Can Solana Catch Hyperliquid in Perps? | Luke & Sam
Monday, 25 May 2026 · 4 min read · Listen to the episode ↗
Luke and Sam examine whether Solana can close the gap on Hyperliquid's dominant position in perpetuals trading, tracing the structural reasons Hyperliquid leads, including purpose-built infrastructure and ingrained maker protections that Solana's general-purpose VM lacks.
Hyperliquid holds dominant market share in perpetuals trading and that lead appears to be compounding, reflected in its token price. Sam attributes this to purpose-built infrastructure, contrasting it with Solana's general-purpose VM and leader-based architecture, which lacks ingrained maker protections and exposes market makers to toxic flow with no priority guarantee. Jito's BAM client is the most prominent effort to address this gap, ordering transactions by fees per compute unit to provide more consistent execution. BAM currently holds around 30 percent of stake, with the standard Jito client holding another 30 percent. The maker priority plugin, the first BAM plugin, launched approximately two weeks before the episode and targets spot markets, with a roadmap of five to ten plugins extending to perps. Whether the plugins add latency is uncertain, though Sam's understanding is that execution stays within the 400-millisecond block window.
Phoenix is considered one of the strongest engineering teams building perps on Solana and has constructed a high-performance perps platform fully within the Solana VM, something that had not been done before. Phoenix shows meaningful improvement over Drift, the closest prior attempt using a central limit order book, but still lags Hyperliquid and Pacifica, both of which are effectively off-chain platforms. Phoenix has been developing its architecture for about a year and has not yet proven it can match off-chain performance while remaining fully within the VM. Luke noted that some teams claimed the main technical challenges were solved a year ago, yet the BAM maker priority plugin only recently shipped, suggesting slow execution on a trend running roughly 18 months. The central question is whether Solana can create an ecosystem where multiple viable Hyperliquid competitors coexist on one chain, with one view being it may be too little too late and another that ongoing development could still close the gap.
Hyperliquid has partially escaped the cyclicality of crypto trading volumes through HIP-3 listings covering real-world assets, stocks, commodities, and pre-IPOs, with execution quality on those instruments able to rival centralized venues for small and mid-sized traders. Hyperliquid's pre-IPO listings have outperformed similar efforts by other platforms, and timely macro events including a silver trading surge, oil volatility, and the Iran conflict drove additional volume. On the regulatory front, US exchanges are currently prohibited from offering perpetual futures to domestic customers. If the CFTC regulates perps in the US, Luke predicts significant activity could flow to Hyperliquid, though it may need to incorporate a KYC model to capture that regulated flow.
Jupiter announced a deal with Securitize and Jump Trading that Sam said could bring high-quality execution for tokenized equities on-chain to Solana. Luke noted that AlphaQ was quoting pre-stock tokenized equity markets on Solana and then stopped based on current incentive structures, and that retail traders in those markets are not effectively arbitraging prices back to Nasdaq. BlackRock and large asset managers are signaling tokenization of stocks and equities pending regulatory clarity, with Luke seeing tokenized spot activity potentially going to Solana given its existing prop AMM infrastructure, though incumbents like Nasdaq and NYSE spinning up nodes remain a competing path. In Solana's prop AMM market, GunFi held around 60 percent market share at the end of last year and has since fallen to around 20 percent, with BuySumFi surpassing it on volumes.
Streth has grown approximately 100 percent from five billion to ten and a half billion dollars over the past two months and has been the primary facility through which Strategy has financed new Bitcoin purchases. Tokenized Streth on Pendle has risen from zero percent of Pendle total TVL at the start of March to close to 27 percent, making Streth-related instruments the largest listings on Pendle and surpassing Athena, USDAI, Khype, and USDG. Streth yield is approximately 11.5 percent, framed as a new benchmark hurdle rate for on-chain opportunities. Two primary issuers compete in the tokenized Streth marketplace: Apex issuing APX USD and APY USD, and Saturn issuing USD80 and S USD80, with Apex the front runner by total deposits and APY USD implied yield priced at 17 to 19 percent fixed. Three primary risks to Streth participation are Bitcoin's compound annual growth rate persistently falling below the dividend obligation, suspension of the dividend payout, and the dividend rate being toggled down, any of which could push Streth well below its par value of 100.
Despite Streth listings driving Pendle's share of deposits from zero to 27 percent, Pendle's overall TVL has continued to decline. Luke characterizes Pendle V2 as primarily an incentive marketplace where the strongest listings are almost always accompanied by significant incentives, making the protocol cyclical and tied to typical crypto market structure. Sam reframed this as cyclical rather than terminal, and Luke conceded Pendle V2 is not necessarily doomed because airdrops will remain a persistent theme in the on-chain economy. Luke argues Boros is better positioned than Pendle V2 to capitalize on high and variable yield, with monthly notional trading volumes elevated and sticky since its August launch. Luke describes the Boros total addressable market as any interest rate that can be inbounded with an oracle, with margin trading allowing notional volumes to be significantly larger than TVL-based comparisons would suggest, and notes Boros can list funding rates on assets including Nvidia, oil, silver, and gold across both Hyperliquid and Binance.
This summary was generated from the episode transcript and can contain mistakes.