<p>The goal of Ethereum has never changed: to build a global, censorship-resistant, permissionless blockchain platform. Ethereum innovates not only in cryptography and economics but also in social technology. The Ethereum ecosystem showcases a model of open and decentralized collaboration. The rise of Layer 2 protocols has increased Ethereum's transaction capacity by 17 times and reduced fees. Currently, Layer 2 faces challenges of scalability and heterogeneity. Scalability can be achieved by increasing Blob space and improving technology. Layer 2 needs to enhance security and interoperability to ensure the economic viability of ETH. The future of Ethereum depends on the active participation of everyone.</p>

Original Title: Scaling Ethereum L1 and L2s in 2025 and beyond  
Original Author: Vitalik  
Original Compilation: Odaily Planet Daily  
How Ethereum's goal has not changed since day one: to build a global, censorship-resistant, permissionless blockchain platform. This is a freely open platform for decentralized applications, aligned with the principles of GNU + Linux, Mozilla, Tor, Wikipedia, as well as many great free and open-source software projects (now referred to as the spirit of revival and cypherpunk). Over the past decade, Ethereum has also developed a characteristic that I greatly appreciate: in addition to innovations in cryptography and economics, Ethereum is also an innovation in social technology. The Ethereum ecosystem as a whole showcases a more open and decentralized way of collaboration. Political philosopher Ahmed Gatnash described his experience at Devcon as follows: ...It gave me a glimpse of what an alternative world might look like—a world with almost no barriers and completely disconnected from traditional systems. Here, the social hierarchy has been subverted, and the highest social status is held by those geeks focused on independently solving the issues they genuinely care about, rather than those playing the game to climb the ladders of traditional institutions and accumulate power. Here, almost all power is soft power. I find this beautiful and incredibly inspiring—it makes one feel that in such a world, anything is possible, and that such a world is actually within reach. Technological projects and social projects are essentially intertwined. If you have a decentralized technological system at time point T, but it is maintained by a centralized social process, there is no guarantee that your technological system will still be decentralized at time point T+1. Similarly, social processes are maintained by technology in various ways: technology attracts users, the ecosystems brought by technology provide retention incentives for developers, technology grounds the community, focusing on building rather than just socializing, and so on. After a decade of effort, under the co-governance of technological and social attributes, Ethereum has shown another important quality: it can provide practical services to people on a large scale. Millions are using ETH or stablecoins as a form of savings, and more people are using these assets for payments: I am one of them. Ethereum has efficient and practical privacy tools that I use to pay for VPN services to protect my internet data. It also has ENS, a robust decentralized alternative for DNS and a broader public key infrastructure. Furthermore, there are user-friendly decentralized Twitter alternatives and DeFi tools on Ethereum, providing hundreds of millions with higher yields and lower risk assets compared to traditional finance. Five years ago, I was reluctant to discuss the use cases of the latter, mainly because the infrastructure and code were not yet mature. At that time, we had just experienced the large-scale, painful smart contract hacks of 2016-2017, and if there was a 5% chance each year of losing 100% of the yield, a 7% annualized yield would be meaningless compared to a 5% yield. Additionally, transaction fees were too high to enable massive adoption of these tools. Today, these tools have proven their resilience over time, the quality of auditing tools has improved, and we are increasingly confident in their security. We know what cannot be done. L2 scaling technology is working, and transaction fees have remained very low for nearly a year. We need to continue enhancing Ethereum's technical and social attributes and its practicality. If we have only the former without the latter, we will become an increasingly ineffective "decentralized" community, merely protesting against the "immoral and wrong actions" of mainstream institutions without being able to truly provide better alternatives. If we have only the latter without the former, we are no different from Wall Street's mentality of "greed is good," while many joined the Ethereum community precisely to escape that mindset. This duality of both technology and practicality has far-reaching implications. In this article, I want to focus on a specific aspect that is crucial for Ethereum users in the short and medium term: Ethereum's scaling strategy.  
The Rise of Layer 2  
Currently, our path to scaling Ethereum is through Layer 2 protocols. The Layer 2 of 2025 has made significant leaps compared to the early experiments of 2019: they have achieved key decentralization milestones, are securing billions of dollars in assets, and have increased Ethereum's transaction capacity by 17 times, while also reducing fees by the same margin. All of this is happening just as a wave of successful applications has emerged: various DeFi platforms, social networks, prediction markets, and novel projects like Worldchain (which now has 10 million users). Moreover, the "enterprise blockchain" movement, which was deemed a dead end due to the failures of consortium chains in the 2010s, has been revitalized with the rise of Layer 2, with Soneium being a prominent example. These successes also demonstrate the social advantages of Ethereum's decentralized and modular scaling approach: the Ethereum Foundation does not need to personally find all the users, but rather there are dozens of independent entities spontaneously driving this forward. These entities have also made critical contributions to the technology; without them, Ethereum could not have achieved its current success. This is why we are finally approaching "escape velocity."  
Challenges: Scaling and Heterogeneity  
Current Layer 2 faces two main challenges:  
1. Scaling: The current "Blob space" barely supports existing Layer 2s and their use cases, but it is far from sufficient to meet future demands.  
2. Heterogeneity issue: Ethereum's original scaling vision was to create a blockchain composed of multiple shards, each shard being a copy of the EVM, handled by a small number of nodes. In theory, Layer 2 is the realization of this vision. However, in practice, there is a key difference: each shard (or a group of shards) is created by different participants, regarded as different chains within the infrastructure, and typically follows different standards.This situation has led to challenges for developers and users in terms of composability and user experience. The first issue is a technically understandable challenge, with a simple solution that is difficult to implement: providing more "Blob space" for Ethereum. Additionally, Ethereum L1 can alleviate pressure in the short term through moderate scaling and improvements in proof of stake, stateless and light verification, storage, EVM, and cryptography. The second issue is a coordination challenge that has already garnered significant public attention. The Ethereum ecosystem is no stranger to complex technical tasks completed through cross-team collaboration—after all, we accomplished The Merge. However, the coordination problems here are more challenging due to the larger number of participants, more diverse goals, and a later start to the process. Even so, our ecosystem has solved many difficult problems in the past, and we can do it again this time. One possible shortcut to scaling is to abandon Layer 2 and achieve much higher Gas limits directly through Layer 1 (whether through multiple shards or a single shard). However, this approach sacrifices too much of the current social structure of Ethereum, which is highly effective in integrating different forms of research, development, and ecosystem building culture. Therefore, we should stick to the existing path, continuing to scale primarily via Layer 2 while ensuring that Layer 2 truly delivers on its promises. This means the following: Layer 1 needs to accelerate the expansion of Blob capacity. Layer 1 also needs to moderately scale the EVM and increase Gas limits to address the activities that Layer 1 will still host even in a Layer 2-dominant environment (such as zero-knowledge proofs, large-scale DeFi, deposit and withdrawal operations, special large-scale exit scenarios, key storage wallets, asset issuance, etc.). Layer 2 needs to continuously enhance security. Layer 2 should provide the same security guarantees as sharding (including censorship resistance, light client verifiability, and no embedded trusted parties, etc.). Layer 2 and wallets need to accelerate improvements and standardize interoperability. This includes chain-specific addresses, messaging and cross-chain bridge standards, efficient cross-chain payments, on-chain configurations, etc. Using Ethereum should feel like using a single ecosystem, rather than 34 different blockchains. The withdrawal times for Layer 2 need to be significantly shortened. As long as the basic interoperability needs are met, the heterogeneity of Layer 2 is beneficial. Some Layer 2 solutions will be based on governance-minimized Rollups, running a copy that is completely the same as Layer 1 EVM; some Layer 2 solutions will attempt different virtual machines; others will function more like servers, leveraging Ethereum to provide additional security for users. We need to cover various types of Layer 2 across this spectrum. We need to explicitly consider the economics of ETH. Even in a Layer 2-dominant world, we need to ensure that ETH can continue to accumulate value and provide solutions for multiple value accumulation models. Next, we will discuss each topic in detail. Scaling: Blob, Blob, or Blob In EIP-4844, there are 3 Blobs per slot, with a data bandwidth of 384kB per slot. A rough estimate shows that this corresponds to 32kB per second, with each transaction on-chain occupying about 150 bytes, allowing us to support approximately 210 transactions per second (TPS). According to data from L2beat, this estimate is nearly spot on. Pectra, scheduled for release in March, will double the number of Blobs to 6 per slot. The current goal of Fusaka is primarily focused on PeerDAS, planning to prioritize only the implementation of PeerDAS and EOF. PeerDAS may potentially increase the number of Blobs by another 2-3 times. The next goal is to continuously increase the number of Blobs. When 2D sampling is reached, the number of Blobs could be increased to 128 per time slot, and further enhancements can be made in the future. Combined with data compression improvements, on-chain TPS could reach 100,000. The above is a restatement of the established roadmap before 2025. The key question is: how do we accelerate this process? My answers are as follows: more clearly prioritize non-Blob functionalities lower. More clearly indicate that Blobs are the target and list relevant peer-to-peer R&D as a priority in talent recruitment. Allow stakers to directly adjust Blob targets, similar to Gas limits. This would enable the Blob targets to increase more rapidly with technological improvements without waiting for hard forks. More aggressive approaches could be considered to increase the number of Blobs more quickly by introducing more trust assumptions for low-resource stakers, but we need to approach this with caution. Enhancing Security: Proof Systems and Native Rollups Currently, there are three Stage 1 Rollups (Optimism, Arbitrum, Ink) and three Stage 2 Rollups (DeGate, zk.money, Fuel). However, most activities still occur on Stage 0 Rollups (i.e., multi-signature schemes). This situation needs to change. One significant reason for the slow progress is that building a reliable proof system and establishing enough confidence to fully depend on its security (abandoning "training wheels") is very challenging. To achieve this, there are two paths: Stage 2 + multiple proof systems + formal verification: Achieving redundancy through multiple proof systems and enhancing security confidence through formal verification (e.g., "verified ZK-EVM projects"). Native Rollup: Integrating the verification of EVM state transition functions into the protocol itself, for instance through precompiled contracts. Currently, both paths need to be advanced in parallel. For "Stage 2 + multiple proof systems + formal verification," the roadmap is relatively clear.Accelerating development can be achieved through enhanced collaboration within the software stack, which not only reduces redundancy but also improves interoperability as a byproduct. For local Rollups, this is still in its early stages, especially requiring more thought on how to maximize the flexibility of precompiled contracts. An ideal goal is to support not just a complete clone of the EVM, but also an EVM that includes arbitrary modifications, allowing modified EVM Rollups to still utilize the precompiled contracts of local Rollups, introducing "custom provers" only for the modified parts. This may involve adapting precompiled contracts, opcodes, state trees, and other components. The objectives of interoperability and standardization aim to make the transfer of assets and usage experiences across different L2s as seamless as interactions between different "shards" on the same blockchain. Currently, there is a relatively clear roadmap in this area: Chain-specific addresses: Addresses should include account information on the chain as well as the identifier of the chain itself. For example, ERC-3770 is an early attempt, and more complex designs have emerged, even migrating the L2 registry to Ethereum L1. Standardizing cross-chain bridges and messaging: There should be standardized methods for validating proofs and passing messages between L2s, and these standards should not rely on trust mechanisms like multi-signature bridges. An ecosystem reliant on multi-signature bridges is unacceptable. If such trust assumptions were absent in the shard design of 2016, they are equally unacceptable today. Accelerating deposit and withdrawal times: The time for "local" messages should be reduced from weeks to minutes (with the ultimate goal being one block time). This requires faster ZK-EVM provers and support for proof aggregation technologies. Reading L1 data from L2 synchronously: For example, L1SLOAD and REMOTESTATICCALL, these functions will greatly simplify interoperability across L2s while aiding the functionality of key store wallets. Shared ordering and other long-term work: One of the reasons why Rollup-based designs hold value is that they can implement features like shared ordering more efficiently. Under the premise of meeting these standards, L2s can differ in security, performance, and design models according to demand. For instance, different virtual machines, ordering models, and trade-offs between scale and security can be explored. However, the security levels of each L2 must be clearly defined for users and developers. To accelerate progress, cross-domain institutions within the ecosystem can take on a larger share of the work, such as the Ethereum Foundation, client development teams, and mainstream application development teams. This will reduce coordination costs, making the adoption of standards an easier decision, as the development workload for each L2 and wallet will subsequently decrease. However, as an extension of the Ethereum ecosystem, L2s and wallets also need to strengthen the "last mile" of development to ensure these features truly reach users. The Economics of ETH: We should adopt a multi-pronged approach that covers all major potential sources of value for ETH as a three-point asset. Key components of this strategy may include: achieving widespread consensus to solidify ETH as the primary asset within the larger (L1 + L2) Ethereum economic system, supporting applications that use ETH as the primary collateral, etc. Encouraging L2s to support ETH and allocating a portion of fees. This could be achieved by burning a portion of fees, permanently staking fees and donating the proceeds to Ethereum ecosystem public goods, or through other various means. Supporting Rollup-based designs, partially serving as a pathway to obtain value through MEV as L1, but not mandating that all Rollups must be based on this design, as this does not apply to all applications and it cannot be assumed that this approach alone can solve all problems. Increasing the number of Blobs, considering setting a minimum Blob price, and treating Blobs as another potential source of income. For example, if the average cost of Blobs over the past 30 days remains constant (driven by demand) while the number of Blobs increases to 128, Ethereum would burn 713,000 ETH annually. However, the demand curve may not necessarily be so favorable, so this cannot be relied upon to solve the problem alone. Conclusion: The Road Ahead: Ethereum has matured in both its technology stack and social ecosystem, leading us towards a more free and open future, where hundreds of millions can benefit from crypto assets and decentralized applications. However, there is still a significant amount of work to be done, and now is the time to double down on efforts. If you are an L2 developer, engage in developing tools to safely scale Blobs, develop code to extend EVM execution, and implement features and standards that ensure L2 interoperability. If you are a wallet developer, you also need to contribute and implement standards to ensure that the ecosystem provides a seamless experience for users while maintaining the same level of security and decentralization as Ethereum L1. If you are an ETH holder or community member, actively participate in these discussions; there are many areas that require deeper thinking and collective brainstorming. The future of Ethereum depends on the active participation of each and every one of us.