The derivatives market, in both traditional finance and crypto, far outstrips the spot market. For example, the volume of BTC derivatives is 17 times higher than its spot volume.

In crypto, perpetual futures contracts largely dominate the derivatives market. These perpetual futures contracts, which have no maturity date, make it easy to take positions with high leverage.

Platforms are competing fiercely to gain a foothold and a dominant position in this huge market. FTX, a fast-growing centralised exchange, saw its abrupt closure in November 2022 as a reminder of the crucial importance of personal custody of assets.

Blockchains make it possible to minimise this risk, as users hold their assets themselves. In addition, they offer the possibility of creating applications that can be audited in real time. In this way, decentralised perpetual trading applications, or "Perp DEX", aim to compete with their centralised counterparts (CEX).

There has been an upward trend in the volume share of DEX futures relative to CEX since FTX fell in November 2022. However, this share is still very low, at just over 3%. In comparison, spot volume on DEX accounts for 13% of that on CEX.

This difference is mainly explained by the performance requirements of derivatives trading, particularly in terms of latency - the time it takes to execute an order. Blockchains, which have to form a consensus between the nodes in their network, are at a disadvantage compared with centralised platforms that do not have this constraint.

As a result, the fastest layer 1 blockchains all have a latency of several hundred milliseconds, while a platform like Coinbase can drop below 1ms. Of course, latency also depends on the distance between the user and the CEX servers or network nodes.

Reliability is another crucial factor: speed loses its relevance if order execution is not guaranteed. For example, on Solana, 42.5% of trades fail.

High-frequency trading firms, like all traders, require both minimal latency and high reliability to optimise the execution of their trades and ensure the performance of their strategies.

Perp DEXs must therefore innovate to try to rival the performance of CEXs, while offering unique features enabled by blockchains. The sector is extremely competitive, with a real race to attract sufficient liquidity and create a network effect, in order to make a lasting mark on this market.

We will therefore compare the different approaches adopted by Perp DEX.

👉 dYdX: the pioneer

dYdX was the first application to offer a good perp onchain trading experience. Initially built as an Ethereum layer 2 using technology from Starkware (StarEx), the company behind the Starknet network, the project later launched "dYdX V4", its own layer 1 based on the Cosmos SDK, in October 2024.

The dYdX Chain is secured by dYdX stakers, remunerated in USDC from trading fees paid by users, which currently represents an annual return of 8%. The network validators also host an onchain order book and act as an oracle to provide price feeds on the network.

The point of an onchain order book is to make it verifiable that orders issued have been executed correctly. A non-verifiable order book means trusting your operator not to front-run trades, manipulate their order or censor them. For example, the CFTC had accused Binance of trading against its users in November 2023.

dYdX launched its V5 in June 2024: the protocol now supports isolated markets to improve risk management and allow the creation of new trading pairs of niche assets such as memecoins. V5 also adds vaults into which users can deposit USDCs that will provide liquidity on the various markets in exchange for returns.

> Read also - Charles d'Haussy (dYdX Foundation): "Only Cosmos allowed us to have total control over our project"

The next update to the protocol should allow any trading pair to be listed without requiring anyone's permission. Still in the spirit of being able to "trade everything", it should also be possible to trade the prediction markets available on Polymarket with leverage. In fact, it is already possible to trade the US presidential elections with leverage on dYdX.

As a layer 1, dYdX must ensure the security of its own chain: the cost of an attack must outweigh the potential gains, for example recovering USDCs from the network or manipulating the order of transactions. Given that the protocol is based on the Tendermint consensus, an attacker could also block the network by controlling 33% of the staked tokens.

A major challenge for dYdX is to reduce the latency of its blockchain, which currently produces a block every 0.98 seconds.

As the chart below shows, dYdX V3 (in light green) has long accounted for more than 80% of DEX Perp volume. However, increasing competition has led to a steady loss of market share, and the launch of the dYdX Chain has not reversed this trend. It remains to be seen whether future updates can change this.

👉 Hyperliquid: the new giant

yperliquid currently dominates the Perp DEX sector with 25% of the market share and is growing steadily.

Like dYdX, Hyperliquid is a layer 1 blockchain optimised for trading, incorporating an onchain order book and its own oracle. To optimise execution speed, Hyperliquid has developed HyperBFT, a consensus algorithm inspired by Hot Shot. A trader co-located with a network node benefits from a median latency of 200 ms, placing Hyperliquid among the fastest layer 1s.

The protocol has integrated vaults enabling various strategies, such as liquidity provision or copy trading. Anyone can deploy their vault with their own strategy.

Hyperliquid is working on implementing support for the Ethereum Virtual Machine (EVM) to enable other applications to be deployed on its blockchain, aiming to create an ecosystem.

It is notable that Hyperliquid has been entirely self-financed by its team, without recourse to external investment. This approach has been particularly well received by its community, at a time when criticism is mounting against projects reserving investment opportunities solely for venture capital funds (VCs).

Hyperliquid's HYPE token has not yet been launched. The protocol has been running a points campaign to drive activity on its app. The token could be launched by the end of the year.

Despite its successes, Hyperliquid has some grey areas, the biggest being its centralisation. All of the chain's validators are managed by the team, potentially giving it total control of the funds.

The decentralisation of the chain, planned with the launch of the token, will bring new challenges. In particular, the protocol will have to contend with the maximum extraction value (MEV), which could adversely affect the user experience. In addition, decentralisation could affect the performance of the blockchain.

The team has deployed its own bridge to Arbitrum, but this stalled for several hours in early October, due to a synchronisation error between the nodes of the two blockchains.

Hyperliquid will therefore have to meet the challenges of decentralisation while maintaining the growth of its business following the distribution of its airdrop.

👉 Jupiter: the preferred trading venue on Solana

Jupiter is the largest spot DEX aggregator on Solana and one of the largest Perp DEXs across all channels.

In contrast to Jupiter, dYdX and Hyperliquid are synthetic Perp DEXs. Their trades are not based on real assets, apart from the USDC, which is their sole source of liquidity. Instead, these platforms use order books, vaults and oracles to manage positions. The advantage of the synthetic model lies in its ability to easily integrate new trading pairs, provided an oracle provides their prices.

Jupiter takes a different approach: traders here interact with a liquidity pool called JLP, made up of real tokens - SOL, ETH, wBTC, USDC and USDT. This model reduces reliance on oracle price manipulation and ensures that each trade is actually backed by the corresponding assets.

This approach was initially popularised by GMX, one of the largest DEX Perps on Arbitrum.

Jupiter has no order book. It is the JLP holders who act as counterparties for traders: they collect trading fees and take reverse positions. This configuration also exposes them to potential losses when traders are predominantly winning.

JLP holders also help to boost the liquidity of pool assets on the spot market. They benefit from the composability of blockchains by being able to use their JLPs on other applications, for example as collateral on Drift, another DEX Perp from Solana.

One of the main limitations of this model is the difficulty of integrating new asset pairs.

Jupiter incorporates Edge, the oracle developed by Chaos Labs, which adjusts the protocol's various risk parameters in real time while providing asset price feeds. In the event of Edge's failure, Jupiter relies on the prices provided by Pyth.

Jupiter takes advantage of the composability offered by a general-purpose blockchain. However, this dependency exposes the application to the risks of congestion caused by other projects on the same chain.

Jupiter's main challenge is to integrate new assets within its application.

Another major challenge is to increase the usefulness of the JUP token beyond simple governance. Currently, JUP holders benefit from rewards offered by projects launched via the LFG launchpad, developed by Jupiter.

👉 Vertex: unifying multichain liquidity

Vertex began as a DEX Perp on Arbitrum. In February 2024, the team launched Edge: an innovative order book connecting liquidity between blockchains in real time (not to be confused with the Edge oracle mentioned earlier).

The protocol consists of two key elements: instances on various blockchains, allowing users to deposit different tokens, and an off-chain sequencer managing a single order book. The latter executes orders in 5 to 15 ms, with transactions then being finalised on their respective chains at their own speed.

Vertex therefore represents a hybrid model between appchains such as dYdX and Hyperliquid, and applications deployed on more generalist channels such as Jupiter, Drift and GMX. Its order book operates independently of the congestion and speed of the blockchains hosting Vertex instances, while benefiting from the composability offered by these generalist blockchains.

Currently, Vertex is deployed on Arbitrum, Base, Mantle, Blast and Sei. Users can deposit various tokens from these blockchains and use them as collateral. These tokens can also be borrowed by other users via a lending pool similar to Aave, generating returns.

As transactions are finalised on their respective chains, all benefit from activity on the various instances of the protocol. These channels can offer their tokens to Vertex, which redistributes them during trading competitions, thus contributing to the liquidity of all the channels. In particular, this synergy has enabled Sei, Mantle and Blast to attract new users.

In addition, other protocols can integrate Edge's multichain order book to benefit from and contribute to its liquidity, as Pear Protocol does.

In the event of a malfunction of the sequencer managing the order book, the protocol would temporarily switch to an MA model in order to ensure exchanges, in all cases users remain the sole owners of their funds and the sequencer has no way of executing orders without the users' signature.

Staking the VRTX token allows users to receive a share of the trading fees generated by the protocol and to reduce their own fees, following the example of Binance. The project recently launched the "VRTX Vertical Program", aimed at gradually improving the usefulness of VRTX and its integration into the application.

The majority of markets on Vertex rely on prices provided by Stork's oracles, while Chainlink's "data streams" currently secure ETH pairs.

> Read our analysis of the different oracle solutions

The major challenge will be to decentralise the sequencer operating the order book to make it resilient and verifiable, while maintaining its current execution speed.

There are also questions about the potential impact that a reorganisation or corruption of one of the chains supported by Edge could have on the entire protocol.

The Big Whale's view 🐳

In contrast to CEXs, Perp DEXs can take advantage of the benefits offered by blockchains, i.e. personal custody, verifiability and composability. However, they face challenges posed by latency and SRM. The aim is to come up with the best compromise.

Layer 2 development represents the next frontier in terms of scaling blockchains. Projects such as MegaETH, Rise, Eclipse or Atlas will be able to create onchain applications with much lower latency than current blockchains. This will be possible thanks to a limited number of validators and an optimisation of the execution function, while relying on Ethereum's security.

Similarly, decentralised specialist services such as EigenLayer's AVS will be able to improve certain specific components of Perp DEX. Vertex's order book is approaching this category.

It is important to note the evolution of other applications, such as Deriv (formerly Lyra), which is in the process of positioning itself as the leader in the onchain options sector.

Onchain trading is still far from being able to compete with CEXs, either in terms of performance or volume. It is crucial for these applications to offer sufficient volume to attract the biggest traders. The major challenge is to manage to grow that volume from scratch with token rewards, and then to sustain that growth organically.

If Perp DEXs manage to gain sufficient traction and allow new assets to be listed permissionless, it will also be a victory for new projects that are currently forced to pay fortunes to be listed on CEXs.

‍