Using Segmented Shards to Scale Ethereum

Image by Ideogram

Years ago, when Ethereum was still using the proof of work consensus mechanism, the roadmap included scaling through layer two blockchains and eventually sharding. Ethereum was designed to be the world’s computer, its use was expected to grow over time which it has.

At the end of 2025, the Fusaka upgrade saw the block gas limit raised from 36 million to 60 million. A quick glance of beaconcha.in/blocks shows that blocks use about 40–60% of that limit.

Transaction fees are not that high at the moment, they are in the range of $0.10 to $0.50 which is much less than in 2020 during the height of the NFT fad when transactions could cost up to $50. The problem is that while those fees are reasonable for a transfer of hundreds to thousands to millions of dollars in value, they are still too high for daily payments.

For Ethereum to become an accepted payments solution, it will have to address privacy concerns. Ethereum transfers are currently pseudonymous which means it is difficult but not impossible to determine the parties involved in a transaction.

Introducing privacy will be computationally expensive. Any private transaction will use a lot more gas than the average transaction today. In addition, quantum-resistant transactions will also be gas intensive and it is only a matter of time before a cryptographic relevant quantum computer is introduced to the world.

The point is that while transaction fees may seem low at the moment, they are only low relative to the value of the underlying transactions and they will increase significantly in the coming years as more gas will be consumed in individual transactions. Gas prices are subject to supply and demand so as more gas is required for a transaction, the higher the prices will be.

Earlier this year, Vitalik called out the layer two chains for not being sufficiently decentralised and rightfully pointed out that Ethereum was scaling on its own. I think that if Ethereum wants to scale at a level that will meet the sort of demand that mainstream usage will create, then a discussion needs to start about a more aggressive scaling. This is where sharding can help.

Sharding, in its original intended form, is running parallel chains instead of the single chain. This was not possible with the proof of work design. Proof of staking enables sharding because it split consensus from execution. Consensus occurs on the beacon chain. With sharding, the parallel chains can all be run separately but still share the same beacon chain.

T here are currently 893,807 validators as of June 1, 2026. This is an incredible number and shows the true value of Ethereum with its decentralisation. This also means that there are more than enough validators for sharding to occur. With sharding, a validator would only be concerned with a parallel chain instead of one main chain. It would be a distribution of validators that would reduce the workload on individual nodes.

Sharding could even be used to introduce a sector-based model. Instead of a single chain that had all transactions, there could be 6 specialised chains. For example, the 6 chains could be for regulated activities, cryptocurrencies, identity, equity DeFi, debt DeFi, and governance. Each parallel chain would be segmented.

By segmenting the chains, each chain could then focus on their direct needs. Pre-compiles and core changes could be introduced that benefit that segment without affecting any of the other chains. Chains could adapt to their segments needs; with a single chain these sorts of needs might be ignored if they presented no value for other segments or if they conflicted with the needs of other segments.

Segmented chains would also change how gas pricing works. If one chain saw a lot of activity resulting in the gas price increasing, the other chains would still have low gas. If this was available during the NFT boom, then anyone trading NFTs could have still paid $50 for a transaction while other users involved in governance could have paid significantly less fees while participating in their protocols.

Image by Ideogram

Sharding could also be beneficial to the layer two ecosystem. The blob capacity could increase and enable layer two chains that are specialised, provide more data availability, and reduced rollup costs.

The model I would suggest is to require a validator to support two of the chains. So if the sharding model started with 6 parallel chains, then 6 times the amount of transactions could be managed while validators only see a doubling in the number of transactions that they would process. With the current numbers, that would mean about 297,935 validators per chain which is still a large and respectable number.

With my hypothetical model, there would be a risk that validators favour one chain over another because of higher gas costs. First, if there was an imbalance in validators between chains, then the average reward would be affected too since more validators would mean less opportunities to propose a block. But the model could also include weighted rewards to further disincentivise one chain over another. Chains with fewer validators could pay slightly higher rewards to encourge more validators to migrate.

Cross-shard communication would be difficult but not impossible. If all validators had to support two chains in the six chain model and they were randomly arranged, then there would be 15 different combinations of validator set ups which means with today’s numbers there would still be about ~60K validators supporting any two given chains. So if a transaction required validating a state on a parallel chain, there would be plenty of validators that could attest to that state.

One obvious risk would be the original double spend problem that Bitcoin solved all those years ago. Ethereum balances are the sum of their transactions but simultaneous transactions on parallel chains could provide an opportunity to cheat the system. To get around this, I would suggest that balances are allocated across shards.

When a user holds ether in their wallet, they can choose to allocate some or all of it to one of the chains. There can be a cross-chain transfer type transaction that could enable users to change their balance allocations. The difficulty here would be that any wallet may have to track balances across six chains but I doubt there would be many users actively transacting on multiple chains and even then they would probably be better off using multiple wallets instead.

With this model, future expansion could be easier to implement by changing the minimum number of chains per validator. In the 6–2 model I described above, a 9–3 model could be introduced by requiring all validators to support a third new chain. It wouldn’t even need to start with the 6–2 model, an 8–2 model could be used with future expansion planned in a 12–3 model.

Overall, if Ethereum really wants to be the world’s computer and place itself as infrastructure for global digital economies, then it needs to start planning for more aggressive scaling than just increasing the block gas limit. By using segmented shards, not only could there be a lot more scaling but localised gas fees could mean that hyperactivity in one segment would not drive up prices for other industries.

Sharding provides an opportunity for innovation to continue in Ethereum without pricing out other silos. It can keep the cost of using the network low so more people can benefit from the technology.

---

Using Segmented Shards to Scale Ethereum was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.