Ethereum enters the super scaling era: What does it mean when L2 throughput surpasses 3,700 ops/sec?

In April 2026, the overall throughput of Ethereum Layer 2 networks first surpassed 3,700 operations per second (ops/sec), representing an increase of over 210% compared to the same period in 2025. This milestone was directly driven by the collaborative optimization of data availability (DA) and execution layers through two core upgrades: Pectra (May 2025) and Fusaka (December 2025). Meanwhile, the Fusaka upgrade extended the destruction mechanism to Blob transactions, pushing the annualized ETH destruction rate from 0.89% to 1.32%. On the fee side, the cost for a single transfer on mainstream L2 networks has fallen to between $0.002 and $0.008, while swap operation fees are approximately $0.01–$0.03, representing reductions of 40%–90%.

What technologies drove the breakthrough of L2 throughput beyond 3,700 ops/sec

The Pectra upgrade includes 11 Ethereum Improvement Proposals (EIPs), making it the largest hard fork since The Merge. Notably, EIP-7691 increased the target number of Blobs per block from 3 to 6, with a hard cap rising from 6 to 9, directly expanding the available channels for L2 to submit data to L1. Additionally, Pectra adjusted other parameters to raise the target Gas limit from 15M to 22.5M, nearly doubling the batch submission capacity for major L2s like Arbitrum, Optimism, and Base. More importantly, the compression algorithm used by L2 sequencers was unified and optimized, increasing the average compression rate of transaction call data before submission to L1 from 32% to 47%.

The Fusaka upgrade further advanced this process. Its core component, PeerDAS (Peer Data Availability Sampling), allows each node to store only 1/8 of Blob data and employs erasure coding, theoretically increasing Blob throughput by 8 times while keeping validator bandwidth and storage burdens manageable. The BPO (Blob-Parameter-Only) fork mechanism also enables Ethereum to independently adjust Blob parameters in phases—progressing from the basic 6/9 to 12/15, and then to 14/21—without waiting for a major annual upgrade. These technological iterations collectively pushed L2’s comprehensive throughput to a new high of 3,700 ops/sec, covering various operations such as cross-chain messaging and state updates.

What mechanisms caused L2 fees to decrease by 40%-90%

Fee reductions are the most direct market feedback from the Pectra and Fusaka upgrades. According to Gate market data (as of April 16, 2026), the average gas price on the Ethereum mainnet remains stable between 8–15 Gwei, while the cost for a single transfer on L2 networks has dropped to between $0.002 and $0.008, with swap operation fees around $0.01–$0.03.

The change in fee structure stems from two core mechanisms. First, expanding Blob data space directly reduces the competitive costs when L2 submits batches to L1. After doubling Blob capacity, the gas cost for L1 data availability dropped below approximately 1 Gwei, with ZK-rollup network fees decreasing by 78%–91%. Second, EIP-7702 smart accounts introduced batch transaction aggregation, allowing users to pay only once for multiple operations (e.g., approval + swap + staking). This improvement lowers the entry barrier for external accounts, enabling wallets to execute smart contract functions and pay gas fees with stablecoins. For high-frequency DeFi users and on-chain gamers, daily interaction costs have decreased from $2–$5 to $0.2–$0.5, directly boosting active addresses.

How the increase in destruction rate to 1.32% impacts ETH’s economic model

The Fusaka upgrade’s economic model change is mainly reflected in EIP-7918. This proposal links Blob base fees with execution layer Gas fees, ensuring that Blob transactions must pay a minimum fee even during low demand, preventing near-free usage of Blob. More importantly, prior to Fusaka, Blob transactions paid only base fees without participating in destruction; after Fusaka, 30% of the base fee in Blob transactions is incorporated into the EIP-1559 destruction mechanism. This adjustment increased ETH’s annualized destruction rate from 0.89% before the upgrade to 1.32% (as of April 15, 2026). Based on current ETH prices (according to Gate data as of April 16, 2026), the daily destroyed ETH value is approximately $3.8 million.

The rise in destruction rate has two structural impacts on Ethereum’s economic model. First, the possibility of the net issuance rate turning negative increases. If daily destruction exceeds validator rewards consistently, ETH supply enters a deflationary phase, reinforcing long-term holder deflation expectations. Second, the operational cost structure of L2 changes—sequencers need to balance throughput and destruction costs, leading some L2s to adjust batch submission frequency to optimize expenses. It’s important to note that the increased destruction rate does not necessarily mean higher user costs, as the absolute Blob fee values remain far below pre-upgrade calldata costs.

What does the 26% growth in DeFi TVL on L2 ecosystem reflect about capital flows

As of April 15, 2026, the total value locked (TVL) in Ethereum L2 DeFi ecosystems reached $38.7 billion, a 26% increase compared to the same period in 2025. This growth rate exceeds the 14% increase in Ethereum mainnet DeFi, indicating capital migration from mainnet to L2. In terms of ecosystem distribution, peak TPS on major L2 networks has stabilized above 1,200, with transaction volumes on leading networks like Base and Arbitrum continuing to rise.

The shift in capital flows reflects the evolution of the L2 competitive landscape. The significant fee reductions lower entry barriers for users, while improved cross-chain interoperability enables more efficient liquidity transfer across different L2s. Notably, fee reductions are activating high-frequency application scenarios previously hindered by high costs, including on-chain order book DEXs, decentralized gaming, and micro-payment systems. Some analyses suggest that Fusaka’s upgrade could further reduce L2 data costs by an additional 40%–60%, which is especially beneficial for high-volume sectors like DeFi and blockchain gaming.

What are the real feedback from developers and application layers on these upgrades

From the developer perspective, Pectra and Fusaka are transforming L2 application development paradigms. EIP-7702’s account abstraction capabilities enable wallets to support gas sponsorship, stablecoin fee payments, and batch transaction aggregation, reducing the cognitive load for ordinary users of crypto applications. Some L2 project teams report that doubling Blob capacity provides more low-cost space for DEX and gaming users, while zero-knowledge proof languages like Cairo may see shorter proof generation cycles as costs decline.

However, the upgrades also introduce new technical challenges. Data from research firm MigaLabs shows that after Fusaka, blocks with 16 or more Blobs exhibit higher missing rate, with the highest observed at 21 Blobs, where the missing rate exceeds three times the network average. This indicates that Ethereum still faces bottlenecks under extreme data loads, and further upward adjustment of Blob parameters must be approached cautiously. Meanwhile, Ethereum co-founder Vitalik Buterin publicly questioned early 2026 that some L2 networks have not truly scaled Ethereum, criticizing their increasing reliance on centralized components, which could undermine the core security and decentralization principles of the mainnet. These debates suggest that the L2 scaling roadmap remains in active revision.

What challenges and opportunities does the future scaling roadmap face

Following Pectra and Fusaka, Ethereum’s scaling roadmap has entered a new phase. According to official plans, the first half of 2026 will see the rollout of Glamsterdam, focusing on improving execution layer efficiency and fairness in block construction; the second half will introduce Hegotá, further optimizing underlying infrastructure. Strategically, Ethereum is shifting from a “rollup-centric” approach toward a dual-track model of “L1 settlement + L2 execution.” L1 emphasizes maximum security and decentralization, while L2 handles execution and throughput expansion.

However, challenges remain. First, Blob expansion faces network stability limits—rapid parameter increases may cause higher block missing rates, affecting overall reliability. Second, decentralization within the L2 ecosystem varies; some sequencers are still controlled by single entities, which conflicts with Ethereum’s core values. Third, as L1’s own throughput continues to improve, the narrative of L2 necessity might be re-evaluated. These issues will be focal points in the Glamsterdam and Hegotá upgrades, engaging the developer community.

Summary

The two major upgrades, Pectra and Fusaka, mark Ethereum’s transition from “concept validation” to “scaling deployment.” The breakthroughs in L2 throughput exceeding 3,700 ops/sec, fee reductions of 40%–90%, and the increase of destruction rate to 1.32% collectively point to a core conclusion: Ethereum is evolving through a “dual architecture of L1 settlement + L2 execution,” maintaining security and decentralization while enabling large-scale applications with high throughput and low fees. Nonetheless, network stability, L2 decentralization, and the economic relationship between L1 and L2 require ongoing optimization. The upcoming Glamsterdam and Hegotá upgrades in 2026 will be critical in determining whether this roadmap can progress from “feasible” to “sustainable.”

FAQ

Q: What is the difference between the 3,700 ops/sec L2 throughput and typical TPS?

3,700 ops/sec (operations per second) includes not only regular transactions but also cross-chain messaging, state updates, data availability sampling, and other on-chain operations. This metric provides a more comprehensive reflection of the actual processing capacity of the L2 ecosystem than simple transaction TPS. Mainstream L2 networks’ peak TPS has stabilized above 1,200, while the overall throughput aggregates all L2 operations.

Q: What are the practical impacts of Pectra and Fusaka upgrades on ordinary users?

The most immediate change is a significant reduction in transaction costs—single transfers now cost between $0.002 and $0.008, swap operations around $0.01–$0.03. Additionally, EIP-7702’s smart accounts allow users to pay gas fees with stablecoins like USDC and support batch transactions, lowering the interaction costs for multi-step operations.

Q: Does the increase in destruction rate to 1.32% necessarily mean ETH will become deflationary?

An increased destruction rate raises the possibility of ETH supply shrinking, but whether it enters a deflationary phase depends on whether daily destruction consistently exceeds validator rewards. After Fusaka, 30% of the base fee in Blob transactions is burned, raising the annualized destruction rate from 0.89% to 1.32%. Currently, ETH remains near the boundary between mild inflation and deflation.

Q: With L2 fees already very low, is there room for further reduction?

Yes. The PeerDAS and BPO mechanisms introduced by Fusaka provide the technical foundation for ongoing Blob capacity expansion, with theoretical throughput potentially increasing by up to 8 times. Analysts estimate that, as these mechanisms are deployed gradually, L2 data costs could decrease by an additional 40%–60%. However, fee reductions will need to be balanced carefully with network stability and decentralization considerations.