Ethereum and Base Split Over EIP-8141 and EIP-8130. Wallets Pay the Bill.

Why can't Ethereum and Base agree on what a wallet should even look like? That's the question a lot of builders were asking on Sept. 14, when Ethlabs researcher Derek Chiang confirmed that talks between the Ethereum team pushing EIP-8141 and the Base team pushing EIP-8130 had broken down. Not slowed. Broken.

Both sides wanted the same headline features. Gas sponsorship, passkeys, flexible authentication, programmable accounts. Nobody walked into that room arguing about the destination. They argued about the road, and the road turned out to be the whole ballgame.

The Two Proposals, Side by Side

EIP-8141 is Ethereum's answer. It introduces Frame Transactions, which slice a transaction into programmable calls that handle validation, execution, and gas payment separately. The goal is to detach accounts from the elliptic-curve keys that still define nearly every address on Ethereum today, open the door to key rotation, and lay a path toward post-quantum authentication. In plain terms, an account becomes an address whose behavior is defined by code.

EIP-8130 goes the other direction. It was written by Coinbase engineer Chris Hunter, and it requires every transaction to declare its authenticator so nodes can size up the validation work before executing arbitrary wallet code. The current draft pitches that structure as a way to make validation predictable and let nodes reject unknown authenticators before execution. Ethereum's proposal trusts accounts. Base's proposal trusts the chain to know what it's getting into.

That tension shows up cleanly in the draft's own adoption profiles. The L1 profile permits authenticators outside a canonical set within defined limits. The L2 profile lets high-throughput chains restrict the native transaction path to approved canonical authenticators. Same document, two different philosophies of who gets to decide.

Here's what matters: this didn't fall apart overnight. In late August, developers were still holding dedicated discussions around Frames and borrowing ideas from EIP-8130, even after Ethereum core developers gave EIP-8141 a strong signal toward inclusion in the Hegotá upgrade. Weeks later, the room emptied.

Chiang's own summary was blunt. "Ethereum wanted to be the best version of Ethereum, and Base wanted to be the best version of Base."

The L2 Bargain Just Got Renegotiated

For years, Ethereum's pitch to holders has been simple. Rollups scale the base layer, consume its data availability, settle to it, and keep apps inside the family. The L1 provides security and settlement. The L2s provide throughput. Everyone compounds together.

That story held while L2s were basically execution extensions of mainnet. It gets shakier once an L2 has its own users, its own commercial priorities, its own ship cadence, and now its own native transaction architecture. Base is Coinbase's property. Coinbase has shareholders, a product roadmap, and a sequencer that collects fees. None of that's a secret, and none of it aligns perfectly with what mainnet wants.

Crypto lawyer Gabriel Shapiro put it more sharply than most. "L2s are great, for the crypto industry and for people who own the sequencer. For Ethereum, they're just kinda like 'less bad' than competing L1s."

That's the uncomfortable part, and frankly it's been true for longer than anyone wanted to admit. Sequencing economics accrue to the rollup. Product decisions accrue to the rollup's team. Ethereum gets settlement demand and data availability fees, which is real revenue, but it isn't the same as owning the user relationship. Shapiro's point is that Ethereum may need to stop leaning on the reflected glow of Base and Robinhood and start arguing for what the base layer uniquely does.

Credit where it's due, though. EIP-8130 isn't a secession document. It still targets cross-chain portability through a common authenticator set, and it keeps ERC-4337 available as an alternative transport on networks that don't support the new transaction type. Base isn't designing for isolation. It's designing for a different default.

What the Builders Actually Say

Ethereum core developer Matt Garnett argued that divergence among L2s was always going to happen because market competition forces rollups to ship faster than the L1 can move. "Market pressure forces them to ship features at a pace that L1 can't match, so incompatibilities accumulate. Time will tell whether that's a strength or weakness."

I think that framing is a little generous. Incompatibilities don't accumulate in a neutral vacuum. They accumulate against the thing with the slower release cycle.

From a risk perspective, the bill lands on wallets. Every wallet team now faces a choice: build two transaction paths, or pick one and deprioritize the other. Those teams are small. They ship where the users are. If Base's structured approach produces cleaner high-throughput UX, you can guess which path gets the polish and which one gets the maintenance ticket.

And the philosophical split isn't cosmetic. Chiang framed mainnet's priority as CROPS, meaning censorship and capture resistance, open-source software, privacy, and security. Those values push toward an account model developers can extend without asking permission and a transaction design that can carry privacy systems and future post-quantum signatures. High-throughput chains need something else entirely. They need account systems that scale while staying legible enough for the chain to know which authenticators and transactions it will accept.

Both requirements are legitimate. That's what makes this hard. Nobody here's being unreasonable. They're being differently reasonable.

What to Watch From Here

Three concrete things.

First, watch the Hegotá upgrade. EIP-8141 had momentum toward inclusion, and if Frame Transactions hold their slot in the roadmap, Ethereum mainnet is locking in its direction regardless of what Base does. If the proposal slips, that's a signal the core devs want more room to reconcile.

Second, watch ERC-4337. If it quietly becomes the default interop layer instead of a fallback, account abstraction moves out of the protocol and into middleware. Middleware is where standards go to fragment politely.

Third, watch which path Coinbase's own wallet products optimize for first. That's the fastest read on where the users actually end up, and it's a signal no governance forum will give you.

Then there's the bigger watch item. Do other L2s pick a side, or do they write their own? Monday's split is a permission slip either way. If three or four more rollups ship bespoke account models over the next year, the interoperability problem stops being a wallet inconvenience and starts being the defining constraint on Ethereum's user experience.

The reality is that the EIP numbers aren't the story. The story is how many wallet teams end up building the same feature twice. If the answer is few, the L1 keeps its gravity and this becomes a footnote. If the answer is most, Ethereum's base layer settles into being a settlement layer that wallets route around, and the L2 bargain gets renegotiated again, this time with the L1 holding a much weaker hand.