Opinion by: Jay Jog, co-founder of Sei Labs 

When CryptoKitties crashed the Ethereum network in 2017, the industry learned a hard lesson about blockchain scalability. Today, with over $100 billion locked in decentralized finance (DeFi) and millions of non-fungible tokens (NFTs) being traded, that lesson is more relevant than ever. The Ethereum Virtual Machine (EVM) — the engine that powers this activity — is reaching its limits.

So far, the crypto community’s answer has been layer 2 solutions — separate chains that process transactions and report back to Ethereum. But what if the community’s been looking for answers in the wrong place?

Layer 2s are not the solution

Layer 2 blockchains have long been touted as the solution to the EVM’s performance challenges, given their ability to offload the computational work from Ethereum to a secondary chain. Layer-2 solutions have proven to be nothing more than a “quick fix” instead of a permanent solution, as many hoped for. As Gemini reported, a new layer 2 appeared every 19 days in 2024, indicating that the competitive landscape is creating more problems instead of solving them.

Layer 2 solutions come with their own challenges, primarily tied to centralization and interoperability. Many of today’s layer 2 blockchains run with centralized sequencers that could expose the network to transaction censorship, transaction reordering and more. Additionally, Vitalik Buterin stated in a recent blog post that layer 2s are struggling to maintain interoperability. This called attention to the disorganized state of layer 2s, further contributing to liquidity fragmentation and a complex user experience. 

Advanced rollup designs have tried to fix these pain points. Recently, there has been a new design called native rollups that is trying to tackle layer 2’s centralization issues. Native rollups take value away from projects, which will significantly deter adoption. Consequently, it is doubtful that native rollups are the answer to all of Ethereum’s urgent problems. 

With just as many challenges as the EVM itself, why rely on layer 2s instead of looking elsewhere? Could there be a better solution? According to L2BEAT, it costs around $95.53 million annually to run all the major L2s. Instead of spending more money on building and running more L2s and interoperability solutions, why not focus on refining the existing foundational layer? 

A more accurate alternative to TPS

To create the most performant layer 1s, the industry must first reevaluate the approach to track blockchain performance. Most blockchains focus on throughput, using transactions per second (TPS) to compare chain performance. While many argue that reaching the most significant transactions per second is the way to enable mainstream adoption for crypto, TPS unfortunately doesn’t allow for apples-to-apples comparisons since different types of transactions require different amounts of compute. 

For example, an Ether transfer requires 21,000 units of gas, whereas an ERC-20 transfer needs 65,000, confirming that TPS conveys zero value when tracking mass transactions and network throughput.

A new standardized performance metric that better reflects network computing capability must be developed to understand a blockchain's full potential. This is where an alternative performance metric called “gas per second” emerges — a measure that evaluates the gas fees required to process transactions, better reflecting different transaction types. While TPS is best served to assess simple ETH transfers, gas per second shows the bigger picture by considering all computational efforts, even for complex transactions. 

Given the novelty of this metric, measuring gas per second across all chains will be a long process but a crucial step in blockchain’s evolution. 

Going back to the basics: Layer 1s

The capability of layer 1s has historically been overlooked, as many Ethereum researchers focused on a rollup-centric roadmap. As the backbone of the entire crypto ecosystem, layer 1s are the key to scaling the EVM. To solve EVM’s scalability challenge, layer 1s must start rebuilding the EVM from scratch with performance in mind above anything else. 

The EVM faces severe network congestion and high gas prices as volume increases. It’s time for layer 1s to scale to onboard the next generation of users. Approaches such as parallelization will help improve throughput and, combined with transforming the EVM’s consensus mechanism and storage solutions, will set a new performance standard for the industry and establish a more developer-friendly environment for projects.

The proper solution to scaling the EVM 

For the past few years, Layer 2s have been presented as the answer to providing the cheapest and fastest way to execute transactions. Layer 2s are not what the EVM truly needs. From day one, Layer 1s have always been the true solution to the EVM’s scalability problem. 

It is time to be open to adopting more accurate performance metrics and divert attention to improving network performance. These changes will pave the way for the EVM to achieve its highest potential, introducing levels of scalability and efficiency never seen before. The EVM is here to stay, but its future depends on the industry to build.