When could 10k TPS on Ethereum Mainnet happen?

On an illustrative 3x YoY path, ~2031—subject to roadmap realities.

Why do we need Layer 1 scaling and Layer 2 scaling?

L1 is the credible neutral settlement layer; L2s provide horizontal scale for speed- and cost-sensitive use cases.

What does 10,000 TPS on mainnet actually mean?

≈1 Ggas/s assuming ≈100,000 gas per tx; a proxy for sustained throughput.

Why convert gas to TPS?

Gas/sec is protocol-native; dividing by ~100k gas/tx yields a comparable TPS estimate.

How do blobs and DAS matter?

Cheaper, abundant data availability lets rollups post more data per block, unlocking higher aggregate throughput.

What's the difference between Layer 2s and Rollups?

Rollups are a type of L2 that post data to Ethereum for security and availability; other L2s use different security models.

Will verification stay accessible?

That's the goal: scale without sacrificing verifiability or pushing hardware requirements too high.

Scaling Ethereum Mainnet to 10,000 TPS

Ethereum is on a clear path to scale. Over the next six years, Ethereum Mainnet throughput is expected to surge toward 10,000 transactions per second (TPS) - roughly 1 gigagas per second - while Layer 2s collectively push the ecosystem toward million-TPS capacity.

How far we've come

Since launch, Ethereum Mainnet has methodically improved efficiency and capacity without compromising decentralization or security. It went through several key upgrades, each contributing to incremental improvements in efficiency and capacity. You can read more about these on our ecosystem page.

From 2015 to today, Ethereum scaled from 0.71 TPS to 24.9 TPS, a 35.1x increase.

Where we're going next

After years of steady gains, the pace is set to accelerate. The goal is to scale by 3x per year with upcoming improvements. This takes today's 24.9 TPS into the thousands before decade's end.

How Ethereum Mainnet scales from here

Ethereum's strategy combines multiple approaches to sustainably increase capacity while preserving its core principles:

EIP-7938 (Dankrad Feist) - proposes a default, exponential gas-limit growth schedule where clients vote automatically to increase L1 capacity over time (subject to coordination and override). Read the spec (here)[https://eips.ethereum.org/EIPS/eip-7938].
Lean Ethereum (Justin Drake) - a design philosophy to streamline consensus, data, and execution, leveraging DAS and real-time zkVMs for “beast mode” performance while staying verifiable. (More)[https://blog.ethereum.org/2025/07/31/lean-ethereum].
More EIPs - parallel efforts improve execution, networking, and data availability. (Slide overview)[https://docs.google.com/presentation/d/1vDkFlUXJ6S94hOGjMkeIO-SP3ZB4ysKPNgu1mskf1No/edit#slide=id.g36e7baa00a101116].

The aim isn't raw TPS alone - it's sustainable, decentralized scale that remains easy to verify.

Horizontal scale with Layer 2s

In addition to scaling execution on Ethereum Mainnet, the Ethereum Ecosystem also scales horizontally via Layer 2 solutions. Layer 2s that handle execution off-chain and post data to Ethereum are called Rollups. As blob capacity grows and data availability sampling matures, Rollups can publish more data per block and scale safely.

These improvements unlock orders of magnitude of additional throughput - toward 1,000,000+ TPS at the ecosystem level.

Conclusion

Ethereum's path to 10k TPS on L1 and million-TPS with L2s is driven by pragmatic upgrades and a modular roadmap. The result: lower fees, higher capacity, and preserved neutrality and verification for users and developers.

Assumptions

100,000 gas per transaction for converting gas/sec → TPS
Illustrative 3x YoY L1 growth curve; 4x YoY for aggregate L2s