Efforts to agree on a shared account abstraction standard between Ethereum and Base collapsed last week, and wallets will now have to support two separate transaction formats.
Ethereum and Base are set to implement different account abstraction standards after efforts to agree on a shared standard broke down last week, a split that pushes the burden of cross-chain compatibility onto wallet developers and underscores a growing philosophical divide between layer-1 and layer-2 networks.
Derek Chiang, founding member and researcher at Ethlabs and a co-author of Ethereum’s EIP-8141 proposal, revealed the breakdown in a Monday post on X, saying interoperability standards became secondary to each chain’s core goals, leading both networks to go their separate ways and “putting the burden on wallets.” Account abstraction allows programmable rules for authorizing transactions and paying fees, the technology behind smart-contract wallets, gas sponsorship, session keys and social recovery.
The practical consequence: wallet developers who want to offer a consistent experience across Ethereum and Base may need to support two distinct native account abstraction formats, testing, auditing and maintaining each.
Two chains, two standards
Ethereum is now advancing Frame Transactions under EIP-8141 as a headliner item of its Hegotá upgrade. The proposal introduces native account abstraction to the base layer and creates a path toward post-quantum authentication, an increasingly prominent concern among protocol researchers. The Ethereum Foundation named EIP-8141 among the must-ship items for Hegotá in a Sept. 7 blog post, elevating it alongside the upgrade’s headline scalability work.
Base, meanwhile, is developing native account abstraction via Keystore under EIP-8130, a specification that is currently live on devnet. The Base approach is optimized for the L2’s scalability-first design, with keystore-style accounts intended to make smart wallet operation cheaper and simpler at high transaction volumes.
Chiang argued the split reflects different priorities rather than simple disagreement. Layer-1s are increasingly focused on censorship resistance, capture resistance, open-source guarantees, privacy and security features, priorities that favor Ethereum’s more conservative, security-first account design. Scalability-focused L2s align more naturally with standards such as EIP-8130, where throughput and cost dominate the requirement set.
The researcher also pushed back on the idea that divergence is inherently bad. Both Ethereum and Base are now “free to innovate on AA to the maximal extent in accordance with their own visions,” he wrote, suggesting that forcing a compromise standard would have diluted both designs.
The interoperability bill comes due at the wallet layer
The cost of that freedom does not disappear; it migrates. Wallet developers, from browser extensions to mobile smart-wallet startups, now face a matrix of account abstraction implementations across the two largest ecosystems by transaction volume. Each format carries its own validation rules, fee-payment semantics and signature schemes. Teams that previously targeted a single emerging standard must now budget for parallel implementations, or ship different wallet capabilities depending on which network a user is on.
The episode is a test of a long-standing claim in the Ethereum community: that the ecosystem’s composability extends across layers through shared standards rather than protocol mandates. With Ethereum and Base, the two most-used networks in the space by many measures, failing to converge on this piece of infrastructure, other L2s are likely to face pressure to pick a side, deepening the fork rather than healing it.
The roadmap context
The split lands at a consequential moment for Ethereum’s upgrade pipeline. Hegotá, which could begin implementation in late 2026 following Glamsterdam, would carry EIP-8141 onto the base layer. Glamsterdam itself, expected on mainnet in the second half of 2026 according to the public roadmap, is designed to improve scalability, harden the L1 and make the network easier to use, and is widely regarded as one of the most consequential upgrades of the year.
Native account abstraction has been a community goal for the better part of a decade, progressing through the ERC-4337 mempool-based workaround and a succession of RIP and EIP proposals. EIP-8141’s elevation to headliner status represents the culmination of that arc, embedding programmable accounts into the protocol itself rather than simulating them with a parallel transaction mempool.
What to watch
Three things will determine how costly the split proves. First, whether wallet SDK providers abstract the difference, letting application developers write once and deploy across both standards. Second, whether Base’s EIP-8130 spec stabilizes on devnet or converges back toward Ethereum’s design as implementation feedback accumulates. Third, whether other layer-2s, Arbitrum, Optimism, Linea and others, publicly align with one camp, which would harden the fork into two de facto account ecosystems.
Chiang’s framing suggests the ecosystem is choosing deliberate divergence over forced convergence. For users, the experience may barely change at first. For the infrastructure layer underneath, the account abstraction unification story just got materially harder.
as a wallet dev this is the worst possible outcome. now i get to implement frame transactions for EIP-8141 AND whatever base ships, twice the audit bill, because two teams could not share a spec
two audit tracks for the same feature is just burning money on duplicate work. wonder how many wallets quietly drop base support and call it a day
already happening. two wallets i use froze base AA support last sprint and neither mentioned it in a changelog
two wallets froze base AA with zero changelog mention and users still blame the wallet when the tx fails. support nightmare incoming
Chiang straight up admitted interop became secondary and people are still shocked wallets are bailing. maintaining two AA stacks is real engineering cost, dropping the smaller chain is just rational budgeting
quietly dropping base support is already the base case, pun intended. small teams will ship whichever chain has the users and file the other one as a permanent backlog item
the audit bill point is real. my team is already scoping a feature flag just to ship both formats, we eat the cost so l1 and l2 can finish their design debate
the feature flag tax is real. we budgeted a full quarter for dual format support and that was the optimistic estimate
two sprints was our estimate and we are nine weeks in. dual AA format support is becoming the new ie6 of wallet dev, permanent backlog nobody can kill
ie6 of wallet dev is exactly it. we feature flagged frame transactions behind a config and the QA load doubled overnight
nine weeks into a two sprint estimate, felt that one in my backlog. dual AA format support is becoming permanent infrastructure debt, nobody will fund killing it and nobody can ship without it
the ie6 comparison is too accurate. dual AA support is permanent headcount now, not a feature flag
Chiang basically admitted each chain cared more about its own roadmap than interop. Honest at least, but the bill lands on the smallest players in the stack.
smallest player always pays for the divorce. same story since the rollup centralization debates, wallets eat the integration bill both ways
audit bill is only half of it. every user whose tx works on ethereum and fails on base becomes a support ticket your team eats
the support ticket point is underrated. users dont care whether EIP-8141 or frame transactions won, they screenshot the failed tx and tag the wallet
users screenshotting the failed tx and tagging the wallet is exactly how it goes. they do not know frame transactions from eip-8141, they just know send worked on mainnet and broke on base
chiang admitting interop became secondary is the whole story. frame transactions vs 8141 in hegota and the wallets pay for the divorce
maintaining two native AA implementations so wallets have to pick a side is the most 2026 ethereum thing possible lol. EIP-8141 should have shipped years ago
every wallet dev reading this just felt their maintenance backlog double
backlog doubled is optimistic tbh. we also get to explain to users why the same tx looks different depending on the chain. fun quarter ahead
The post-quantum angle of Frame Transactions is being overlooked. That alone justifies pushing 8141 into Hegotá instead of waiting for Base to converge.
finally someone bringing up the post quantum angle. frame transactions without those paths means redoing the entire AA layer again in three years
everyone mad but eip-8141 bundling native AA with post quantum paths in the hegota upgrade is the more ambitious design. base just wanted to ship faster
framing it as base shipping faster ignores that shortcut specs are exactly why interop talks kept stalling. 8141 bundling AA with post quantum paths was always the heavier lift
bundling post quantum paths into an AA upgrade is also why the 8141 timeline slipped. base shipping a simpler format first aint the sin people make it