Inside Ethereum’s race to achieve 10 million TPS

Ethereum is picking up speed again, not in price charts where it is currently struggling, but in the silent machinery beneath it.

Two parallel breakthroughs, one on the protocol layer and the other in cryptography, are redefining how fast and how light the world’s most-used blockchain can run.

Together, they sketch a future where anyone, from institutions to small-time validators, can participate on the network in real time without needing supercomputers or deep pockets.

Fusaka upgrade

The first significant milestone on that path is Fusaka, Ethereum’s upcoming hard fork, tentatively expected in December.

The planned upgrade blends improvements to Ethereum’s execution and consensus layers in one coordinated release.

Unlike Dencun, which introduced “blobs” to help rollups scale, Fusaka isn’t chasing raw throughput.

Instead, its role is subtler, focusing on making the network lighter, cheaper, and more efficient.

Fusaka implements 12 Ethereum Improvement Proposals (EIPs) aimed at streamlining validator workloads and improving how rollups post their data.

The centerpiece, EIP-7594, or PeerDAS, lets validators confirm data availability by sampling portions of rollup data instead of downloading it fully.

While this doesn’t directly raise TPS, it changes how efficiently Ethereum handles data. More rollup information can now fit per block without increasing node requirements.

Developers expect the upgrade to lower rollup transaction costs and make it easier for small operators to run validators.

Notably, it also raises the gas limit from 45 million to 60 million, a 33% bump that gives Layer-2s more headroom to publish compressed transaction data.

Meanwhile, the rollout is already underway. Fusaka passed early tests on Holesky and Sepolia, and will undergo its final trial on the Hoodi testnet later this month.

Real-time proving

While Fusaka lays the groundwork, the real spectacle is happening in the proving arena.

On Oct. 15, Ethereum scaling firm Brevis unveiled Pico Prism, a new zero-knowledge Ethereum Virtual Machine (zkEVM) capable of producing cryptographic proofs almost as fast as the network creates blocks.

In testing, the system achieved 99.9% real-time proving, generating full block proofs in under 12 seconds.

Ethereum Foundation researcher Justin Drake pointed out that this represents a leap from May’s performance, when the SP1 Hypercube setup could only prove 94% of blocks within the same window.

According to him, the improvement cuts average proof latency to 6.9 seconds, meaning block verification can keep pace with block production. Notably, this is a prerequisite for Ethereum’s long-term goal of sub-second settlement.

Drake furthered that this development, alongside the imminent Fusaka upgrade, will make on-premise proving viable for the first time.

He said:

Scalability roadmap

Drake believes these developments fit into his long-term projection of “gigagas L1, teragas L2.”

In this scenario, Ethereum’s throughput on its base layer for high-value activities like payments and trading rises to 10,000 transactions per second (TPS).

On the other hand, the network can scale up to 10 million TPS across its layer-2 networks to handle everything else. Drake said:

Rising technical debt

Ethereum’s march toward faster, cheaper transactions comes with a quieter problem of its technical debt piling up.

Ethereum developer Federico Carrone, better known as Fede’s Intern, cautions that many of the network’s core development tools, especially the Solidity programming language, are losing momentum.

Solidity is the foundation of Ethereum’s DeFi ecosystem. According to DeFiLlama, it is responsible for more than 86% of the smart contract language used in the blockchain network’s over $200 billion DeFi protocols.

His concerns echo those of Paradigm CTO Georgios Konstantopoulos, who had previously said Solidity’s ecosystem was “in a problematic state.”

However, Carrone sees the issue as technical and economic for the blockchain network.

He argued that maintaining complex infrastructure depends on time, continuity, and deep expertise, which cannot be obtained overnight.

In addition, Carrone noted that Ethereum’s planned gas-limit increase under the Fusaka upgrade poses another risk.

Carrone warned that many execution clients have not significantly improved their performance and may struggle to process larger blocks.

Considering all these issues, he concluded: