Introduction
Perennial is a DeFi-native perpetuals trading protocol that emerged to tackle the limitations of early decentralized derivatives platforms. Traditional decentralized perpetual exchanges typically use either AMM or CLOB models. AMMs provided simplicity but suffered from capital inefficiency and slippage issues, while on-chain order books struggled with throughput on general-purpose chains.
The solution is a hybrid infrastructure combining the strengths of both AMMs and CLOBs (such as using off-chain price oracles or partial off-chain order matching). Perennial uses an intent-based trading mechanism designed to unify liquidity across venues and deliver a centralized exchange (CEX)-like trading experience on-chain. Price feeds for any asset can be supported, as well as for any derivative that can be computed out of it, no matter how exotic.
The protocol features a permissionless on-chain layer where other protocols can build on top of, seamlessly integrating intents to attract market makers and off-chain solvers. This approach enhances liquidity conditions, fosters network effects, and optimizes trading efficiency.
Key Takeaways
Hybrid efficiency-oriented design: The protocol architecture seeks to take the best out of two different but complementary trading models, synthetic AMMs and intent-based systems; the former enables permissionless composability and on-chain settlement, while the latter enhances market efficiency by aggregating liquidity and improving trade execution.
Customizable payoff curves: Perennial can support any price feed and, therefore, enable trading for any asset or derivative computed from it (simple long/short, power perps, convex payoff positions like squeeth and power perps, pairs trading, etc).It has already integrated some forex (JPY, EUR, GBP) and commodity (GOLD) market assets and can potentially support indexes and stocks.
Flexible Liquidity Provisioning Solutions: Liquidity providers on Perennial have the freedom to operate with high leverage (up to 100x) and can tailor their risk exposure and hedging strategies to suit their preferences. For those seeking a more hands-off approach, passive liquidity providers can use vaults built on top of the protocol, which simplify the process and eliminate complexity.
Tailor-made for developers: Perennial was built with flexibility in mind, designed as a minimalistic primitive to facilitate protocol integrations. Partners building on top of this base layer can allocate a part of the collected fees to the product owner (e.g. individual organizations, decentralized autonomous organizations (DAOs), or other protocols).
Permissionless and composable: On Perennial, both the integration and creation of new markets involve a permissionless process. As long as a price feed exists, any market can be created for an underlying asset, as well as parameterized to enable custom payoff structures.
Background
Perennial was developed to address the inefficiencies of early DeFi derivatives, which relied on simple but extremely inefficient mechanisms, heavy incentivization, and lacked scalability for further growth. Early derivatives models on Ethereum mainnet were constrained by high computational costs, preventing them from competing with centralized platforms. Hence, most solutions resulted in worse UX and more expensive operations, making it infeasible for large players to deploy capital in size. As a result, only a few protocols like dYdX stood out during the early DeFi days, which they achieved by building their own custom rollup instead of deploying on top of Ethereum.
The second generation of derivatives on Layer 2 solutions (L2s) reduced costs and enabled greater experimentation, but challenges remain in achieving widespread adoption and overcoming inherent limitations. For instance, the increasing number of L2s has resulted in liquidity and user fragmentation across ecosystems, reducing market depth. In addition, operational challenges, including infrastructure scalability or reliance on oracle providers and off-chain data sources, have limited its seamless adoption. Lastly, the dependence on liquidity pools of most AMM-based solutions since the launch of GMX, while providing a constant liquidity solution independent of direct intermediaries, exposed traders to wider price spreads, transferred excessive risk to Liquidity Providers and has often resulted in high funding costs as Open Interest imbalances are borne by the protocol (i.e. by Liquidity Providers).
Next, a third generation of derivatives has emerged built on appchains like Aevo, Derive, dYdX, or Hyperliquid. This represents a more focused approach to derivatives trading by integrating order books and tailoring infrastructure to meet the specific demands of these markets. However, this further exacerbates the issue of liquidity and user fragmentation, competing for liquidity against existing chains. Their bespoke infrastructure and customizability also comes with increased development and operational costs, as they must maintain order books, matching engines, and other liquidation engines. Furthermore, those using order books need to reach onboarding agreements with market makers in order to manage liquidity on their chain exchanges. Also, interoperability with these chains still remains a challenge, and most L2 sequencers are still centralized and managed by their respective teams, almost never sharing revenue with users or token holders.
Perennial seeks to overcome these challenges not only by offering improvements over existing DeFi primitives or trading products, but also enabling unique use cases that are only possible on-chain. The protocol adopts a minimalistic and unopinionated design that scales with network effects: as more protocols build on top and the liquidity improves, an increasing number of builders and traders feel more attracted to that liquidity.
The protocol aims to offer a Web2-like UX while leveraging the unique benefits that come from building on-chain. Its main features include:
- Perennial Rollup on Base (L3) – a dedicated Optimism-stack rollup that gives the protocol full control of block timing, data layout and fees. It delivers sub-second finality, negligible gas costs and throughput measured in tens of thousands of matched orders per second, while inheriting Base / Ethereum security.
- Zero price impact from big buys/sells, as trades are based on external reference prices (provided by oracles, namely Pyth) and liquidity is synthetic. However, simulating price impact is necessary because, without it, the market would simply reflect external feeds, limiting price discovery and arbitrage opportunities. By adjusting internal prices based on trade volumes or open interest, Perennial ensures a dynamic and functional trading environment.
- Composable infrastructure, acting as a permissionless liquidity layer on top of which any protocol can build. This shared liquidity becomes more efficient as it gets used by more and more protocols. As an example,Cryptex Finance uses Perennial as infrastructure to launch indexes such as the Total market cap perps, other NFT/RWA indexes, etc.
Perennial was co-founded by Kevin Britz and Arjun Rao, both of whom have extensive experience in the cryptocurrency and technology sectors. Prior to establishing Perennial, they co-founded Astro Wallet in 2017, which was later acquired by Coinbase in 2019. Both Britz and Rao have also held advisory roles at Coinbase, the former as Engineering Manager and the latter as Staff SWE. Other team members’ backgrounds include the likes of Manifold Trading, Gemini, Gearbox, Cega, Maker, Coinbase, and more.
In December 2022, Perennial secured $12 million in a seed funding round co-led by Polychain Capital and Variant. Additional investors included Archetype, Coinbase Ventures, Scalar Capital, and Robot Ventures.

Sector Outlook
Perpetual DEXs are among the most adopted and rapidly growing categories in DeFi. While they gained substantial traction after the collapse of FTX, the pace of development and innovation in this space over the past two years has been remarkable.
Since the advent of L2s and appchains, the sector has rapidly evolved to accommodate for a wide variety of models, tools, and complementary services, ranging from Central Limit Order Books (CLOBs) to on-chain AMMs. Technological evolution was then followed by a good level of adoption, as evidenced by the steady growth of the CEX/DEX perp volume ratio

Nonetheless, the rise of perpetual swaps reflects the growing maturity of the crypto market, as they enhance liquidity, foster more efficient price discovery and support more sophisticated financial strategies. Beyond catering to the demand for speculative leveraged trading , these tools enable effective risk management by allowing traders to hedge positions, transfer exposures, and navigate financial uncertainty more effectively.
Before delving into the infrastructural details of the different perp models developed over time, let’s understand what perpetual markets are and how they work.
Perpetual Swaps
Perpetual swaps, or perpetual futures (perps for short), are derivative contracts that let traders speculate on the price movements of an underlying asset (such as stocks, crypto, bonds, or commodities) without requiring them to own or deliver the asset. Unlike traditional futures, perpetual futures have no expiration date, offering continuous exposure to the asset, which gives the instrument its “perpetual” nature.
To maintain the contract’s price in line with the underlying asset, these markets employ a funding rate mechanism. This involves periodic payments between market participants, where the side with greater demand compensates the opposing side to keep positions open and maintain price alignment between future and spot markets.

Like any other derivative, the value of a perpetual contract is driven by supply and demand for the contract itself, rather than the price of the underlying asset it aims to track. When the price of the perpetual contract (known as the mark price) diverges significantly from the spot price of the underlying asset (typically represented by the index price, a weighted average of prices from various centralized exchanges), the funding rate increases accordingly.
The benefits that the category offers over spot markets are many and can be summarized as follows:
- The availability of leverage, often reaching 100-1000x, depending on the underlying asset.
- The ability to profit from both upward and downward price movements by taking long or short positions.
- Generally deeper liquidity, resulting in narrower spreads and reduced price impact, and covering an expanding array of assets beyond crypto, including equities, traditional financial indexes, commodities, and forex.
- Enhanced flexibility for users’ crypto portfolios, enabling effective hedging and delta-neutral strategies.
- A strong use case combined with the potential for high profit margins, even with minimal trading volumes, when managed efficiently.
In order to fully understand the pros and cons of each paradigm, we first need to establish objective evaluation criteria, such as the following:
| Liquidity
Management |
Liquidity sourcing | Evaluates how liquidity is provided (AMM Liquidity Provisioning
or Market-Making managed liquidity on CEXs) |
| Access to cross
market liquidity |
Measures integration with external markets for better prices. | |
| LIquidity Dynamics | How liquidity responds to market changes, such as high volatility or
imbalances in long and short positions |
|
| Capital costs,
efficiency, and management |
Fee structure | Evaluates the costs incurred by users, including trading and funding rates,
and any additional costs for liquidity provision or collateral management |
| Price discovery
mechanism |
Assesses how the model determines asset prices, such as relying
on oracles, market makers, or AMM-based algorithms |
|
| Cost efficiency | Looks at how well the model minimizes expenses like gas
fees, operational costs, and trading costs for its user base |
|
| Capital efficiency | Examines how the model utilizes collateral and liquidity. | |
| Trade matching | Reviews how trades are executed and matched | |
| Security | Oracle dependence | Analyzes the reliance on price oracles for market operations and pricing |
| Permission degree | Considers how permissionless or open the model is, both in terms
of user access and protocol operations |
|
| Centralization risks | Identifies potential points of control or centralization | |
| UX | Order types | Assesses the range and flexibility of available order types |
| Ease of use | Evaluates how intuitive and accessible the platform is for users | |
| Price Transparency | Reviews how clearly and accurately the model displays pricing
information, including fees, spreads, and price impacts |
|
| Execution speed | Measures the speed of trade execution, influenced by factors like network
congestion, off-on-chain mechanisms, and the underlying chain’s performance |
Management
or Market-Making managed liquidity on CEXs)
market liquidity
imbalances in long and short positions
efficiency, and
management
and any additional costs for liquidity provision or collateral management
mechanism
on oracles, market makers, or AMM-based algorithms
fees, operational costs, and trading costs for its user base
of user access and protocol operations
information, including fees, spreads, and price impacts
congestion, off-on-chain mechanisms, and the underlying chain’s performance
As the market matures, the demand for robust derivatives infrastructure has only grown, making perpetual DEXs one of the most prolific areas of innovation in the crypto space. Additionally, the shift away from centralized exchanges due to custody risks and the advancements in execution performance brought by appchains and Layer 2 solutions have laid the foundation for decentralized leveraged trading to thrive.
To meet these demands, many trading models have emerged, each tackling the challenges of on-chain derivatives in different ways.
- Virtual/synthetic AMM-based approaches, like virtual AMMs inspired by Perpetual Protocol, simulate perpetual trading without actual spot exchanges. They add a price impact function on top of a virtual price curve that often takes inspiration from Uniswap’s constant product formula. While they can offer self-custody, composability and ongoing liquidity at any price level, LPs take on big risks during volatile periods with large imbalances between longs/shorts that might even lead to insolvency in some cases. E.g., In May 2022, the extreme volatility experienced during the collapse of UST and LUNA led to significant price deviations in the Perpetual Protocol v1 markets. The protocol liquidation mechanisms failed to function properly, and the insurance fund, which acted as a backstop solution, was completely drained, resulting in more than $5.7 million in losses due to an economic design flaw and delays in oracle price feeds.
- Oracle-based LP driven systems, popularized by GMX, rely on price feeds from CEXs, offering cost efficiency and composability but exposing users to risks such as price manipulation and system downtime.
- CLOB-based exchanges running on custom L1/2s that offload trading activity to off-chain matching engines, along with other computational complexities that live in off-chain backend systems, although often maintaining on-chain settlement. Even though this paradigm solves many disadvantages associated with AMM models, it still introduces new risks, starting with centralization and higher operating and maintenance costs, mostly coming from the need to onboard market makers to the exchange.
- Fully on-chain order books, on the other hand, prioritize transparency and integrity but face technical challenges such as blockchain latency and throughput constraints, leading projects like Hyperliquid to develop custom Layer 1 chains to overcome these limitations.
- Among these models, intent-based trading is a newer approach gaining significant traction. This model transforms trading into a decentralized, bilateral and over-the-counter experience. Users submit intents specifying their desired trading conditions, such as opening a position at a particular price, which are then broadcast to market makers. These market makers fulfill the intent orders by locking collateral and establishing on-chain agreements while hedging delta exposure, and thus remain market neutral, on other trading platforms (such as Binance). This design reduces the need for high-throughput on-chain interactions, as market makers only engage when executing an intent. It also allows them to source liquidity globally, offering better pricing and deeper markets for users. Compared to traditional AMMs or centralized order books, intent-based trading combines flexibility, efficiency, and improved asset exposure.
Different Types of AMM models
Over time, many teams have refined the AMM-based model, leading to the development of a more structured sub-category that can be further divided into:
| AMM model | Description | Protocols adopting it |
|---|---|---|
| vAMM | AMM-based models designed to simulate the liquidity depth of an AMM and its operational mechanics (for pricing in and out of trading positions) through the definition of virtual parameters |
Perp Protocol, Drift (as a building block within a hybrid design), NftPerp |
| Multi-asset LP oracle based | Relies on oracle price feeds to execute trades and leverages an AMM-based pool consisting of a basket of assets (mainly stable and main crypto assets) as a counterparty for spot swaps and leveraged trades opened on the platform | GMX, HMX, Jupiter |
| Synthetic, oracle-based peer to pool AMM |
Offers a comprehensive and easily integrated framework (thanks to the SDK developed by the team) for creating derivative products with two-sided, customizable payoffs, as long as they can be derived from the oracle feed itself |
Perennial (as a component within a hybrid design), Synfutures |
| Hybrid design | Integrates multiple design approaches in an attempt to harness the best features of each while at the same time trying to mitigate each one’s shortcomings. AMM is widely used as a default or backstop solution because it offers continuous liquidity to traders | Drift, Synfutures, Vertex, Perennial |
Although there are many different paradigms available today, each with its own unique features, it is still possible to identify common traits across the whole category:

Continuous liquidity availability as long as there are LPers within pools (thus ensuring that users can always trade, even in low-volume markets).
Automated, permissionless, non custodial and interest-bearing MM solutions suitable for any user type, i.e. the LP positions. Liquidity pools basically enable anyone to act as a market maker by depositing assets, decentralizing the liquidity supply and reducing dependence on professional market makers (MMs).
Composable nature. LP positions of perp AMM models are generally composable within the DeFi ecosystem, thus increasing the versatility offered to traders.
This wide range of different AMM models for perpetual futures trading emerges from the ongoing challenge of balancing performance with decentralization. This trade-off is a recurring theme for developers working in the space, as they strive to create solutions that deliver both high efficiency and a commitment to decentralized principles.
CLOB-Based Systems
The Central Limit Order Book (CLOB) model, widely regarded as the gold standard for performance, excels in speed and efficiency. Its ability to handle high-frequency trades and execute transactions with minimal delay makes it attractive for traders seeking a seamless experience. However, this performance is often achieved at the expense of decentralization. CLOB systems depend heavily on professional liquidity providers, i.e. market makers who maintain the order book’s depth and ensure smooth trading. Moreover, the batching of transactions in these systems is typically handled by centralized sequencers, often under the control of the platform’s team. This introduces a layer of centralization that can conflict with the ethos of DeFi. Furthermore, the infrastructure supporting these systems, optimized for low latency and high throughput, can further undermine decentralization by managing off chain the computational complexities of the matching engine and creating barriers to entry for smaller, independent nodes. The major players in this subcategory are Hyperliquid, which dominates the perp-on-chain volume ranking, Aevo, Paradex, and Derive (formerly Lyra).
Intent-centric Paradigms
Intent-based auction systems are a much newer model drawing on the extensive research and development work on the spot side: the most relevant technological contributions on the topic are undoubtedly those produced by Flashbots, Cow Protocol, 1Inch, and Uniswap-Across. Although perp infrastructures took its cue from the technological contribution made by spot dex projects, the two systems are very different because they meet different needs.
Spot Intent Models
Generally, each spot version of intent centric architecture has the following components:
An intent centric execution environment, either through the definition of a set of rules for communicating intent orders or through the adoption of third-party developed standards.
An auction model with design choices inspired by performance optimization (dutch or batch style, potential exploitation of peer-to-peer matching solutions, more or less extensive information sharing with bidders/solvers, limited or open participation in the auction process, etc).

A network of Solvers specialized in the type of service offered (mostly MM activities, thus Market Makers), i.e. bonded third parties executing trades on behalf of users through a delegated execution model that focuses on optimizing trade routes.
The main features offered by this system to the perp dex category are multiple and can be summarized as follows:

- Extreme flexibility in creating, editing, and deleting orders thanks to intents
- Liquidity aggregation by default
- No need for LPers
- MEV risk transfer to qualified third parties (i.e. solvers)
- Exploitation of both on chain and off chain liquidity sources – Market Makers’ (MMs) private inventories, Cex liquidity, AMM liquidity pools, etc
Perps Intent Models
Intent-based perpetual models allow users to submit intents (declarations of their desired trading actions) which are then broadcast to a network of market makers (MMs) or solvers, effectively functioning as a decentralized OTC desk. Market makers fulfill these intents by forming on-chain agreements, where both parties lock collateral to ensure security and reliability.
Unlike systems reliant on passive liquidity providers (LPs), intent-based models leverage market makers who operate across many different venues, including centralized exchanges like Binance or Bybit. These MMs can hedge and arbitrage price differentials efficiently, often offering better prices for trades due to their ability to mitigate risks and capitalize on cross-market opportunities.
The power of intent-based models lies in their ability to generate a self-reinforcing liquidity flywheel. As the network of solvers or market makers grows and diversifies, liquidity increases, fostering greater competition. This competition drives improved order fills, offering traders better pricing and faster execution. The enhanced trading experience attracts more order flow, further incentivizing market makers to join the network. This continuous feedback loop strengthens the ecosystem, creating a scalable and highly efficient trading environment.
Comparing Models
Let’s now see how GMX AMM (as a reference example for all oracle-based AMM models) and Perennial’s new hybrid infrastructure meet the evaluation criteria examined above and what their key features are.
Perennial offers a hybrid design, which is based on the AMM model but also leveraging the flexibility of intents. Notably, Liquidity providers (LPs) benefit from limited risk exposure, as their liability is confined to the imbalance between long and short positions only, and not the total open interest.
| AMM (GMX as reference model) | Hybrid (Perennial after V2.3 update as reference model) | ||
| Liquidity
Management |
Liquidity sourcing | Passive LP-based liquidity (via GLP pool) | Passive LP-based liquidity (through Vaults) + Active market
maker liquidity (intent-based solvers handle trades) |
| Access to cross
market liquidity |
None for AMMs (liquidity remains on-chain) | Accessible but the number of external liquidity resources
available changes for each protocol |
|
| LIquidity Dynamics | Static pools constrained by AMM rules | A blend of static and dynamic sources (Hybrid) | |
| Capital costs,
efficiency, and management |
Fee structure | Fixed fees per trade based on pool mechanics | Flat notional trade fee or spreads from solvers |
| Price discovery
mechanism |
Oracle-based pricing | Mixed oracle and market-driven pricing | |
| Cost efficiency | Low slippage, but gas-dependent | Context-dependent (whether the trade was matched to a solver or against the pool) | |
| Capital efficiency | Low due to the over-collateralization required to cover extreme market events | High efficiency through intent-based hedging, solver optimization, and
infrastructure design (with liquidity providers only netting longs with shorts) |
|
| Trade matching | Automated through pool mechanics (AMM) | Both Automated and bilateral matching via AMM and off-chain solvers respectively | |
| Security | Oracle dependence | Highly reliant on price oracles to align pool prices with market prices. | Balances reliability using oracles for AMM components and
external price feeds (from CEXs) for solving service. |
| Permission degree | Fully permissionless | Hybrid systems may involve partial permissioning | |
| Centralization risks | Decentralized by design, with risks primarily tied to the platform’s governance or infrastructure | Centralization risks depend on the relative weighting of AMM and intent-basedcomponents, offering potential for greater decentralization than intent models alone | |
| UX | Order types | Usually limited to basic orders | Combines AMM simplicity with intent-based order flexibility,
offering a broader spectrum of options |
| Ease of use | Simple interfaces and predictable mechanics make it highly accessible for retail users, though advanced traders may find it a bit restrictive | Balances simplicity for basic users with advanced options for
experienced traders, depending on implementation |
|
| Price Transparency | High transparency as pricing is deterministic and based on pool reserves | Offers a mix of variable transparency and opacity depending on the
weight of the various components on the overall infrastructure |
|
| Execution speed | Deterministic and fast, limited only by blockchain congestion and gas fees | Combines on-chain and off-chain execution, with speed varying by mechanism weighting |
Management
maker liquidity (intent-based solvers handle trades)
market liquidity
available changes for each protocol
efficiency, and
management
mechanism
infrastructure design (with liquidity providers only netting longs with shorts)
external price feeds (from CEXs) for solving service.
offering a broader spectrum of options
experienced traders, depending on implementation
weight of the various components on the overall infrastructure
The next comparative analysis is between pure intent-based protocols, such as Symm.io, and hybrid models like Perennial (after the V 2.3 update).
| Intents (Symm.io as reference model) | Hybrid (Perennial after V2.3 update as reference model) | ||
| Liquidity
Management |
Liquidity sourcing | Fully intent-based, relying on active market makers (solvers) to provide liquidity dynamically | Passive LP-based liquidity (through Vaults) + Active market maker liquidity (intent-based solvers handle trades) |
| Access to cross
market liquidity |
Symm.io solvers actively hedge and arbitrage across external venues (e.g., CEXs), giving the platform access to deep and competitive liquidity | Accessible but the number of external liquidity resources available changes for each protocol | |
| LIquidity Dynamics | Highly dynamic. Liquidity responds in real-time to market demand through solvers, adapting effectively to shifts in trade flows and volatility | A blend of static and dynamic sources (Hybrid) | |
| Capital costs,
efficiency, and management |
Fee structure | Fixed fees or spreads from market makers | Flat notional trade fee or spreads from solvers |
| Price discovery
mechanism |
Solver-driven pricing with minimal reliance on oracles | Mixed oracle and market-driven pricing | |
| Cost efficiency | High. Symm.io reduces on-chain transactions through off-chain intent matching, minimizing gas costs and operational overhead | Context-dependent (whether the trade was matched to a solver or against the pool) | |
| Capital efficiency | Solvers dynamically deploy capital across multiple venues, avoiding over-collateralization and maximizing liquidity usage | High efficiency through intent-based hedging, solver optimization, and infrastructure design (with liquidity providers only offsetting longs with shorts) | |
| Trade matching | Off-chain bilateral matching via solvers | Both Automated and bilateral matching via AMM and off-chain solvers respectively (all trades are then settled into the AMM) | |
| Security | Oracle dependence | Minimal oracle reliance, as solvers – market makers dynamically price trades | Balances reliability using oracles for AMM components and external price feeds (from CEXs) for solving service |
| Permission degree | Permissionless approach but high barrier of entry (+ KYC) to join solving service | Hybrid systems may involve partial permissioning | |
| Centralization risks | Higher risks due to reliance on a smaller network of solvers – market makers, which can lead to collusion or limited competition. | Dependent on the relative weight of AMM and intent based components on Perennial traders/makers/LPs can always fall back on
the AMM at any time to open and close their positions when the solver APIs are inactive or solvers quote poor prices |
|
| UX | Order types | Potential support for any type of order supported on CEX (atm limited to basic order types) | Combines AMM simplicity with intent-based order flexibility, offering a broader spectrum of options (atm the offer is limited to basic order types and intent trigger orders) |
| Ease of use | Complex, requiring familiarity with off-chain matching and solver dynamics, which can deter casual users
(Symm.io acts as a backend infra and does not directly offer trading services, which are instead provided by frontends, such as IntentX) |
Balances simplicity for basic users with advanced options for
experienced traders, depending on implementation |
|
| Price Transparency | Dynamic and solver-driven pricing can feel opaque, as users rely on off-chain decision-making | Offers a mix of variable transparency and opacity depending on the
weight of the various components on the overall infrastructure |
|
| Execution speed | Typically faster due to off-chain matching but may face delays from solver availability or high load | Combines on-chain and off-chain execution, with speed varying by mechanism weighting |
Management
market liquidity
efficiency, and
management
mechanism
the AMM at any time to open and close their positions when the solver APIs are inactive or solvers quote poor prices
(Symm.io acts as a backend infra and does not directly offer trading services,
which are instead provided by frontends, such as IntentX)
experienced traders, depending on implementation
weight of the various components on the overall infrastructure
Another interesting comparative analysis is that between 2 different hybrid models: Drift on Solana and Perennial on EVM-based chains. This will allow us to better assess the impact of the design choices of the respective teams on the final product.
| Hybrid (Drift Protocol as reference model) | Hybrid (Perennial after V2.3 update as reference model) | ||
| Liquidity
Management |
Liquidity sourcing | Triple source of liquidity (aka Liquidity Trifecta: JIT Auctions, Orderbook, AMM as backstop solution) | Passive LP-based liquidity (through Vaults) + Active market maker liquidity (intent-based solvers handle trades) |
| Access to cross
market liquidity |
High. Market makers hedge trades on CEXs and other external venues | Accessible but the number of external liquidity resources available changes for each protocol | |
| LIquidity Dynamics | Extremely dynamic due to market makers’ activity | A blend of static and dynamic sources (Hybrid) | |
| Capital costs,
efficiency, and management |
Fee structure | Tiered fee structure for all futures markets based on an on-chain calculation of 30-Day Volume or Insurance Fund Stake | Flat notional trade fee or spreads from solvers |
| Price discovery
mechanism |
Oracle-based and market-driven pricing mechanism | Mixed oracle and market-driven pricing | |
| Cost efficiency | Very high. Drift leverages Solana’s ultra-low transaction fees and efficient infrastructure, minimizing costs for traders | Context-dependent (whether the trade was matched to a solver or against the pool) | |
| Capital efficiency | High. Drift’s cross-margined risk engine allows traders to use collateral across multiple positions, optimizing capital utilization | High efficiency through intent-based hedging, solver optimization, and infrastructure design (with liquidity providers only netting longs with shorts) | |
| Trade matching | High due to on-chain intents matched via solvers (and AMM acting as a backstop solution). Fast and efficient trade matching using Solana’s low-latency infrastructure ensures seamless execution | Both Automated and bilateral matching via AMM and off-chain solvers respectively | |
| Security | Oracle dependence | While Drift relies primarily on Market Makers – solvers for price discovery, oracles may
still play a secondary role in certain scenarios, such as: providing benchmark data to validate solver-driven prices; supporting liquidation calculations and funding rate adjustments where necessary |
Balances reliability using oracles for AMM components and external price feeds (from CEXs) for solving service. |
| Permission degree | Permissionless for traders and general users, but certain infrastructure components,
like the solver network, are semi-permissioned to maintain efficiency and security |
Hybrid systems may involve partial permissioning | |
| Centralization risks | Moderate because Drift relies primarily on a network of market makers (although the AMM acts as a backstop solution) | Centralization risks depend on the relative weighting of AMM and intent-basedcomponents, offering potential for greater decentralization than intent models alone | |
| UX | Order types | Very wide offer of order types (Stop Market, Stop Limit, TP Market, TP Limit, and Oracle Limit orders. | Combines AMM simplicity with intent-based order flexibility, offering a broader spectrum of options |
| Ease of use | Moderate to high. Drift’s user-friendly interface and seamless trading experience are enhanced by Solana’s speed. However, advanced features may require familiarity from less experienced traders | Balances simplicity for basic users with advanced options for experienced traders (atm limited to basic order types) | |
| Price Transparency | Moderate. Pricing is dynamically determined by market makers | Offers a mix of variable transparency and opacity depending on the weight of the various components on the overall infrastructure | |
| Execution speed | The protocol achieves high execution speed, leveraging Solana’s blockchain infrastructure and its unique design | Combines on-chain and off-chain execution, ensuring good execution speed |
Management
market liquidity
efficiency, and
management
mechanism
still play a secondary role in certain scenarios, such as: providing benchmark data to validate solver-driven prices; supporting liquidation calculations and funding rate adjustments where necessary
like the solver network, are semi-permissioned to maintain efficiency and security
The last comparison specifically covers the topic of fees, which are a crucial aspect of any perp dex.
| Aspect | Hyperliquid | Intent X | Perennial |
| Fee structure | Rolling 14-day trading volume, with sub-accounts aggregated under the master account. | Multiple fees: platform fee, liquidator
fee, CVA, funding rate, and spread considerations. |
Flat trade fee on notional size and a dynamic fee based on price impact and skew change, balancing markets and incentivizing neutrality. |
| Trading Fees | Tiered and volume-based; referral discounts / rewards for the first $25M in trading volume. | 2.5 bps taker fee; 0.2 bps maker rebate. Referrers earn 10% of referees’ taker fees. | Position-dependent fees to balance skew. Positions reducing skew incur lower fees; those increasing skew pay more. |
| Unique Fees | Maker rebates continuously paid on each trade. | CVA (penalty for liquidation risks);
negotiated funding rate during intent requests. |
Dynamic fees / rebates based on skew change. |
| Incentives | Maker rebates; referral rewards.
Hyperliquid offers a 4% fee discount for referred traders and referrers earn 10% of their taker fees. |
Flat fees with referral incentives. IntentX rewards traders with a 15% xINTX bonus on trading volume and affiliates with a matching 15% bonus. Affiliates also earn 25% of trading fees daily in xINTX. | Rewards trades that reduce skew.
External incentives funded by Arbitrum. |
fee, CVA, funding rate, and spread
considerations.
negotiated funding rate during intent
requests.
Hyperliquid offers a 4% fee discount for referred traders and referrers earn 10% of their taker fees.
External incentives funded by Arbitrum.
The Perennial Stack
Perennial’s architecture has undergone substantial evolution, progressing through multiple iterations—V1, V2, and V2.3—each improving key areas such as market creation, liquidity management, and user experience, with a focus on scalability and efficiency.
Perennial V1 laid the foundational infrastructure by introducing permissionless market creation, exceptional composability offering third-party projects the ability to create any structured product, with any payoff curve as long as it can be derived from an available price feed, and a shared liquidity layer that allowed liquidity providers to customize their MM experience through advanced LP solutions. Although the features introduced with the first iteration were promising, the initial version faced limitations. Long execution times, high fees, and inadequate incentives for market makers hindered its scalability and adoption, despite its original design.
To overcome these challenges, Perennial V2 redefined the protocol’s framework by focusing on improving capital efficiency and optimizing liquidity utilization. The most notable advancement was the introduction of a mechanism that netted long and short positions within each market, reducing the liquidity providers’ obligation to cover total exposure.
Unlike other AMMs where LPs underwrite all OI at any given time, LPs on Perennial only need to cover the imbalance. This shift enhanced capital efficiency dramatically, making markets up to an order of magnitude more efficient in handling open interest relative to the liquidity provided.

The latest update, V2.3, introduces Intent-Based Trading, enabling traders to specify exact prices for executing trades, similar to limit orders.
Intent orders are basically users-signed messages representing a desired outcome (the input assets & related amounts) instead of a specific path and which allows solvers to execute a trade on their behalf using their specified parameters.
This new type of order has taken hold in the DeFi landscape, both on the spot and perp side, because it has a number of key features that are worth knowing about.
Intents are:
- Flexible in creating, editing and deleting.
- MEV-resistant.
- Suitable for integrating Omnichain functionality.
- Cost-effective.
- Designed to be fully customizable and composable.
- Capital-efficient.

Solvers are the backbone of intent-based trading platforms, bridging the gap between user intent and the technical complexity of trade execution: they are trade path optimization algorithms that take the user’s intent, query multiple liquidity sources (including on-chain and off-chain order books, AMMs, and dex), and identify the best execution strategy. These software on perp intent centric platforms are usually operated by Market Makers who can satisfy user intent internally (leveraging private inventories) and hedge delta risk on other trading venues.
Intent orders on Perennial are settled directly by solvers into the AMM, ensuring precise execution without slippage. While the feature relies on solvers for filling trades, users retain the option to interact directly with the AMM for several reasons: the AMM can provide better pricing when solver liquidity is insufficient, remain accessible if solver infrastructure is unavailable, support liquidations when solvers are unwilling or unable to act, and enable advanced composability for complex trades (such as structured products or vaults).
The update also incorporates Collateral Accounts and Intent Trigger Orders, allowing traders to authorize keepers to execute trades and manage collateral via signed messages. These features streamline the trading process and allow integrators to offer one-click trading functionality, improving the overall UX.
In addition to these new features, V2.3 improves existing functionalities.
Oracle Fee Adjustments dynamically align settlement fees with chain gas prices, ensuring balanced keeper compensation. Oracle providers can now claim fees separately via the SubOracle Factory, creating a more transparent fee system.
Market Fee Improvements separate trading fees and price impact calculations, offering users clearer insights into costs incurred during trades.
Fee Unwrapping enhancements allow users to batch and convert their market fees from $DSU to $USDC atomically, simplifying fee claims through the MultiInvoker contract.
On-chain metadata improvements eliminate the need for off-chain sources to map market names, oracle IDs, and sub-oracle addresses. With these mappings now fully available on-chain, data integrity and accessibility are significantly enhanced.
These updates collectively advance Perennial’s goal of creating a capital-efficient and user-friendly trading platform. By refining its infrastructure and features, V2.3 enhances functionality, transparency, and usability for both individual traders and integrators.
Base Rollup
Perennial now operates on its own dedicated rollup (an L3 atop Base, which is an Optimism-stack L2) that users can bridge to via Decent, Relay or Superbridge.

To add the chain: Chain ID: 1424
Explorer: https://explorer.perennial.foundation
Public RPC: https://rpc.perennial.foundation
This proprietary chain is purpose-built for perpetuals trading. By controlling the execution environment, Perennial achieves several advantages:
- Lower Latency & Faster Execution: Blocks on dedicated chains can be produced faster than on a general L1 or L2, reducing the time to finality for trades. This enables near-instant order matching and confirmation, approaching the feel of centralized exchanges. Perennial’s move mirrors dYdX’s shift to a sovereign chain in pursuit of higher throughput.
- Cheaper Transactions: Operating on an app-specific rollup means users pay little to no gas for each trade. Instead of Ethereum’s typical transaction fees, Perennial can charge only a small trading fee for executed orders.
- Better User Experience: The dedicated chain’s optimizations (faster execution and lower costs) directly translate to a smoother user experience. Trades can be executed with “one-click” simplicity and confirmed almost immediately, without users worrying about slow network times or high fees.
- Scalability: Because no other dApps share this rollup, Perennial can scale usage without competing for block space with unrelated contracts. High trade volumes or complex strategies can be handled as the chain’s full capacity is devoted to Perennial’s transactions. This isolation means better performance at scale, as throughput can grow with the protocol’s demand.
Perp.fun Frontend
Perp.fun is a streamlined user interface aimed at both casual and professional traders seeking fast execution with minimal friction. It provides a clean, streamlined experience where users can execute trades with minimal steps, leveraging the “one-click trading” capability that Perennial introduced alongside Intents.

Casual traders get reduced friction (e.g. intuitive order forms, single collateral account across all trades, and quick confirmations), while professional traders benefit from performance improvements (fast updates, potentially integrated charts and advanced order types, all running on the high-speed rollup). By offering Perpfun in addition to API integrations and third-party UIs, Perennial caters to a broad spectrum of users without compromising on its core features tailored for delivering a CEX-like trading experience in a decentralized manner.
Intent-Based Trading Mechanism
As we have seen above, Perennial’s intent mechanism enables makers and takers to match orders off-chain via solvers, with all trades ultimately settled on-chain through the Perennial Automated Market Maker (AMM). This model eliminates the need for traditional liquidity pools to back multiple markets simultaneously and gives value back through fee payments to the chain on which it is deployed (unlike other protocols building their L2 appchains that extract value and liquidity out of the original chain). Instead, takers’ orders are matched with opposite maker orders, functioning similarly to a Central Limit Order Book (CLOB). For instance, a taker opening a long position can be matched with a maker closing a long or opening a short. This approach enhances liquidity and efficiency by directly aligning complementary intents.
Intent Order-Book Functionality
Market makers post arrays of order sizes and prices to a Solver API, ensuring that takers receive guaranteed execution at quoted prices. This setup mimics traditional order books, providing real-time visibility into depth and pricing while reducing latency by avoiding round trips to makers after an order is placed. The aggregated data supports transparency and enables integrators to display quasi-order books in user interfaces.
Collateral Management and Accounts
Perennial employs Collateral Accounts to simplify collateral handling, allowing users to manage funds across markets through signed intents, which can be batched for efficiency. These accounts are optional but enhance operational flexibility. Each user’s account is deterministically linked to their externally owned account (EOA). Interactions are conducted using EIP-712 messages submitted via relayers, streamlining processes such as compensation for keepers who maintain system functionality.
Perennial uses Digital Standard Unit (DSU) as its collateral, a wrapped version of USDC designed to combine USDC’s stability with enhanced censorship resistance. Users primarily interact with USDC, with automatic wrapping and unwrapping to DSU managed by the platform.
Trigger Orders and Execution Flow
Trigger orders enable users to set predefined conditions for trade execution, such as specific price limits, which are stored on-chain and executed by keepers when triggered. The execution flow involves continuous pricing updates by market makers, valid for up to 30 seconds, allowing traders to interact with real-time data for reduced latency. Orders are aggregated by the Solver API and relayed to makers, ensuring timely and efficient execution.

Fill Rate and Limitations
Perennial ensures high execution reliability and responsiveness through a Request for Quote (RFQ) flow, where market makers are expected to maintain a minimum fill rate, such as 90%. Failure to meet this standard may harm a maker’s reputation and reduce their prioritization within the network.
Users cannot hold both long and short positions in the same market from a single wallet. Switching from a long to a short position requires first closing the existing position. These limitations currently necessitate separate wallets for managing net-long and net-short positions but are expected to be addressed in future updates.
Liquidations
Liquidations are triggered when a position’s collateral falls below its maintenance requirement. During liquidation, the account is locked, preventing further trades or liquidations until the process is resolved.
The maintenance requirement is calculated based on the maximum exposure, assuming 100% utilization of a position, ensuring worst-case scenarios are always covered. Liquidators are incentivized through fees derived from the collateral, calculated as a percentage of the maintenance requirement and capped by predefined bounds.
Efficiency and Market Impact
Perennial’s architecture significantly reduces capital requirements through mechanisms like netting opposing positions, intent-based matching, and real-time pricing aggregation. This improves liquidity utilization and provides competitive execution with minimal slippage and reduced operational complexity. The intent-based model allows for composability, enabling integrators and developers to build new solutions on the protocol’s foundation.
Conclusion
Perennial introduces a hybrid model for decentralized derivatives trading, combining Automated Market Maker (AMM) efficiency with intent-based trading mechanisms to address key inefficiencies in existing DeFi models.
Its design emphasizes capital efficiency through netting long and short positions, reducing liquidity demands on Liquidity Providers. The integration of intent-based orders facilitates flexible and precise trade execution while leveraging solvers for liquidity optimization and aggregation across multiple sources.
The platform’s architecture supports modular, permissionless market creation, enabling the development of any derivative product as long as its payoff can be derived from a price feed. By implementing features like collateral accounts, dynamic fee adjustments, and intent-triggered orders, Perennial enhances both operational flexibility and user experience. However, the reliance on solvers and oracles could also introduce risks of centralization and further external dependencies.
Perennial’s technical innovations position it as a scalable infrastructure layer for on-chain derivatives, offering composability and customization for developers while improving efficiency and reducing operational complexity for users.
References
Perennial’s official documentation.
Perennial’s official Medium blog.
“Insider Playing #4” (2024), published by Rakki Otoko on his Substack page.
“GMX, A Can-Be-Evil “AMM”” (2022), published by Deri Protocol on its Medium page.
“The Beautiful Equilibrium of GMX” (2022), published by Alex (@thiccy) on his Substack page.
Drift’s official documentation.
Symm.io’s official documentation.
Disclosures
Alea Research is engaged in a commercial relationship with Perennial and this report was commissioned as part of that engagement.
Members of the Alea Research team, including those directly involved in the analysis above, may have positions in the tokens discussed.
This content is provided for educational purposes only and does not constitute financial or investment advice. You should do your own research and only invest what you can afford to lose. Alea Research is a research platform and not an investment or financial advisor.


