Beyond Directional Exposure: The Relative Value Primitive

A digital asset can be modelled with regulatory precision and still be mostly an assumption because absolute pricing is a wrong starting point. In equities, valuation has a frame – cash flows, margins, cost of capital and terminal growth. Models can still be wrong, but the exercise is bounded. In crypto, that frame is much looser.

A DCF (discounted cash flow model, a valuation method projecting future cash flows and discounting them to present value) on any major liquid token can be made to say almost anything. Terminal value, fee capture, monetary premium, token velocity, emissions and regulatory discount can move an output by orders of magnitude. High-confidence dollar valuations for crypto assets are structurally unreliable.

Relative value asks for less and often says more. It does not need to know what ETH is worth in dollars. It asks whether one asset is cheap or expensive against another asset with a comparable economic role, buyer base, narrative or risk factor. It needs to know whether ETH (Ether, the native token of the Ethereum blockchain) should trade rich or cheap against SOL (Solana, a high-throughput layer-1 blockchain). It does not need a perfect fair value for HYPE, the native token of Hyperliquid. It needs to know whether HYPE deserves to outperform another execution venue given liquidity, revenue capture, builder mindshare, user alignment and ecosystem momentum. The target no longer depends on market level but on the relationship between two trading legs.

That is why a pair spread is a cleaner expression of belief than outright valuation. A long-only trade says an asset should go up. A long/short trade – long an asset and simultaneously short another – says one asset should outperform another, a statement that carries more information, forces a cleaner comparison set and survives more market environments. If the whole market falls, but the long leg falls less, the spread works. If the market rises but the long leg compounds faster, the spread works again. The trade becomes less dependent on calling the market and more dependent on understanding relative strength, pocketing gains across different market conditions by focusing on asset relationships rather than market direction.

The intellectual foundation is established. Gatev, Goetzmann and Rouwenhorst (2006) documented the opportunity across nearly four decades of US equity data as the systematic exploitation of "temporary mispricing between close economic substitutes", with pairs portfolios nearly factor-neutral by construction. Crypto generates those substitutes continuously – competing ecosystems, overlapping narratives comparable risk factor exposures – and dislocates them faster and more violently than equities ever did.

Crypto no longer rewards beta 

In previous cycles, the playbook was simple: Bitcoin breaks out, ETH follows, liquidity rolls into higher-beta majors and the long tail rips. Direction mattered more than discrimination. Since 2023, that dynamic has broken down. The market is more fragmented, more narrative-driven and more dispersion-heavy. Some assets can rally while majors bleed. Breadth can narrow while isolated pockets of strength expand.

Understanding why requires decomposing what actually drives returns in this environment. Every asset's return now has two components. The first is systematic – dollar liquidity, Bitcoin direction, perp funding shocks, ETF flows, macro risk appetite and collateral stress. It hits everything simultaneously. The second is idiosyncratic – valuation, revenue, protocol growth, execution quality, ecosystem momentum, token supply and narrative attention. Over long horizons, idiosyncratic factors dominate. Strong assets separate from weak ones. Over short horizons, the systematic component overwhelms everything else.

Those shocks matter because crypto is structurally path-dependent. Violent moves trigger liquidations, which push prices further, which mark collateral lower, which trigger the next wave. The loss does not come from being wrong on fundamentals. It comes from being unable to survive the path between entry and thesis realization.

Pair trading lives inside that gap. The long leg carries the idiosyncratic view. The short leg hedges part of the shared market component. When volatility spikes and assets move together, the short leg offsets losses on the long leg. When the panic fades and assets separate again, the long leg has room to outperform. Correlations rise precisely during market stress – making the hedge most valuable exactly when outright exposure becomes hardest to hold.

Pair trading does not eliminate risk but changes it considerably. The investor no longer owns the full market direction – only the spread between two assets. That spread can still move against the position. But the market's common shock carries less power to kill the trade before the thesis has time to work.

(Source: TradingView, metrics compiled by Sandmark as of 6 Mar)

The house of onchain pair trading

Pear Protocol (a non-custodial onchain execution layer purpose-built for pair, basket and relative value trades) compresses all the operational complexity of relative value into one execution surface. What started as a feature built on top of GMX on Arbitrum has become a standalone execution layer, enabling simultaneous long and short positions through a single unified interface.

The core insight is execution. Running a long/short position manually across two assets means managing two separate positions, two funding streams, two liquidation risks and a spread that exists nowhere on your screen. Pear collapses that into one interface. The PnL shown is the spread. The chart is the ratio. The stop-loss triggers on the relationship, not on either leg in isolation. And the protocol does not let you open a trade without defining a pair, a ratio and, implicitly, a thesis. That constraint forces the discipline that separates relative value from speculation.

The rebalancing mechanic is where the product earns its sophistication. As prices move, the initial weights drift – the long leg outgrows the short, or the short compounds against you, and what started as a relative trade quietly becomes a directional one. Pear's rebalancing snaps the position back to its original ratio before that drift becomes the trade. Beta weighting extends the logic further: legs are sized by market sensitivity, not notional equality, so the hedge actually hedges.

And the construction goes beyond pairs. Baskets – long HYPE, ZEC, BTC against short ETH, ARB, ZK, or a broader cohort – or even indices let traders express sector-level views with the same spread mechanics. The short leg of a basket dampens the systematic component across multiple assets simultaneously. It is closer to how a crypto hedge fund would construct a relative value book than anything previously available onchain.

Despite early integrations including GMX, Pear execution mostly runs on Hyperliquid and Symmio's intent-based architecture – tight spreads, deep liquidity and best-in-class execution, onchain and non-custodial. Institutional-quality strategy construction on institutional-quality execution rails.

As of 4 Jun, Pear Protocol has processed $1.69bn in cumulative volume since launch. In 2025, it delivered $853mn and 2026 is already tracking ahead of that pace with $344mn in the first five months. May crossed $100mn for only the second time in protocol history, against a backdrop of broadly declining perpetual volumes across the space.

Since its 25 Jun integration, Hyperliquid has accounted for 86% of all volume, clearing $868mn of the $1.01bn transacted on Pear since then – with GMX and Vertex essentially zeroed out and Symmio's accounting for the residual share. Since 26 Mar, Hyperliquid execution venue accounts for ~98% of the traded volume on Pear Protocol. 

(Source: Pear Protocol's Dune Dashboard, as of 6 Apr)

Users are at their highest levels since launch. Daily active users hit 139 on 2 Jun – an all-time high – with June 2026 averaging 118 per day, roughly double the Q4 2025 run rate. The catalyst is the Pear Agent launch, deployed around a HYPE-incentivized campaign, which pulled a new cohort of traders into the interface. Pear Agent is an AI-powered trading signal bot and conversational assistant developed by Pear Protocol, which continuously monitors market conditions and identifies high-probability pair trading opportunities based on statistical signals.

New user additions, which were effectively zero through most of 2025, have started printing 10 new users per day consistently over the past couple of weeks. The funnel is opening as the protocol's feature set matures.

(Source: DeFiLlama, as of 6 Apr)

Revenue follows the volume curve with a lag. As Hyperliquid concentration deepened and OI compounded through early 2026, the revenue line responded: March at $77k, April at $68k, May at $85.5k – the strongest month since October 2024's all-time high of $106k, and up 216% on October 2025 and more than 3x January's $28.9k.

June is already tracking above that trajectory, opening at pace – $25.4k in the first five days, nearly 30% of May's full month – putting the protocol on track for a new monthly all-time high if the run rate holds. At a daily run rate of ~$6.3k, it implies a monthly ~$190k revenue print, more than doubling May's headline and establishing a new all-time high for the protocol.

Open interest (OI) has tripled in three months – from $3.7mn in late March to a peak of $9.6mn earlier this week, an all-time high. The OI composition is more informative than the headline. The two largest buckets are structured, including a long DeFi / short underperforming Layer-1s and memecoins position at $1.65mn OI, and a long/short tokenized stock basket at $1.24mn – the latter enabled by HIP-3 markets (Hyperliquid Improvement Proposal 3, which allows qualified builders to permissionlessly deploy custom perpetual futures markets) and a signal that sophisticated capital is using Pear's infrastructure to express views well beyond crypto. Capital flowing into Pear is being deployed in exactly the structured relative value strategies the protocol was designed for, not simple directional bets.

(Source: PearData, as of 6 May)

Long/short the probability curve 

The most forward-looking product on Pear's roadmap – and the first of its kind – asks a simple question: what if the trigger for a long/short trade was not a price level, but a probability threshold?

Prediction markets have matured into precise, continuously updating probability-pricing machines. Pear is integrating those probability curves – Kalshi, Polymarket (a leading decentralized prediction market platform) and HIP-4 Outcome Markets – directly into its trigger configuration. You pick the event, you set the probability that makes you act and you define the trade. Pear watches the odds and executes the moment your condition is met.

If the probability of Hormuz closing crosses your threshold, long oil. If SpaceX's implied valuation drops below $2tn, short Tesla. If NVDA earnings beat with sufficient confidence, long a basket of tokenized semis and short whatever is historically inversely correlated. The position is armed before the headline hits. By the time the chart has moved, the trade is already on.

There is a second-order dimension worth noting. Prediction market curves move ahead of public information – informed flow rushes to reprice odds before news reaches a broader audience. Being positioned on the right side of a rising probability curve is, in effect, a structural hedge against the market moving before you can react.

Tokenized equities and commodities extend the spectrum further. The universe of triggerable relationships is no longer crypto-only – basket trades on equities or commodities, conditional on event probabilities, executable onchain with institutional-grade rails. 

PEAR's evolving tokenomics

The PEAR token was launched on Arbitrum, a natural starting point given the protocol's origins as a feature built on top of GMX. But the product has since outgrown that foundation. With volume, users and revenue now almost entirely concentrated on Hyperliquid, keeping the token on Arbitrum creates a structural disconnect between where the protocol lives and where value accrues. The planned migration to Hyperliquid closes that gap.

The strategic logic goes beyond venue alignment. In a discussion with Sandmark, Huf was explicit about why Hyperliquid is the right home, citing the venue's established integrations, its loyal user base and the economic loop where traders recoup fees through token buybacks – keeping them on the platform and generating more fees. Deploying PEAR on the same stack replicates that dynamic one layer up. Traders become token holders, token holders become traders, and a self-reinforcing loop compounds at the protocol level, where 80% of protocol revenue goes to token buybacks and the remaining 20% to the treasury.

The roadmap extends the surface area further. New products on Base expand distribution beyond the Hyperliquid ecosystem, and an upcoming mini-app integration acts as a distribution play reaching Coinbase’s existing user base through an embedded experience – ensuring the protocol is positioned ahead of where attention is moving, not behind it.

From signal to allocation strategy 

Pear Agent is the first step toward a fully integrated discovery-and-execution platform – moving through sizing and correlation data, tradable universe signals and a freshly announced conversational AI trading assistant, each phase building toward the same destination: the vault. The fourth and final phase packages Pear Agent's mean-reversion and statistical arbitrage signals into a deployable instrument that anyone can allocate to, closing the loop between strategy discovery and capital deployment.

A vault infrastructure is the logical next step for the protocol. Planned to launch on HyperCore, the offering will be curated and initially restricted to vetted managers featuring flexible withdrawal policies, performance fee customization and full composability, with excess collateral lendable onchain and positions managed through Pear Pro, a dedicated long/short terminal built for professional execution.

HyperCore vaults are not smart contract wrappers or EVM primitives. They are native exchange-level pooled capital vehicles executing directly within HyperCore's onchain order book, with access to the same features as the DEX itself – liquidations, high-throughput market-making and the full execution stack. That is a fundamentally different primitive from simple token rebalancing vaults. With Pear as the execution layer on top, the vault effectively becomes an onchain hedge fund: long/short strategies, beta-weighted positions, basket construction and spread management – all accessible from a wallet. At a point when onchain yield is structurally compressed and allocators are actively searching for return streams that do not depend on directional crypto exposure, a curated vault offering built on this stack is not a nice-to-have. Asset managers like Bitwise and 21Shares are already building curator strategies, and the surrounding infrastructure has started to catch up with the demand.

The perpetual DEX market is consolidating into an oligopoly. As dominant venues scale their liquidity and execution moats, fee competition alone stops working – incentive programmes end, mercenary capital rotates out and what remains is either a sticky user base or irrelevance. The dynamic mirrors what played out in Layer-2s, with a survival of the fittest where the winners are not the cheapest but the most defensible. Hyperliquid has reaffirmed itself as the dominant venue. The rest are competing for residual flow on the same pie.

Differentiation no longer comes from fees or execution speed. It comes from a use case. The biggest breakthroughs in perpetual derivatives have come from expanding the universe of what can be traded and how – TradeXYZ proved it with RWA perpetuals and pre-IPO markets. Protocols that built a genuine niche and a specific enough use case to grow a sticky user base are the ones still standing when incentives run out.

Pear Protocol squares neatly into that category and goes one step further, sitting at the crossroads of two of the most significant financial innovations of the decade. Prediction markets have become the most precise probability-pricing tool ever built, and perpetuals remain the most capital-efficient synthetic exposure instrument in existence, pinning spot to futures efficiently through funding rates around the clock. Both are now attracting institutional inflows at scale, and for the first time, crypto products built for retail are being adopted by institutions rather than the other way around.

A product that combines both, atop institutional-grade execution rails and the infrastructure depth of a relative value fund, has arguably a whole new market to tap into.