In this article, both hard liquidity support models and synthetic models are discussed.

By Chaos Labs

Compiled by: TechFlow

Written by Research Analyst @0xGeeGee

In both traditional finance and cryptocurrency, the scale of the derivatives market far exceeds that of the spot market. For example, as of now, the daily spot trading volume of Bitcoin is about US$4 billion, while its derivatives trading volume is as high as US$53.89 billion (data source: Cryptoquant.com).

Bitcoin: Volume Ratio (Spot vs Derivatives) — Source: CryptoQuant

This trend has accelerated since the beginning of 2021 and has continued to this day. The derivatives market in traditional finance has long surpassed the spot market, and the derivatives market in the centralized exchanges (CEX) of cryptocurrencies is following closely. In the field of decentralized finance (DeFi), derivatives have not yet surpassed the spot market of decentralized exchanges (DEX). For example, in the past 24 hours, @Uniswap v3 facilitated $1.3 billion in spot transactions, while @HyperliquidX processed about $1 billion in derivatives transactions (data source: Coingecko Data).

Nevertheless, the gap is narrowing, and it is clear that as the ecosystem matures, on-chain derivatives may eventually surpass the spot market, just like other mature markets. Although market demand is tilting towards derivatives, this growth needs to be supported by secure and efficient trading platforms and models.

Derivatives trading volume — Source: DefiLlama

Understanding the different models that underpin derivatives markets is critical to building the infrastructure to support this shift. In this article, I will discuss hard liquidity support models and synthetic models.

Hard Liquidity Support Model

In the hard liquidity backing model, traders trade with real assets, tokens, or stablecoins deposited in liquidity pools. These assets are actually loaned to traders to open margin positions. @GMX_IO, @JupiterExchange, @GearboxProtocol's PURE, and @Contango_xyz are some examples of this approach.

Liquidity providers (LPs) earn trading fees by depositing hard assets and may be rewarded as counterparties to traders. Therefore, LPs' earnings depend on the performance of the assets in the pool, the utilization rate of the pool, and in a model where there is no mechanism to balance long and short trading volume, the profits and losses of traders.

advantage:

Lower risk of bankruptcy: Since transactions are backed by real assets, the risk of system bankruptcy is lower.

DeFi composability: Hard-backed models such as GMX and Jupiter allow re-hypothecation of liquidity pool tokens: $GLP and $JLP tokens can be used as collateral or staked in other DeFi applications, improving capital efficiency.

Lower trading/market making incentive requirements: Since LPs act as counterparties or market makers, direct incentives are less important. Although LPs are usually rewarded through token incentives in the early stages, in the long run, the return on providing liquidity comes mainly from trading fees, reducing the difficulty of designing a balanced trading incentive plan.

Deepening market liquidity: The hard-backed model facilitates the deepening of market liquidity by requiring a basket of liquidity backed by real assets. Over the past few years, this has also made protocols like GMX one of the most efficient places to exchange spot assets, as liquidity is concentrated in pools that can serve both derivatives and spot markets.

From the screenshot of DefiLlama, we can see the number of protocols and pools including GLP and JLP earnings.

Within this category, different sub-models emerge based on how liquidity is captured and shared:

GMX v1 and Jupiter: These protocols use a global shared liquidity pool, where all assets are pooled together. This model ensures deep liquidity and enhances composability by allowing liquidity providers to use a single token in different DeFi protocols.

GMX v2 and Gearbox’s PURE: Introduced isolated liquidity pools with a modular architecture, with each asset or market having its own liquidity pool. This reduces the systemic risk of the protocol, enabling it to support longer-tail, higher-risk assets. The risk and return of each asset are independent, preventing a single asset from affecting the liquidity of the entire protocol and forming different risk/return characteristics.

In this “hard liquidity support” model, we can also see Contango in action. While it is not a standalone model, Contango runs on top of existing lending protocols such as Aave to provide a margin decentralized exchange experience. It uses real assets borrowed from lending pools and flash loan functions to create leveraged positions.

Synthetic Model

While hard liquidity-backed models guarantee security and composability by requiring real assets as collateral, synthetic models take a different approach.

In synthetic models, trades are typically not backed by real assets; instead, these systems rely on order book matching, liquidity vaults, and price oracles to create and manage positions.

There are a variety of synthetic model designs - some rely on peer-to-peer order book matching, with liquidity provided by active market makers, which can be professional or managed through algorithmic vaults, and liquidity can be globally shared or market-segregated; others take a purely synthetic approach, with the protocol itself acting as the counterparty.

What is a Liquidity Vault?

In the synthetic derivatives model, a liquidity vault is a centralized liquidity mechanism that provides a source of funds needed for trading, whether directly supporting synthetic positions or acting as a market maker. Although the structure of liquidity vaults may vary slightly from protocol to protocol, their main purpose is to provide liquidity for trading.

These liquidity vaults are usually managed by professional market makers (such as Bluefin stablecoin pool) or algorithms (such as Hyperliquid, dYdX unlimited, Elixir pool). In some models, they are purely passive counterparty pools (such as Gains Trade). Usually, these pools are open to the public, allowing the public to provide liquidity and receive rewards for participating in platform activities.

Liquidity vaults can be shared across listed markets, like in Hyperliquid, or partially segregated, like in @dYdX unlimited, @SynFuturesDeFi, and @bluefinapp, approaches with similar risks and benefits to those mentioned previously.

Some protocols, such as Bluefin, adopt a hybrid model that combines a global liquidity vault managed by market makers and segregated algorithmic pools.

In synthetic models, liquidity is usually provided by a combination of active users (peer-to-peer matching), liquidity vaults (as backup), and market makers who quote bids and asks on the order book. As mentioned earlier, in some purely synthetic models, such as @GainsNetwork_io, the liquidity vault itself serves as the counterparty to all trades, eliminating the need for direct order matching.

Advantages:

The synthetic model has different trade-offs than the hard liquidity backing model, but also brings a range of advantages:

Capital Efficiency: Synthetic models are highly capital efficient because they do not require direct 1:1 physical asset backing. The system can be run with fewer assets as long as there is enough liquidity to cover the potential outcomes of active trading.

Asset flexibility: These systems offer more flexibility in the assets they trade, as positions are synthetic. There is no need to provide direct liquidity for each asset, which allows for greater diversity in trading pairs and the ability to list new assets more quickly — even semi-permissionedly.

This is especially true in Hyperliquid’s pre-launch markets, as the assets traded in these markets don’t even really exist yet.

Better price execution: Since trades are purely synthetic, it is possible to achieve better price execution, especially when market makers are active on the order book.

However, these models also have some significant shortcomings:

Dependence on Oracles: Synthetic models are highly dependent on price oracles, which makes them more susceptible to related issues such as oracle manipulation or delays.

Lack of liquidity contribution: Unlike hard-backed models, synthetic trading does not contribute to the global spot liquidity of the asset as liquidity only exists in the derivatives’ order book.

Although decentralized exchanges still account for a smaller share of overall perpetual swap volume than centralized exchanges (about 2% of the market), diverse models are laying the foundation for real growth in the future. The combination of these models, coupled with continued improvements in capital efficiency and risk management, will be key for decentralized exchanges to expand their share of the derivatives market.

Perpetual contract trading volume distribution - Source: GSR Annual Report

Chaos Labs Contributions

Chaos Labs plays an important role in risk management of hard-backed liquidity and synthetic models, meeting the specific needs of our partner platforms such as @GMX_IO, @dYdX, @SynFuturesDeFi, @JupiterExchange, @OstiumLabs, and @Bluefinapp.

As a long-time risk analytics provider, Chaos Labs helps protocols manage leverage limits, liquidation thresholds, collateral requirements, and overall platform health through real-time risk assessments and simulations.

Chaos Labs’ latest product, Edge Network, introduces a decentralized oracle system that helps mitigate oracle-related risks, ensuring that both synthetic and hard-backed models benefit from real-time, accurate price data. Edge is already being used as a primary oracle by well-known platforms like Jupiter.

Chaos Labs also works with partners to develop optimized liquidity incentive plans to ensure a smooth trading experience and attract more liquidity.

Finally, Chaos Labs also provides public dashboards for monitoring risk parameters for platforms such as GMX, Jupiter, Bluefin, and dYdX.