Network Externalities and the Advancing Scale of Layer-2 Protocols

By combining user growth and the subsequent advantages that accrue as a result of network growth, our findings predict that there will continue to be exponential growth in Ethereum-based L2 networks due to their growing number of users. In addition, the continued increase in Ethereum-based L2 networks will lead to more and more developers being attracted to develop applications and services. This is a new type of network effect, as opposed to social networks, although it produces the same type of outcome: increased liquidity, developer tooling, oracle and bridge infrastructure, etc. This creates large, growing economies of scale and, similar to how other markets utilizing network effects operate, there will be a "winner-takes-most" advantage for early movers.

The volume of growth also serves to illustrate the scale of this shift. As of the beginning of 2025, the Ethereum L2 networks collectively processed over 90 million daily transactions, compared to approximately 1 million on Ethereum's base layer. In addition, Base, an Ethereum L2 developed by Coinbase, launched in August 2023 and achieved the highest Total Value Locked (TVL) of any Ethereum L2 within approximately 18 months of its launch, reaching $10 billion in TVL before exceeding Arbitrum's TVL of approximately $17 billion in the first quarter of 2025. The volume of TVL that this network generates demonstrates that liquidity creates additional liquidity, an economic representation of network externalities.

The additional factor which reinforces this pattern is the cost associated with conducting transactions on Ethereum L2 networks. The Dencun upgrade that took place in March 2024 introduced two important changes: EIP-4844 and proto-danksharding. Because of these changes, the average fees charged on Optimism and Base were lowered to less than $0.01 per transaction. Many users who had previously been priced out of Ethereum activity have now returned because of lower fees. The result is a compounding effect, where reducing cost reduces barriers to entry, which increases transaction volume, which in turn increases the revenue generated by sequencers, which funds further infrastructure development.

As the L2 ecosystem matures, it is becoming increasingly important for developers using Base—and any other Layer-2s—to have a reliable low-latency infrastructure for their applications to function. Developing applications on Base requires developers to have reliable, consistent access to node infrastructure providers via remote procedure call (RPC) providers. Having an efficient Base RPC provider has become a requirement for any application that seeks to be competitive in today's marketplace. This segment of providers operates similarly to cloud computing providers, through economies of scale. The economic efficiencies afforded by the economies of scale associated with this provider segment will likely lead to a consolidation of the market around providers able to provide lower latency, higher uptime, and lower costs for RPC services, thereby increasing the switching costs and "stickiness" of using RPC providers for enterprise developers.

In addition to generating RPC provider market consolidation, the reliance of a growing number of applications on a given provider's ecosystem increases the amount of capital that flows into it, thereby strengthening the provider's market position. Even without formal barriers to entry, the economic efficiencies of lock-in effects and the feedback loops that develop around these providers make it reasonable to expect that the emergence of concentrated provider marketplaces will continue to occur.

The aforementioned economic forces do not operate in a vacuum. The economic forces are affected by the governance decisions made in connection with revenue distributions of protocols, decision-making authority regarding upgrades to protocols, and decision-making authority over the operation of sequencers. The combined market capitalisation of the ARB and OP governance tokens for Arbitrum and Optimism, respectively, is projected to exceed tens of billions of dollars by the end of 2025. However, it is anticipated that both Arbitrum and Optimism will become more centralized with respect to the way sequencers operate as the volume of transactions and the number of active users increase. This tension between the concentration of value and stakeholder dependency upon these ecosystems will create pressure on the providers of these networks to make credible commitments to censorship resistance while the underlying infrastructures continue to trend toward concentration, and this pressure will likely increase as these ecosystems mature. Regulators in many major jurisdictions have started to take a close look at entities that may have the ability to unilaterally control transaction ordering at scale.

Layer-2s are an example of a much larger phenomenon occurring within digital infrastructure markets—the combination of network effects and user adoption producing increasing returns that favour early entrants with access to capital. The resulting concentration within these markets, as evidenced by the ways in which total value locked is distributed across various L2 networks and RPC providers as well as the control of sequencers, creates questions surrounding the governance and regulatory policies relating to L2s. As L2 ecosystems scale, it will be critical to find a way to harmonise the efficiencies gained through an increase in the concentration of infrastructure with the goals of decentralisation and censorship resistance that served as the initial motivation for developing the technologies.