Ethereum, Solana and Hyperliquid capture value in markedly different ways, according to ARK Invest digital asset researcher Lorenzo Valente, who has compared the three layer 1 networks with McDonald’s, Chipotle and In-N-Out.
Valente set out the comparison in an essay on 3 September, arguing that investors should not assess every blockchain using the same valuation framework. His restaurant analogies are an analytical model rather than ARK investment recommendations or objective classifications of the networks.
Ethereum, in his view, resembles a franchise system. Solana operates more like a company-owned chain, while Hyperliquid has a concentrated model built around its trading platform, infrastructure and use of fees to buy its native HYPE token.
Ethereum’s scaling strategy relies heavily on layer 2 networks such as Arbitrum, Base and OP Mainnet. These systems process transactions away from Ethereum’s main execution layer, then periodically submit data or proofs to the network to inherit elements of its security and settlement infrastructure.
Valente compared that structure with McDonald’s franchising model. Ethereum supplies the brand, technical standards, developer ecosystem and settlement layer, while independent teams fund and run individual layer 2 networks.
The arrangement allows Ethereum to expand without paying for every new execution network itself. Separate operators can develop specialist products, attract users and test different technologies while continuing to settle transactions on Ethereum.
However, the model also limits how much economic activity flows back to the main network. Layer 2 operators charge users transaction fees, while their payments to Ethereum mainly cover data availability and settlement.
The introduction of EIP-4844 in March 2024 created dedicated blob space for rollup data. Blobs made it cheaper for layer 2 networks to submit information to Ethereum, lowering costs for users but also reducing the fees paid to Ethereum when available blob capacity was greater than demand.
Valente described the outcome as a successful franchise network that does not collect enough rent. Ethereum controls valuable settlement infrastructure, he argued, but prices access at a level close to its operating cost.
The comparison is not exact. Ethereum does not sign commercial franchise contracts with layer 2 operators, cannot impose royalties or dictate their products, and cannot stop them using alternative data availability services. Any attempt to increase the minimum blob fee would require technical assessment and approval through Ethereum’s decentralised governance process.
Developers have considered changes to blob pricing as demand and capacity develop. A higher fee floor could increase revenue for Ethereum, although it might also raise layer 2 costs or encourage operators to move to competing systems.
Solana keeps execution in one network
Solana was compared with Chipotle because applications operate inside one vertically integrated environment rather than relying on independent rollups as the main scaling route.
Transactions involving Jupiter, token launches, stablecoin transfers and other applications share Solana’s execution layer. Users pay base and priority fees, while validators can receive additional value from transaction ordering and Jito tips.
That design leaves more of the fee flow within the Solana ecosystem. Validators and their delegators receive compensation, while part of the base fee is burned. The link between network use and value capture is therefore more direct than when execution takes place on an independently operated layer 2.
As with Chipotle’s company-owned restaurants, the network retains greater control over the customer experience and revenue. The trade-off is that it must fund expansion itself and carry the consequences of operational problems.
Solana’s unified architecture gives it direct control over execution, fee markets and performance upgrades. But congestion or a network failure can affect applications across the ecosystem at the same time.
The network has invested in additional validator clients, including Firedancer, to improve performance and reduce reliance on a single major software implementation. Solana’s Firedancer and Alpenglow upgrades could improve performance and validator diversity, although their eventual impact will depend on deployment and the willingness of operators to adopt them.
Valente said Solana’s integrated model provides stronger fee retention than Ethereum’s rollup structure. That conclusion depends on which revenues and costs are counted. Validator rewards include token issuance, while application fees do not automatically benefit every SOL holder equally.
Hyperliquid links fees to HYPE purchases
Hyperliquid received the In-N-Out comparison because it combines a narrow product focus, internally developed infrastructure and limited dependence on external capital.
Its original offering centred on perpetual futures trading through an onchain order book. The platform developed its own consensus system, HyperBFT, and operates its trading infrastructure through HyperCore. It has since added HyperEVM for general smart contract applications, although derivatives remain a major source of activity and revenue.
Valente argued that Hyperliquid has the most direct value-capture mechanism of the three networks. Trading fees go to the protocol, while its Assistance Fund uses most eligible revenue to purchase HYPE on the market.
That process is not the same as a conventional company share buyback. HYPE is a crypto token, not equity, and ownership does not provide the legal claims associated with company stock. Nevertheless, the purchases can create recurring market demand when trading activity generates enough fees.
The Assistance Fund directs most protocol trading fees into HYPE purchases. Crypto.news reported in May that the fund had spent more than $1.3bn on purchases since the mechanism began, based on available protocol and market data.
More recent research found that Hyperliquid and Pump.fun represented nearly 90% of tracked crypto token repurchases during 2026. Those figures relate only to the period examined and do not guarantee future demand.
Hyperliquid has also expanded through HIP-3, which allows approved builders to launch perpetual markets using its underlying infrastructure. Official documentation says spot and HIP-3 deployers can retain up to 50% of the fees generated by their assets.
Valente likened this to a tightly controlled restaurant operator allowing outside builders to add products while retaining ownership of the infrastructure and customer relationship.
Each model brings a different risk. Ethereum’s principal advantage is distribution: independent layer 2 teams provide external funding, engineering resources and access to major companies. The cost is less control over users, execution revenue and the future direction of those networks.
Solana keeps more activity within one system, potentially improving fee capture and coordination. It must also support a wider technical surface and absorb risks that can affect the whole network.
Hyperliquid has the clearest connection between product revenue and token purchases, but also the highest concentration risk. Activity, leadership and revenue are closely tied to one trading ecosystem.
Valente warned that builders responsible for a significant share of HIP-3 activity could eventually demand more favourable fee terms. Revenue could also fall during a prolonged downturn in derivatives trading.
The comparison does not determine which token will perform best. Valuations are also influenced by issuance, liquidity, governance, competition, regulation and demand for applications operating on each network.
There was no verified market move directly linked to Valente’s essay. ETH, SOL and HYPE trade continuously and are affected by wider crypto prices, leverage, protocol activity and macroeconomic conditions.
Ethereum’s value-capture debate is likely to centre partly on blob demand and pricing. Developers can change capacity or fee settings, but any adjustment requires testing and support from the community. More settlement revenue would need to be weighed against the importance of keeping layer 2 transactions affordable.
Solana’s model will be tested by network upgrades, greater diversity among validator clients and its ability to handle increased activity without repeated congestion. Institutional products and consumer applications could also alter the make-up of its fees.
For Hyperliquid, adoption of HIP-3 will indicate whether the network can expand beyond its internally developed markets while maintaining its share of revenue. Trading volumes and Assistance Fund purchases will remain key measures of whether the model can endure.
Valente’s central argument is that different layer 1 networks have different routes to success and different ways in which they could fail: Ethereum prioritises external ecosystem expansion, Solana unified execution, and Hyperliquid direct product revenue.
