We don't talk enough about the moment a narrative dies. Not because it fails — but because it succeeds. For the past four years, Ethereum has been the story of "scaling." Rollups multiplied like rabbits. L2 beat every metric that mattered. Transaction costs dropped from $150 to pocket change. The word "blobs" entered our vocabulary. And then, somewhere along the way, the question changed.
It's no longer "Can Ethereum scale?" It's "What was all that scaling for?"
A title crossed my desk last week: "From Glamsterdam to Hegotá." Two invented cities — the first a shimmering mix of Glamour and Amsterdam, the second a dialectical fusion of Hegel and Bogotá. At first read, they feel like playful placeholders. But they map something real. Amsterdam was the old Ethereum — Devcon after Devcon, a European core of builders, protocols, and ideals. Bogotá is the new frontier — retail users in emerging markets, real usage under real constraints, a decentralization that isn't just architectural but geographic.
The shift in those two syllables tells us more about Ethereum's next phase than any roadmap. Scaling was the thesis. What comes after is the antithesis — and the synthesis hasn't been written yet.
Context: The Rollup-Centric Roadmap Is No Longer a Roadmap — It's a Reality
Let's ground this in what's already shipped. EIP-4844 went live in March 2024, giving rollups dedicated blob space. That single change cut L2 data publication costs by over 90%. Arbitrum, Optimism, Base, zkSync — they all hit production. TVL across L2s now routinely exceeds the L1's DeFi TVL. Active addresses on L2s are higher than on Ethereum mainnet.
Scaling happened. I can say that without a single asterisk because I've watched it from inside the ecosystem since 2017 — back when the hottest DeFi project was a vampire attack on Uniswap, and people genuinely thought the DAO hack defined Ethereum forever. The bear market didn't end curiosity; it sharpened it. During 2022's bloodbath, I spent my weekends simulating ZK proof generation times and writing about recursive SNARKs until my eyes blurred. I know what it feels like to want scaling to be the answer to everything.
But here's the uncomfortable truth: Scaling has made Ethereum more fragmented, not less.
The problem isn't throughput anymore. The problem is that users now have to navigate a multiverse of chains — each with its own bridge, its own token standards, its own security assumptions, its own UX quirks. Ask any newcomer to move from Arbitrum to Base without a bridge aggregator and watch them give up. This isn't a technology failure. It's an ecosystem coordination failure.
Core: The Next Ethereum Isn't a Chain — It's a Layer of Intent and Trust
So what does Ethereum actually need to solve next? Let me walk through where the evidence points.
1. Interoperability Without the Bridges-of-Death
Bridges remain the graveyard of DeFi. Since 2020, over $2.8 billion has been stolen from cross-chain bridges. The root cause isn't sloppy code in every case — it's the fundamental architecture of locking assets on one chain and minting wrapped proxies on another. That model introduces a trusted third party, exactly what we're supposed to be eliminating.
The answer isn't a better bridge. It's a different computational paradigm: intent-centric transactions. Instead of users manually hopping chains and signing multiple transactions, users state what they want — "Swap 10 ETH for USDC on Base" — and specialized solvers compete to execute that intent across the entire ecosystem. The user doesn't care which chain the liquidity lives on; they care that the settlement happens within their slippage tolerance.
ERC-7683 is already laying the groundwork for this. But standards alone don't unify a fractured community. Based sequencing — where L2s share Ethereum's sequencer infrastructure — could native interop, but it faces political pushback from teams that want to keep their sequencing MEV.
Here's my contrarian take, grounded in three years of protocol PM work: the technical solutions are mostly here. The missing piece is economic alignment. L2s profit from fragmentation because each chain captures its own fees, its own MEV, its own token premium. We don't have a technical interoperability problem. We have an incentive incompatibility problem. And that's harder to solve than writing another precompile.
2. User Experience Has Been Treated Like a Villain, But It's the Exit
The bear market didn't kill builders' spirit, but it did expose something uncomfortable: the average user doesn't care about "trustless composability." They care about whether the app loads in under two seconds and whether they can recover their account if they lose their phone.
Account abstraction via ERC-4337 was supposed to fix the private key nightmare. And it works — technically. But adoption has been lukewarm because application developers are stuck in the paradigm of "wallet as afterthought." We built the rails for smart accounts, and then forgot to build the on-ramps.
I realized this most clearly in 2025, when I was building TruthLayer, a decentralized registry for AI-generated media. We had to decide between a custody-style wallet and full account abstraction. The engineers wanted AA — it was cleaner, more aligned with long-term architecture. The user research kept pointing to the same pain point: "I don't want to manage 12 recovery phrases." We eventually built a social recovery layer on top of ERC-4337. It took six extra weeks of work. The retention increase? 40%. Nobody congratulated us for removing private keys. They just said, "Finally, it feels like an app."
That's the next phase of Ethereum: not more nodes, not faster blocks, but making the technology invisible enough that real users stop noticing they're on-chain.
3. Value Capture Is Becoming a God-Tier Problem
L2s use Ethereum for security and data availability. But do they send fees back to ETH holders? Almost nothing. The "ultrasound money" narrative assumed that more L2 activity → more ETH burned → more value accrual. Reality: EIP-4844 left L1 gas fees historically low, and L2s operate with sub-cent fees that mostly stay within the L2 ecosystem.
If ETH is no longer the gas currency that burns, what justifies its multi-thousand-dollar market cap? Staking yields, sure — but 3% isn't a growth narrative. DeFi collateral — but collateral can be a basket of yield-bearing assets. The honest answer is that Ethereum's token economics haven't caught up with its architecture.
Some L2s are experimenting with revenue-sharing back to the base layer. But as an observer who's seen six different "alignment" proposals stall, I can tell you: this is the problem that will define Ethereum's next decade. If L1 becomes a public good that L2s consume for free, we're essentially creating a tragedy-of-the-commons on a global scale.
4. The Global South Is Not a Margin Note — It's the Thesis
This is where "Hegotá" becomes more than a clever portmanteau. In Bogotá — and in Nairobi, where I'm writing this — the problem with Ethereum was never throughput. It's accessibility. It's the cost of bytes (not gas), the cost of smartphones, the cost of stablecoin on-ramps.
The 2024 Bitcoin ETF opened the floodgates of institutional interest, but the next billion users aren't in pension funds. They're in Lagos, Mumbai, Buenos Aires, and Jakarta. They need remittances that don't lose 8% to Western Union, savings accounts that don't evaporate with inflation, and identity systems that don't require a birth certificate.
The beauty of scaling isn't that you can now run a decentralized exchange for 99% cheaper. It's that a farmer in rural Kenya can hold tokenized farmland income through a wallet that auto-creates itself with biometrics, protected by social recovery. That's not a "use case." That's a new human right.
Ethereum's next-phase roadmap should be measured not by TVL or transaction throughput, but by how many people outside the crypto echo chamber actually use something built on it without knowing it's called "Ethereum."
Contrarian: The Real Risk Is That Ethereum Becomes the AOL of Web3
Here's where I stop to annoy the maximalists.
Ethereum won the L1 wars. It's the settlements layer, the security root, the most battle-tested chain in existence. But winning a category doesn't mean you win the future. AOL won the internet dial-up race. IBM won the mainframe. Nokia won feature phones.
What kills dominant platforms isn't a rival that plays the same game better. It's a paradigm shift that makes the old platform's strengths irrelevant.
If Ethereum focuses all its energy on perfecting L1-L2 interoperability while a newer chain offers complete vertical integration — where users never need to know which chain they're on, where the wallet is the chain, where the UX is indistinguishable from Venmo — the technical excellence of Ethereum might become a museum piece.
Solana understood this earlier than we'd like to accept. They took the opposite path: one chain, one state machine, one unified liquidity pool. It's less philosophically elegant than a rollup ecosystem. It's also simpler for 99% of users.
Ethereum's diversification into L2s was necessary — you can't compress the world's computation into a single chain without hitting physical limits. But we've treated diversity as an end in itself, forgetting that diversity only creates value when it's coordinated diversity. The next phase requires ruthless product integration at the top, not just new primitives at the bottom.
So my contrarian conclusion: Ethereum shouldn't be building more infrastructure. It should be building fewer, thicker interfaces that abstract away the entire infra stack. The winning application of the next cycle won't be a DEX or a lending protocol. It'll be a super-app that says, "Stop thinking about chains. Here's your money, your identity, your community."
Takeaway: The Dialectical Synthesis Is Human, Not Technical
Hegel's dialectic doesn't end with a perfect thesis. The synthesis is always messier, more complex, more contested than either side ever imagined.
Glamsterdam's scaling thesis gave us seven L2s, blobs, and sub-cent transactions. Hegotá's antithesis gives us real users, real regulators, and real-world constraints that no testnet can simulate. The synthesis — the "Ethereum that solves what comes next" — won't be built by protocol researchers alone.
It'll be built by people like the trio I hacked next to at a Nairobi discord sprint last month: a frontend dev who'd never written a smart contract, a market-maker who thinks ZK proofs are a band name, and a storyteller who just wants her grandmother's savings to survive another election cycle.
The next Ethereum isn't a product. It's a promise — that we don't just scale the technology, but scale our imagination of what a decentralized society looks like.
We don't get there by adding more sequencers.
We get there by remembering that the point was never the blockchain. It was always about the people who, after the noise fades, are finally free to build what they actually need.
The bear market didn't kill Ethereum's spirit. It killed the excuse that we're still waiting for scalability to arrive. The waiting is over. The building of the next phase is already underway — whether we'll recognize it in time is the only remaining question.
About me: I'm Chris Thompson, a protocol PM living in Nairobi. I audit code as a hobby and write about the humanity between transactions. This is what I see from the edge of the map.
