Ethereum has effectively been running in single-lane mode: one long queue of transactions, executed strictly in order. Even with multi-core CPUs, validation stays sequential. BALs (Block-level Access Lists, EIP-7928) change that. Each block carries a precise map of every account and storage slot it will touch, removing any guesswork during execution. They also include the post-transaction state diff for every transaction. Not just what a transaction accesses, but what the state becomes immediately after. With these diffs, clients can detect conflicts deterministically, resolve dependencies, and execute transactions in parallel. Parallelism becomes safe, predictable, and consistent across clients. BALs are slated for the Glamsterdam hardfork in 2026. This upgrade will be the foundation for several future scalability improvements across the execution layer. This is how Ethereum scales using the hardware everyone already has: BALs turn single-lane validation into multi-lane execution, unlocking far higher throughput without compromising decentralization. For more, visit blockaccesslist.xyz, read the EIP, or reach out with questions.

Nov 21, 2025 · 7:10 PM UTC

38
43
246
27,572
Sort replies: Relevant Recent Liked
Replying to @nero_eth
.xyz instead of .pics ? 🥺
1
5
267
One can call it a pivot 😂 No, .pics is more about transparency through dashboards while this is more an info page for the community, client implementers, testoors, researchers, etc. One place with everything relevant to follow along.
1
5
260
Replying to @nero_eth
Hey! There is pre-warming (TM) and optimistic execution (block-STM and friends). Both will still be relevant for old blocks, block production or blocks that already got their BAL's pruned!
1
2
229
Yes, true, good point!
151
Replying to @nero_eth
With zk proving of blocks coming up, why do BALs matter?
1
1
358
You can use the BAL to maintain state while _not_ reexecuting the full block but only verify the succint proof.
2
1
3
349
Replying to @nero_eth
Finally Ethereum is learning from Solana. Good to see progress.
1
44
Note that it's not only the state locations but in addition the BAL contains a post-tx state diff. The latter is essentially a tool to resolve dependencies among transactions.
1
38
Replying to @nero_eth
Such a good EIP! Gemini 3 gets it, I think. It generated this whiteboard diagram based on the EIP text.
4
3
33
1,919
Replying to @nero_eth
Glad to see progress on this front!
4
524
Replying to @nero_eth
What a great design
3
335
Replying to @nero_eth
parallel execution on eth is huge step
1
38
Replying to @nero_eth
Big step, could finally let Ethereum parallelize more and speed things up without breaking consensus.
66
Replying to @nero_eth
Pure throughput king
26
Replying to @nero_eth
Very cool to see this! Do you have any insights on how much this would change the requirements on validators?
58
Replying to @nero_eth
ethereum really be running like a one-way highway sometimes.
37
Replying to @nero_eth
BALs unlock true parallelism but the key is deterministic cross client state diffs and solid conflict resolution
2
111
Replying to @nero_eth
Huge win, finally safe parallel execution on Ethereum, this will unlock real throughput gains and finally make hardware matter
2
128
Replying to @nero_eth
such an amazing website
1
23
Replying to @nero_eth
You have described very clearly the current "single-lane" model, which is a bottleneck despite the capabilities of modern hardware. BALs is exactly the mechanism that puts Ethereum into multi-lane mode.
1
133
Replying to @nero_eth
Interesting point. Hope future updates will unlock more parallel processing for better scalability.
1
132
Replying to @nero_eth
eth zoom!
11
Replying to @nero_eth
This looks good
27
Replying to @nero_eth
The EVM is fundamentally limited ... HAHAHAHAHAHAHAHA Retards, pls lose your money on Soylana.
19
Replying to @nero_eth
wild, bal's basically handing validators a turbo button for state diffs eth nodes on a potato pc could validate blocks in seconds, no sweat
29
Replying to @nero_eth
Please (now) enable per txn account balance/nonce/code read index. It would let us move away from hacks like rsp-subblock for zkevm, while doing parallel execution shard generation. It's time we revisit this, and you did the analysis, it's only 2-3% overhead
38
Replying to @nero_eth
Multi-lane ETH gonna feel like flipping turbo mode on, fren
36
Replying to @nero_eth
Massive shift
7
Replying to @nero_eth
This flips validation into a beast mode upgrade.
56
Replying to @nero_eth
If this lands as designed, validators and dapps win big, can't wait to see the ecosystem build on it
6