Friday 27 March 2026, 03:04 PM

Ethereum's Glamsterdam upgrade: How EIP-7928 and zkEVMs will scale Layer-1

Ethereum's H1 2026 Glamsterdam upgrade uses EIP-7928 access lists and zkEVM proofs to replace transaction re-execution, exponentially scaling L1 throughput.

I’ve spent the last decade watching the familiar Silicon Valley cycle play out: we build something incredible, user demand skyrockets, the system bottlenecks, we apply a few band-aids, and eventually, we just have to rip out the plumbing and rewrite the core architecture.

That’s exactly where Ethereum finds itself right now. We are currently navigating the H1 2026 Glamsterdam upgrade, and honestly, it feels like the moment we finally trade our dial-up modems for a dedicated fiber connection. It’s the most consequential architectural shift we’ve seen since The Merge, and it pivots the network's focus aggressively back to Layer-1 scaling.

With network usage hitting all-time highs recently and ETH trading at a bit of a discount while everyone holds their breath for the pending BlackRock staked ETH ETF decision, Glamsterdam is arriving at the perfect time. We're looking at a potential 10x throughput increase—up to 10,000 TPS—by safely pushing the block gas limit to 200 million. Even better? We’re looking at an estimated 78.6% reduction in L1 gas fees.

Here is a look at the mechanics making this massive leap possible and why it opens up so many exciting possibilities for the ecosystem.

The death of sequential execution

Imagine trying to run a modern cloud backend where every single server has to sequentially re-calculate every single user action, one by one, before moving forward. It’s a nightmare for scalability. For years, Ethereum has operated on sequential transaction re-execution. Glamsterdam is tossing that out the window in favor of parallel processing and cryptographic proof verification.

The headliner for the execution layer is EIP-7928, which introduces Block-Level Access Lists (BALs). Core developers confirmed this approach back in August 2025, and it’s a brilliant piece of practical engineering. By forcing block builders to map all state interactions upfront, validators can pre-fetch data and execute 60-80% of non-conflicting transactions in parallel. Historical data analysis validated this, showing that BALs only add a modest average size overhead of roughly 70 KiB per block. That is a tiny footprint for a massive UX and performance gain.

The zkEVM breakthrough that changes everything

Parallel processing is great, but the real magic happens when you pair it with zero-knowledge proofs.

On January 26, 2026, the Ethereum Foundation’s zkEVM team published a concrete roadmap detailing the integration of zkEVM proofs into the core protocol via EIP-8025. This was the moment we officially shifted from theoretical whiteboard discussions to active implementation.

Why now? Because back in December 2025, the zero-knowledge proving ecosystem achieved the kind of 100x optimization milestone that makes tech nerds like us giddy. Proving latency collapsed from 16 minutes down to just 16 seconds. Modern zkVMs can now prove 99% of all Ethereum blocks in under 10 seconds on target hardware. Real-time Layer-1 proving is no longer a pipe dream; it’s a viable reality.

To make this work, EIP-7732 enshrines Proposer-Builder Separation (ePBS) into the consensus layer, extending the block validation window to 6-9 seconds. This crucial time buffer gives nodes the breathing room to opt-in as "ZK Validators" that verify zkEVM proofs instead of re-executing blocks.

Democratizing the validator layer

One of the most beautiful side effects of replacing heavy execution with lightweight proof verification is what it does for the little guy. By dropping the computational burden, we are democratizing hardware requirements. Independent stakers can now validate on consumer-grade machines, keeping the network decentralized and accessible.

And Ethereum isn't compromising its strict client diversity standards to get there. The protocol will enforce a multi-proof requirement where validators must verify 3 out of 5 independent proofs generated by different zkVM implementations—think SP1, RISC Zero, and Jolt—before accepting a block. It’s a highly resilient, fault-tolerant approach to network security.

Navigating the bumps in the road

Of course, a transition this massive carries elevated risks, and we have to navigate them thoughtfully.

Developers are actively working through the "free option" problem in ePBS and setting up mitigations against potential "prover killer" attacks designed to exhaust GPU hardware. There is also the very real challenge of ensuring the decentralized prover market remains economically viable so we don't accidentally hand control over to a few centralized institutions.

Fortunately, the core devs are showing excellent product management discipline. Just recently, on March 27, 2026, they officially voted to delay the highly anticipated "frame transactions" proposal from the subsequent Hegotá upgrade. It was marked as "under consideration" because its sheer complexity threatened to delay the broader network roadmap. In the startup world, we call that ruthless prioritization, and it's exactly what you want to see from a team managing a global settlement layer.

The transition to Glamsterdam will be phased, starting with a 10% opt-in rate this year before fully deprecating sequential execution once 128-bit provable security is formalized. It paves the way for true stateless clients in Hegotá and the eventual realization of "Native Rollups."

We are entering a wildly playful, open-ended era of Ethereum infrastructure. The sandbox just got a whole lot bigger, and I can't wait to see what we all build on top of it.