Exchange or broker? Hyperliquid's business model dilemma

Title: Hyperliquid at the Crossroads: Robinhood or Nasdaq Economics

Author: @shaundadevens

Compilation: Peggy, BlockBeats

Editor’s Note: As Hyperliquid’s trading volume approaches that of traditional exchanges, what truly matters is no longer just “how large” the volume is, but which layer of the market structure it chooses to stand on. This article references the division of roles between “Broker vs. Exchange” in traditional finance to analyze why Hyperliquid actively adopts a low-fee market layer positioning, and how Builder Codes, HIP-3, while expanding the ecosystem, create long-term pressure on platform fees.

Hyperliquid’s path reflects a core issue faced by the entire crypto trading infrastructure: after scaling up, how should profits be distributed?

Below is the original text:

Hyperliquid is handling perpetual contract trading volumes close to Nasdaq levels, but its profit structure also exhibits characteristics akin to “Nasdaq-level” operations.

In the past 30 days, Hyperliquid cleared $205.6 billion in nominal perpetual contract trading volume (annualized quarterly at about $617 billion), but only generated $80.3 million in fee income, with an implied fee rate of approximately 3.9 basis points (bps).

This means Hyperliquid’s monetization approach is more similar to a wholesale execution venue rather than a high-fee retail trading platform.

By comparison, Coinbase recorded $295 billion in trading volume in Q3 2025, with $1.046 billion in trading revenue, implying an effective fee rate of about 35.5 bps.

Robinhood’s monetization logic in crypto is similar: its $80 billion in nominal crypto assets trading volume generated $26.8 million in trading revenue, with an implied fee rate of about 33.5 bps; meanwhile, Robinhood’s nominal stock trading volume in Q3 2025 reached as high as $6.47 trillion.

Overall, Hyperliquid has entered the top-tier trading infrastructure scale in terms of volume, but in terms of fee rates and business model, it resembles a low-commission execution layer aimed at professional traders rather than a retail-oriented platform.

The gap is not only reflected in fee levels but also in the breadth of monetization dimensions. Retail platforms often profit from multiple revenue “interfaces.” In Q3 2025, Robinhood achieved $730 million in trading-related income, plus $456 million in net interest income, and $88 million in other income (mainly from Gold subscription services).

In contrast, Hyperliquid currently relies much more heavily on trading fees, which are structurally compressed into single-digit basis points at the protocol level. This means Hyperliquid’s revenue model is more concentrated, more singular, and closer to a low-fee, high-turnover infrastructure role rather than a retail platform with deep monetization across multiple product lines.

This can essentially be explained by positioning differences: Coinbase and Robinhood are broker/distribution businesses that leverage their balance sheets and subscription systems for multi-layer monetization; whereas Hyperliquid is closer to an exchange layer. In traditional financial market structures, profit pools are naturally split between these two layers.

Broker-Dealer vs. Exchange Model

In traditional finance (TradFi), the core distinction is between the distribution layer and the market layer.

Retail platforms like Robinhood and Coinbase are situated in the distribution layer, capable of capturing high-margin monetization opportunities; while exchanges like Nasdaq are in the market layer, with pricing power structurally limited, and their execution services are pushed toward a commodity-like economic model due to competition.

Broker / Dealer = Distribution Capability + Customer Asset Balance Sheet

Brokers hold the customer relationships. Most users do not connect directly to Nasdaq but access the market through brokers. Brokers handle account opening, custody, margin and risk management, customer support, tax documents, etc., then route orders to specific trading venues.

It is this “relationship ownership” that allows brokers to monetize beyond trading:

• Funds and asset balances: cash sweep spreads, margin lending, securities lending
• Product bundling: subscription services, feature packages, bank cards / advisory products
• Routing economics: brokers control order flow and can embed payment or revenue-sharing mechanisms in the routing chain

This is why brokers often earn more than trading venues: the profit pool is truly concentrated in “distribution + balances.”

Exchange = Matching + Rules + Infrastructure, Limited in Fees

Exchanges operate the trading venue itself: matching engines, market rules, deterministic execution, and infrastructure connectivity. Their main monetization methods include:

• Trading fees (continuously compressed in high-liquidity products)
• Rebates / liquidity incentives (often returning most of the nominal fee to market makers to compete for liquidity)
• Market data, network connectivity, colocation services
• Listing fees and index licensing

Robinhood’s order routing mechanism clearly demonstrates this structure: user relationships are held by the broker (Robinhood Securities), and orders are routed to third-party market centers, with economic benefits distributed along the chain.

The truly high-margin layer is at the distribution end, controlling customer acquisition, user relationships, and all monetization interfaces around execution (such as order flow payments, margin, securities lending, and subscriptions).

Nasdaq itself operates in the low-margin (thin-margin) layer. Its core products are essentially highly commoditized execution capabilities and queue access rights, with pricing power strictly limited by mechanisms.

The reason is: to compete for liquidity, trading venues often need to return large portions of nominal fees as maker rebates; regulators impose caps on access fees, limiting revenue potential; and order routing is highly elastic, allowing funds and orders to switch rapidly between venues, making it difficult for any single venue to raise prices.

This is vividly reflected in Nasdaq’s financial disclosures: the net profit it captures from cash stock trading is usually only a few thousandths of a dollar per share. This directly illustrates the structural compression of profit margins at the market layer.

The strategic consequences of this low-margin environment are also clearly reflected in Nasdaq’s revenue structure changes.

In 2024, Nasdaq’s Market Services revenue was $1.02 billion, accounting for 22% of total revenue of $4.649 billion; this ratio was as high as 39.4% in 2014 and still 35% in 2019.

This persistent decline aligns with Nasdaq’s strategic shift from highly dependent on market volatility and profit-limited execution services toward more recurring, predictable software and data businesses. In other words, the structurally low profit margins at the exchange layer have driven Nasdaq to gradually shift its growth focus from “matching and execution” to “technology, data, and service-oriented products.”

Hyperliquid as the “Market Layer”

Hyperliquid’s effective fee rate of about 4 bps aligns closely with its deliberately chosen market layer positioning. It is building an on-chain “Nasdaq-style” trading infrastructure:

Centered on HyperCore, with high-throughput matching, margin, and clearing systems, employing maker/taker pricing and maker rebates, aiming to maximize execution quality and shared liquidity rather than multi-layer monetization targeting retail users.

In other words, Hyperliquid’s design focus is not on subscriptions, balances, or distribution income, but on providing commoditized yet highly efficient execution and settlement capabilities — a typical feature of the market layer, and an inevitable result of its low-fee structure.

This is reflected in two structural aspects that most crypto trading platforms have yet to fully realize but are very typical in traditional finance:

First, the permissionless broker/distribution layer (Builder Codes).

Builder Codes allow third-party trading interfaces to be built on top of core trading venues and to collect economic benefits independently. Builder fees have clear caps: perpetual contracts up to 0.1% (10 bps), spot up to 1%, and fees can be set at the order level.

This mechanism creates a competitive market for the distribution layer, rather than a monopoly of a single official app controlling user entry and monetization.

Second, the permissionless listing/product layer (HIP-3).

In traditional finance, exchanges typically control listing approval and product creation. HIP-3 externalizes this function: developers can deploy perpetual contracts inheriting HyperCore matching engine and API capabilities, with market definition and operation managed by the deployer.

Economically, HIP-3 clarifies revenue sharing between the trading venue and product layer: deployers of spot and HIP-3 perpetual contracts can retain up to 50% of the trading fees generated by their deployed assets.

Builder Codes have already demonstrated results at the distribution end: as of mid-December, about one-third of users are not trading via the native interface but through third-party frontends.

The problem is, this structure conducive to distribution expansion also exerts ongoing pressure on the fees at the trading venue layer:

1. Price compression.

Multiple frontends selling the same underlying liquidity will naturally drive prices down to the lowest combined trading cost; Builder fees can be flexibly adjusted at the order level, further pushing prices toward the lower bound.

2. Loss of monetization opportunities.

Frontends control account opening, product bundling, subscriptions, and the full trading workflow, capturing high-margin spaces at the broker layer; Hyperliquid can only retain thinner exchange layer fees.

3. Strategic routing risks.

If frontends evolve into true cross-venue routers, Hyperliquid may be forced into wholesale execution competition, only able to defend by lowering fees or increasing maker rebates.

Overall, Hyperliquid is consciously choosing a low-profit market layer position (via HIP-3 and Builder Codes), while allowing a high-margin broker layer to grow above it.

If Builder frontends continue to expand, they will increasingly determine user-facing pricing, control user retention and monetization interfaces, and gain bargaining power at the routing layer, exerting long-term pressure on Hyperliquid’s fee rates.

Defending Distribution Rights and Introducing Non-Exchange Profit Pools

The most immediate risk is commodification.

If third-party frontends can sustain lower prices than the native interface over the long term, and even achieve cross-venue routing, Hyperliquid will be pushed toward a wholesale execution economic model.

Recent design adjustments show that Hyperliquid is attempting to avoid this outcome while expanding new revenue sources.

Distribution Defense: Maintaining Native Frontend’s Economic Competitiveness

A previously proposed staking discount scheme allows Builders to stake HYPE for up to 40% fee discounts, effectively providing a structural cheaper path for third-party frontends compared to Hyperliquid’s native interface. Reversing this scheme cancels the direct subsidy that external distribution “undercuts” pricing.

Meanwhile, HIP-3 markets were initially positioned mainly for Builder distribution, not prominently displayed on the main frontend; but now, these markets are beginning to be shown on Hyperliquid’s native frontend with strict listing standards.

This signals very clearly: Hyperliquid still maintains a permissionless Builder layer, but will not sacrifice its core distribution rights.

USDH: Shifting from Trading Monetization to “Capital Float” Monetization

The launch of USDH aims to reclaim stablecoin reserve yields that would otherwise be captured outside the system. Its disclosed structure is a 50/50 split of reserve income: 50% to Hyperliquid, 50% for USDH ecosystem growth. Additionally, the trading fee discounts offered on USDH-related markets further reinforce this orientation: Hyperliquid is willing to accept lower margins on individual trades in exchange for a larger, more sticky profit pool tied to balances.

In effect, this introduces a kind of annuity-like income stream for the protocol, with growth dependent on the monetary base rather than just nominal trading volume.

Portfolio Margin: Introducing Prime Broker-like Financing Economics

Portfolio margin unifies margin for spot and perpetual contracts, allowing different exposures to offset each other, and introduces native lending cycles.

Hyperliquid will retain 10% of the interest paid by borrowers, making the protocol’s economics increasingly dependent on leverage usage and interest rates, rather than just trading volume. This aligns more with a broker / prime broker revenue model than a pure exchange logic.

Path to a “Broker-Style” Economic Model for Hyperliquid

In throughput, Hyperliquid has already reached the scale of top-tier trading venues; but in monetization, it still resembles a market layer: extremely high nominal volume combined with effective fee rates in the single digits of basis points. The gap with Coinbase and Robinhood is structural.

Retail platforms are at the broker layer, holding user relationships and funds, capable of monetizing multiple profit pools (financing, idle cash, subscriptions); while pure exchanges sell execution services, which under liquidity and routing competition tend to become commoditized, with net capture continuously compressed. Nasdaq exemplifies this structural constraint in TradFi.

Early Hyperliquid clearly tilted toward a trading venue prototype. By splitting the distribution layer (Builder Codes) and product creation layer (HIP-3), it accelerated ecosystem expansion and market coverage; but this architecture also risks pushing economic viability outward: if third-party frontends decide on integrated pricing and can route across venues, Hyperliquid risks being squeezed into a thin-profit wholesale execution track.

However, recent moves show a conscious shift: without abandoning the advantages of unified execution and clearing, it aims to defend distribution rights and expand revenue sources into “balance-based” profit pools.

Specifically: the protocol no longer subsidizes external frontends to be cheaper structurally than native UI; HIP-3 is more natively integrated; and it introduces balance sheet-like revenue sources.

USDH pulls reserve yields back into the ecosystem (50/50 split, with fee discounts on USDH markets); portfolio margin introduces financing economics through a 10% cut of borrowing interest.

Overall, Hyperliquid is converging toward a hybrid model: using execution as the base, layered with distribution defense and balance-driven profit pools. This reduces the risk of being trapped in low-basis wholesale venues, while maintaining unified execution and clearing advantages, and moving toward a broker-like revenue structure.

Looking ahead to 2026, the unresolved question is: can Hyperliquid further evolve toward a broker-style economy without destroying its “outsourcing-friendly” model? USDH is the clearest test case: at around $100 million in supply, when the protocol does not control distribution, external issuance expansion appears slow. An obvious alternative path could be a default at the UI level—such as automatically converting about $4 billion in USDC holdings into native stablecoins (similar to Binance’s automatic conversion to BUSD).

If Hyperliquid truly wants to access the broker profit pool, it may also need broker-like behaviors: stronger control, tighter native product integration, and clearer boundaries with the ecosystem team in distribution and balance competition.