a portfolio of multiple positions and expect margin logic at the level of a professional trading system,” AFX representatives comment.
The platform offers three types of orders. Market orders allow traders to pre-set acceptable slippage in the range of 0.5% to 5%. Limit orders support the main execution modes—GTC, IOC, FOK, and Post-Only. Conditional orders (stop-market and stop-limit) can trigger either at the last transaction price or at mark price. The roadmap also includes TWAP execution and a hedge mode with simultaneous long and short positions on the same trading pair, but these features were not available at launch.
Base fees are 0.01% for makers and 0.06% for takers. In the VIP program, rates decrease as the 30-day trading volume increases—both the main account and sub-accounts are considered. Users at the highest VIP levels also receive a share of the platform's commission income.
The referral program operates through wallet linking: after following a unique link, a user is tied to the inviter, and the reward size is recalculated daily based on the total trading volume of the referred network.
Betting on Quant Traders and Market Makers
According to the AFX team, the majority of on-chain derivatives volume in the coming years will be generated not by retail traders through an interface, but by market makers, quant traders, copy trading communities, and AI agents.
Hyperliquid has a mature API infrastructure and detailed documentation, but it has limitations that advanced users encounter. Each account by default allows 1,000 open orders, with a ceiling of 5,000 as volume increases. Upon reaching the 1,000 order limit, the platform may reject reduce-only and trigger orders.
“AFX is designed around the needs of professional clients from day one. The long-term goal is to attract not those seeking a decentralized alternative to Binance or Bybit, but those who want to conduct full-fledged on-chain trading operations without sacrificing execution quality,” AFX emphasizes.
Economics Without Venture Capital
In terms of token distribution, AFX emphasizes the absence of a venture round, private allocations, and pressure from future unlocks. The project targets active traders, liquidity providers, and the community rather than early-round investors. The token itself has not yet been released.
At launch, the team initiated trader incentive programs: a VIP system with reduced maker and taker fees based on trading volume, as well as a referral model with wallet linking and commission distribution.
This approach resonates with Hyperliquid's model: first, launch the infrastructure and gather liquidity, then distribute the token through user activity rather than through traditional venture rounds.
“Hyperliquid has one of the strongest communities in DeFi, and this is one reason it became the segment leader. AFX can build on the lessons of the first wave with a clearer focus on community-driven economics,” AFX asserts.
Trading activity before the token's release forms a base for future genesis distribution, and traders accumulating volume and liquidity now are counting on an airdrop. Hyperliquid followed a similar path, with its distribution in November 2024 being the most generous in history, peaking at a valuation of ~$10.8 billion. Since then, the platform has consistently grown in trading volume, and the HYPE token continues to reach historical highs as of June 2026.
However, the experiences of dozens of other projects show that liquidity attracted for the drop often dissipates in the first weeks after distribution.
For AFX, two critical factors will be the structure of the genesis distribution and the ability to retain professional participants after the early-stage incentives are exhausted. Without the latter, even a technically perfect order book risks being left with a thin order book.
AFX vs Hyperliquid: Summary Table
ParameterAFXHyperliquidPositioningSovereign Trading Layer for professional on-chain derivativesOn-chain order book perp-DEX + spot + application ecosystemOrder BookFully on-chainFully on-chain, matching by price-time priorityConsensus and ArchitectureMysticeti DAG BFT, ABCI + Cosmos SDK, modular structureHyperBFT, HyperCore + HyperEVMLatency~120 ms P500,2s median for colocated clientsThroughput50,000+ TPS; theoretical ceiling 100,000–200,000 TPSHyperCore optimized for high-performance on-chain tradingExecution ModelExecution and consensus are separatedHyperCore processes trading logic nativelyMarginReuse of unrealized PnL, cross-positions, real-time risk controlMargin tiers and maintenance margin systemMEV ProtectionDedicated mempool and execution architectureOn-chain order book and transaction ordering rulesTarget AudienceProfessional traders, quants, HFT, market makers, trading communitiesRetail traders, professional traders, developers, ecosystem usersWhat Time Will Show
The technological foundation of AFX addresses specific pain points of on-chain perps: the three-component mark price reduces the risk of price manipulation, the four-stage liquidation system with LP Vault adds an additional buffer between losing positions and other market participants, and the reuse of unrealized profits in cross-margin mode enhances capital efficiency for professional trading strategies.
The coming months will reveal what was not visible from marketing materials:
- Will AFX maintain the claimed latency and mark price behavior during moments comparable to the JELLYJELLY and POPCAT episodes on Hyperliquid?
- Can the LP Vault absorb liquidated positions in the first truly massive cascade without triggering ADL?
- Will professional market makers arrive before and after the token's genesis distribution?
- What will the token distribution structure look like?
The gap from Hyperliquid is unlikely to close quickly in the near term. The segment leader has built a complete ecosystem, and its token continues to reach historical highs above $60. The interest in such platforms and the growth of the largest player may also signal positive trends for AFX, confirming sustained demand for on-chain derivatives.