Synthetix: How to outperform CEXs

Discover how Pyth powers Synthetix Perps with secure, cost-competitive, and user-friendly trading across Optimism and Base. After implementing Pyth oracles in Perps V2, Synthetix's trading volumes soared from $5B to over $50B in just over a year.

“We've been able to reduce fees to 5 to 10 basis points, been able to increase open interest limits, support 23 markets as of right now…users are getting the best perps trading experience out there, in a fully decentralized and trustless protocol.” — Matt Losquadro, Core Contributor

Context: Establishing truly competitive on-chain trading

Perpetual futures have become a core product in crypto markets, attracting users chasing leverage, arbitrage, and exposure to trending assets. In this space, DeFi protocols compete directly with both decentralized and centralized platforms—on fees, liquidity depth, and market breadth.

For Synthetix, success required more than self-custody and on-chain execution. To capture market share, it needed to match CEX-level performance while supporting a wide range of hot and exotic markets. Pyth’s on-demand prices and historical data (Benchmarks) enabled Synthetix to launch Perps V2 with taker-maker fees as low as 2–15bps—unlocking speed, cost-efficiency, and scale across DeFi.

Challenge: Breaking the security, cost, and user experience trilemma

Synthetix faced a core tradeoff between fees and price update latency.

Higher Fees — neutralizes front-running but imposes higher costs on traders.

Lower Fees — necessary for competitiveness and alignment with CEX-style fees.

Price Update Speeds — lower latencies support greater protocol security.

The first iteration of Synthetix Perps utilized legacy push oracles which introduced slow price updates, which opened the protocol up to front-runners. Fees were raised to neutralize this issue, but at the cost of user experience, namely poor pricing for traders. Cost prohibitive fees (50bps and higher) restricted initial adoption of Synthetix Perps.

A shift to the Pyth pull oracle design was proposed: empowering protocol users to request or pull the newest price update on-chain addresses the front-running issues.

Finally, the Synthetix protocol required accurate price reporting to ensure that its perpetual futures trade at prices that truly reflect the markets. This required a reliable source of high-resolution data that matches the price discovery happening in the real world.

Solution: Making Synthetix Perps competitive

The launch of Perps V2 marked a major shift to off-chain pricing with on-chain verification, powered by the Pyth price layer. When a user initiates a trade, a reference price is pulled off-chain using timestamped historical data (Benchmarks). That price is later validated on-chain at execution—enabling both precision and transparency.

This transition eliminated the need for cost-prohibitive fees used to deter front-running in Perps V1. As a result, Synthetix now offers some of the most competitive perps fees in DeFi—ranging from just 2 to 15bps across crypto, forex, and commodity pairs. With Pyth’s low-latency pricing infrastructure, Synthetix Perps V2 delivers a trading experience that rivals centralized platforms while remaining fully on-chain.

Impact: The SNX ecosystem empowered

Since integrating Pyth and transitioning to off-chain oracles via SIP-285, Synthetix Perps has experienced rapid adoption and protocol growth. Trading volumes surged from $5B to over $40B within twelve months—a 700% YoY increase.

Today, Synthetix Perps offers unmatched capital efficiency, with up to 50x leverage—exceeding the 10–20x ceilings of many competing platforms. Maker-taker fees are as low as 2–15bps for crypto, and just 1–2bps for gold and silver.

The protocol has now fully migrated from legacy oracles, with the Andromeda release launching Synthetix V3 and Perps V3 entirely powered by Pyth. This upgrade brought improved pricing accuracy, MEV-resistant liquidations, cross-margining, and NFT-based account control—cementing Synthetix’s leadership in on-chain derivatives.