In April 2026, the combined throughput of Ethereum Layer 2 (L2) networks surpassed 3,700 operations per second (ops/sec) for the first time, marking a year-over-year increase of more than 210% compared to 2025. This surge was directly driven by the Pectra upgrade, which delivered dual optimizations for data availability and the execution layer. Pectra introduced an enhanced version of EIP-4844, doubling the number of Blobs per block from 6 to 12, and, through EIP-7691, raised the target gas limit from 15 million to 22.5 million.
These parameter changes nearly doubled the batch submission capacity for leading L2s like Arbitrum, Optimism, and Base. More importantly, L2 sequencers saw unified improvements to their compression algorithms, increasing the average compression rate of transaction call data before submission to L1 from 32% to 47%. These technical advancements translated directly into user experience: peak TPS on major L2 networks now consistently exceeds 1,200, while aggregate throughput (including cross-chain messaging, state updates, and other operations) has reached 3,700 ops/sec.
How the Pectra Upgrade Cut Mainstream L2 Transaction Fees by 40% to 90%
The most immediate market response to the Pectra upgrade has been a drop in fees. According to Gate market data (as of April 15, 2026), average gas prices on Ethereum mainnet remain stable between 8–15 Gwei, but single transfer fees on L2s have fallen to just $0.002–$0.008, with swap operations costing around $0.01–$0.03. Compared to pre-upgrade levels, fees on Optimism-based networks have dropped about 42%, while ZK-rollup networks have seen reductions of 78%–91%. Two main mechanisms drove this change: first, the expanded Blob data space lowered the competition cost for L2s submitting batches to L1; second, EIP-7702 introduced batch transaction aggregation for smart accounts, allowing users to pay L2 fees just once for multi-step actions (like approve + swap + stake). For high-frequency DeFi users and on-chain gamers, daily interaction costs have dropped from $2–$5 to just $0.2–$0.5, directly boosting the number of active addresses.
Annualized Burn Rate Rises to 1.32%: What On-Chain Economic Changes Did the Fusaka Upgrade Bring?
The Fusaka upgrade was deployed in Q1 2026, with its core change being the extension of EIP-1559’s burn mechanism to include Blob transactions submitted from L2 to L1. Previously, Blob transactions only paid a base fee and were excluded from burning; after Fusaka, 30% of the base fee from Blob transactions is now burned. This adjustment pushed Ethereum’s annualized burn rate from 0.89% pre-upgrade to 1.32% (as of April 15, 2026). At the current ETH price (per Gate market data, around $2,100), this equates to about $3.8 million worth of ETH burned daily. The increased burn rate has two main effects on the network’s economic model: first, it raises the likelihood of net issuance turning negative, strengthening deflation expectations for long-term holders; second, it changes the cost structure for L2 operations—sequencers must now rebalance between throughput and burn costs, with some L2s adjusting batch submission frequency to optimize expenses. Importantly, a higher burn rate does not mean higher user costs, as absolute Blob fees remain far below pre-upgrade Calldata fees.
DeFi TVL Up 26% Year-over-Year: Where Is the New Capital Flowing in L2 Ecosystems?
As of April 15, 2026, total value locked (TVL) in Ethereum L2 DeFi ecosystems reached $38.7 billion, a 26% increase from the same period in 2025. This growth outpaces Ethereum mainnet DeFi, which grew by 14%, signaling a migration of capital to L2s. In terms of distribution, Arbitrum still commands 41% of L2 TVL, but Base and ZKsync Era have seen their shares rise from 12% to 18% and from 7% to 13%, respectively, over the past six months. The three main destinations for new capital are: RWA (real-world asset) protocols deployed on L2s, with quarterly TVL growth of 47%; perpetual DEXs, which have seen daily trading volume rise to $2.2 billion thanks to lower fees; and liquid restaking protocols, whose L2 deployments offer annual yields 1.2–1.8 percentage points higher than mainnet versions, attracting about $1.9 billion in net inflows. Notably, as TVL grows, cross-chain bridge liquidity between L2s has increased by 63%, and the cost of moving funds across L2s has dropped below $0.05, further solidifying L2s as a unified liquidity layer.
How Pectra’s EIP-7702 Smart Accounts Lower the Barrier for Everyday Users
The most user-facing change in the Pectra upgrade is standardized support for smart accounts (EIP-7702). This proposal allows externally owned accounts (EOAs) to temporarily gain smart contract execution capabilities without deploying a new contract account. For users, this means: the ability to set automatic gas limit adjustment rules to avoid failed transactions during network volatility; support for social recovery (resetting private keys with approval from 3–5 guardians); and batch authorization and transaction merging. According to Dune Analytics, 90 days after EIP-7702 went live, 1.87 million addresses on L2s had enabled smart account features—23% of all active addresses. For new users, creating a smart account is nearly identical to creating a traditional EOA, but with significant improvements in security and convenience. This change has directly lowered the psychological barrier for non-technical users to enter crypto, and is a key driver behind the 41% growth in active L2 addresses over the past three months.
From Glamsterdam to Hegotá: Ethereum’s Roadmap to 10,000 TPS
Pectra and Fusaka are short-term scaling solutions, but Ethereum developers aim to surpass 10,000 ops/sec L2 throughput with the Hegotá upgrade (expected Q1 2027). The roadmap unfolds in three phases: Glamsterdam (Q3 2026) will introduce an initial version of parallel execution, allowing non-conflicting transactions to be processed simultaneously within the same block—projected to boost L1 gas throughput by 30%–40%. Next, Hegotá’s core upgrades—State Lease and History Expiry—will reduce full node storage requirements, lowering the barrier to running a node by over 60% and paving the way for decentralized sequencers. The final goal is native L2 cross-chain interoperability, where L2s can communicate directly without routing messages through L1, reducing latency from the current 10–15 minutes to under 12 seconds. Glamsterdam’s testnet is already live, with parallel execution reducing simulated block processing times from 2.1 seconds to 1.3 seconds.
After the Throughput Breakthrough: Remaining Bottlenecks and Risks for Ethereum L2 Scaling
Despite impressive metrics, Ethereum L2 scaling still faces three unresolved structural challenges. First, data availability (DA) layer dependency: over 80% of L2 transaction data currently relies on Ethereum mainnet Blob space, so a spike in Blob demand could cause fees to rebound sharply. Second, sequencer decentralization lags behind: most leading L2s still use a single sequencer node, posing risks of transaction censorship and centralized MEV extraction. Third, cross-L2 interoperability standards remain ununified—assets transferred between L2s still require 7–15 minutes for finality and carry smart contract risks from third-party bridges. On the governance side, some EIPs in the Pectra upgrade (such as EIP-7623, which adjusts call data fees) have sparked community debates over whether L2s are "overusing mainnet resources." How quickly these bottlenecks are resolved will directly impact whether Ethereum L2s can hit the 10,000 ops/sec target by 2027, or if they’ll be partially replaced in certain scenarios by monolithic chains like Solana or modular DA solutions like Celestia.
The High-Performance L2 Era: Is the Competitive Logic for Other Blockchains Changing?
Ethereum L2s breaking through 3,700 ops/sec is reshaping the competitive landscape among blockchains. In the past, performance was the main argument for challenger chains to attack Ethereum; now, L2s’ aggregate performance rivals or even exceeds Solana’s single-chain TPS (roughly 2,500–4,000 actual TPS). The focus is shifting from "how many transactions per second" to "the scale and security of ecosystem assets." Currently, Ethereum L2s support DeFi TVL that’s 4.7 times larger than the second-largest blockchain, and stablecoin issuance accounts for over 56% of the market. This gap is difficult to close by simply increasing TPS. On the other hand, modular chains (like Celestia + Eclipse) are trying to offer similar experiences at lower DA costs, with per-GB data publishing costing about 1/15th of Ethereum’s rate. However, these solutions still lag in security assumptions (requiring trust in external DA validator sets) and ecosystem maturity. Overall, Ethereum L2’s performance leap hasn’t ended competition—it’s shifted the focus to "balancing security and performance" and "the cost of developer and asset migration."
Summary
In April 2026, Ethereum L2s reached a historic high of 3,700 ops/sec in aggregate throughput. The Pectra upgrade, through Blob scaling and EIP-7702 smart accounts, slashed transaction fees by 40%–90% while lowering the barrier for new users. The Fusaka upgrade extended the burn mechanism to Blob transactions, raising the annualized burn rate to 1.32%. Over the past year, DeFi TVL grew 26%, with new capital clearly flowing into L2 ecosystems. The Glamsterdam and Hegotá upgrades are set to introduce parallel execution and state leasing, targeting 10,000 ops/sec. However, data availability dependency, lagging sequencer decentralization, and missing cross-L2 interoperability standards remain critical bottlenecks. High-performance L2s are reshaping blockchain competition, but asset scale and security remain Ethereum’s strongest moat.
FAQ
Q: What are the core changes in the Pectra upgrade? Do regular users need to take any action?
The Pectra upgrade includes three main changes: increasing Blob count from 6 to 12, raising the mainnet gas limit to 22.5 million, and standardizing EIP-7702 smart accounts. Regular users don’t need to do anything—L2 networks will automatically offer lower fees. If you want to use social recovery or batch transaction features of smart accounts, you can proactively upgrade your account type in supported L2 wallets.
Q: Does an annualized burn rate of 1.32% mean ETH is guaranteed to be deflationary?
Not necessarily. The 1.32% annualized burn rate only refers to the proportion of ETH burned via EIP-1559 and Fusaka mechanisms relative to circulating supply. Deflation also depends on validator staking rewards. Currently, ETH’s net annual inflation rate is about 0.22% (issuance of 1.54% minus burn of 1.32%). Only if the burn rate consistently exceeds 1.54% will the total supply actually decrease.
Q: L2 fees are already very low—why keep scaling to 10,000 ops/sec?
Low fees and high throughput solve different problems. The current fee reduction is mainly due to expanded Blob space, but if active L2 users grow from millions to tens of millions, Blob demand could push fees up again. The 10,000 ops/sec goal is designed to support applications like on-chain gaming, decentralized social, and high-frequency trading that require thousands of operations per second, while also accommodating user growth over the next 5–10 years.
Q: Did the Fusaka upgrade increase operating costs for L2 projects?
Yes, but the increase is manageable. Since Fusaka now burns 30% of Blob base fees, about 30% of the total daily Blob costs paid by L2 projects are burned and not received by validators. At current Blob fee levels, mainstream L2s have seen daily operating costs rise by about 12%–18% on average. However, because Blob fees are still over 90% lower than pre-upgrade Calldata fees, most L2s haven’t changed the fees they charge users.
