Which L1 captures more value?

ARK Invest digital asset researcher Lorenzo Valente compared Ethereum, Solana and Hyperliquid with three U.S. restaurant businesses on Sept. 3, arguing that their different operating structures require separate valuation frameworks.

In an essay, Valente likened Ethereum to McDonald’s, Solana to Chipotle and Hyperliquid to In-N-Out. The comparisons address how each blockchain expands, controls its infrastructure and directs revenue toward its native asset.

Valente argued that Ethereum operates like a franchise network because independent layer 2 teams build their own systems while relying on Ethereum for settlement. Solana resembles a company-owned chain because applications execute inside one integrated environment. Hyperliquid offers a more concentrated structure built around its trading venue, consensus system and fee-funded HYPE purchases.

The restaurant comparisons are Valente’s analytical framework. They are not ARK investment recommendations or objective classifications of the three networks.

Ethereum resembles a franchise that charges limited rent

Ethereum’s scaling roadmap allows layer 2 networks such as Arbitrum, Base and OP Mainnet to process transactions away from its main execution layer. Those networks periodically submit data or proofs to Ethereum to inherit parts of its security and settlement infrastructure.

Valente compared this arrangement with McDonald’s franchising model. Ethereum provides the brand, standards, developer ecosystem and settlement layer. Independent teams finance and operate the layer 2 networks, just as franchise operators fund and manage individual restaurants.

This structure helps Ethereum expand without financing every new execution network itself. Separate teams can develop specialized products, attract users and experiment with different technologies while continuing to settle on Ethereum.

However, Valente argued that Ethereum captures too little of the economic activity generated by those networks. Layer 2 operators collect transaction fees from users but pay Ethereum primarily for data availability and settlement.

EIP-4844 introduced separate blob space for rollup data in March 2024. Blobs reduced the cost of submitting layer 2 data to Ethereum, making transactions cheaper for users. They also lowered the fees that rollups paid to the main network when blob capacity exceeded demand.

Valente described this as Ethereum building a successful franchise network but failing to collect enough rent. In his view, Ethereum owns valuable settlement infrastructure but prices access too close to its operating cost.

The analogy has limits. Ethereum does not sign commercial franchise agreements with layer 2 operators. It also cannot impose royalties, control their products or prevent them from using alternative data availability services. Any proposal to raise the minimum blob fee would require technical review and acceptance through Ethereum’s decentralized governance process.

Ethereum developers have considered changes to blob pricing as demand and capacity evolve. A higher fee floor could increase payments to Ethereum, but it could also increase layer 2 transaction costs or encourage operators to use competing systems.

Solana keeps more activity inside one environment

Valente compared Solana with Chipotle because both follow what he described as a vertically integrated model. Solana processes application activity directly through its base network instead of making external rollups the main route for scaling.

Trades on Jupiter, token launches, stablecoin transfers and other application transactions share the same execution environment. Users pay base and priority fees, while validators may receive additional value through transaction ordering and Jito tips.

This structure keeps more of the fee flow within the Solana network. Validators and their delegators receive compensation, while part of the base fee is burned. The relationship between network use and value capture is therefore more direct than it is when execution occurs on an independent layer 2.

Valente compared that arrangement with Chipotle owning and operating its restaurants. The company controls the customer experience and retains store revenue, but it must also finance expansion and absorb operational failures.

Solana faces a similar trade-off. Its unified architecture provides direct control over execution, fee markets and performance upgrades. It also means congestion or network disruption can affect applications across the ecosystem simultaneously.

The network has invested in additional validator clients, including Firedancer, to improve performance and reduce its dependence on one main software implementation. Solana’s Firedancer and Alpenglow upgrades could strengthen performance and validator diversity, although their full effects depend on deployment and operator adoption.

Valente argued that Solana’s integrated model produces better fee retention than Ethereum’s rollup structure. That assessment depends on which revenues and costs are included. Validator rewards involve token issuance, while application fees do not automatically accrue equally to every SOL holder.

Hyperliquid creates the shortest fee-capture chain

Hyperliquid received the In-N-Out comparison because it combines a focused product range, internal infrastructure and limited reliance on outside capital. Its original product centered on perpetual futures trading through an onchain order book.

The platform built its own consensus system, HyperBFT, and operates its trading infrastructure through HyperCore. It later added HyperEVM for general smart contract applications, but derivatives remain a major source of activity and revenue.

Valente argued that Hyperliquid has the shortest value-capture path among the three networks. Trading fees flow into the protocol, and the Assistance Fund uses most eligible revenue to purchase HYPE from the market.

The model differs from a conventional corporate share repurchase. HYPE is a crypto token rather than equity, and holding it does not grant the same legal claims as owning company stock. Assistance Fund purchases can still create recurring market demand when trading activity generates sufficient fees.

Hyperliquid’s Assistance Fund directs most protocol trading fees into HYPE purchases. Crypto.news reported in May that the fund had used more than $1.3 billion for purchases since the mechanism began, based on available protocol and market data.

More recent research found that Hyperliquid and Pump.fun accounted for nearly 90% of tracked crypto token repurchases during 2026. Those figures measure purchases during the examined period and should not be interpreted as guaranteed future demand.

Hyperliquid has also expanded through HIP-3, which lets approved builders deploy perpetual markets while using its underlying infrastructure. Official documentation says spot and HIP-3 deployers may retain up to 50% of fees generated by their deployed assets.

Valente compared the arrangement with a tightly controlled restaurant operator allowing outside builders to introduce products without surrendering its infrastructure or customer relationship.

Different models produce different concentration risks

Ethereum’s main advantage under Valente’s framework is distribution. Independent layer 2 teams provide external capital, engineering capacity and access to large companies. The cost is weaker control over users, execution revenue and the behavior of those networks.

Solana retains more activity inside one system. This can strengthen fee capture and product coordination, but the network must support a broader technical surface and absorb system-wide operational risks.

Hyperliquid offers the most direct relationship between product revenue and token purchases. It also carries the greatest concentration risk of the three models because activity, leadership and revenue remain closely connected to one trading ecosystem.

Valente warned that builders responsible for a large share of HIP-3 trading could eventually seek better fee terms. Revenue may also weaken during a prolonged decline in derivatives activity.

The comparison does not establish which token will outperform. Valuations also depend on issuance, liquidity, governance, competition, regulation and demand for the products running on each network.

No verified market move could be attributed directly to Valente’s essay. ETH, SOL and HYPE trade continuously and respond to broader crypto prices, leverage, protocol activity and macroeconomic conditions.

What happens next

Ethereum’s value-capture debate will focus partly on blob demand and pricing. Developers can adjust capacity or fee parameters, but changes require testing and community support. Higher settlement revenue would need to be balanced against affordable layer 2 transactions.

Solana’s model will be tested by network upgrades, validator-client diversity and its ability to support higher activity without recurring congestion. The expansion of institutional products and consumer applications could also change its fee composition.

For Hyperliquid, HIP-3 adoption will show whether the network can expand beyond its internally developed markets while preserving its revenue share. Trading volumes and Assistance Fund purchases will remain important measures of the model’s durability.

Valente’s central argument is that investors should not value every layer 1 network using identical metrics. Ethereum emphasizes external ecosystem expansion, Solana emphasizes unified execution and Hyperliquid emphasizes direct product revenue. Each model can succeed, he said, but each carries a different path to failure.

FAQs