Linea scored 51 out of 60 on my rubric. Scroll scored 50. Scroll got the build week.
That one-point gap wasn't post-rationalized. The pick came down to two things the rubric doesn't score directly: faucet auth requirements and timing. I'll get to both.
Two chains that didn't make the shortlist
Five candidates went through the blocker check. Two were eliminated. Polygon zkEVM was announced as sunset in June 2025. PancakeSwap pulled support the following month and developer momentum has been draining out since. Building on it now would produce an article about a chain in hospice care. Kakarot was simpler: the GitHub repo was archived January 9, 2025, the team was acquired by Zama, and they pivoted to FHE. No deployable mainnet. Both gone before we talk scores.
The remaining four
zkSync Era made the cut but scored the lowest at 40.5 out of 60. The foundry-zksync fork is still alpha, tests run roughly 17× slower than mainline Foundry, and the native account abstraction model means msg.sender behaves differently in smart wallet contexts. Real build risk for a quick-turnaround week.
Taiko is technically the purest: Type 1 zkEVM, identical EVM, zero code changes, decentralized sequencer through Ethereum L1 validators. It scored 49. TVL sits around $20M, and the Hoodi testnet only replaced Hekla in September 2025, so the wagmi chain definition needs manual configuration. Newer testnet story than I wanted.
That left Linea at 51 and Scroll at 50.
Why Linea almost won
Linea is built by Consensys, the same company behind MetaMask, and it ships pre-configured in MetaMask's default network list. Docs are professionally maintained. TVL is the highest of the qualified candidates. The proof system uses lattice-based cryptography (Vortex/Plonk) instead of elliptic-curve SNARKs, which is a genuinely interesting post-quantum angle. If the criterion were "biggest chain with the most traction," Linea wins.
Why Scroll won
Two things.
First: faucets. W1 on Base went smoothly partly because the testnet faucet only needed a wallet address. I learned to check this before scoring, not after. Linea's cleanest path to Sepolia ETH requires a free Infura account. Not social-gated or ENS-gated, but you still need an account. Scroll's Telegram bot takes a wallet address and sends ETH, nothing else. That difference is small on paper. After last week, I notice it.
Second: timing. In April 2025, Scroll replaced its entire prover stack. The original architecture used hand-written halo2 arithmetic circuits to prove EVM execution, one circuit gadget per opcode, with a hard capacity ceiling per batch. Euclid swapped it out for OpenVM, a RISC-V zkVM built by Axiom in collaboration with Scroll and the Ethereum Foundation's research team. Instead of proving EVM execution directly, the system compiles to RISC-V and proves that. Throughput went up 5×, transaction costs dropped roughly 50%, Scroll hit Stage 1 ZK Rollup status on L2Beat. They also moved from a custom zktrie to Ethereum's standard Merkle-Patricia Trie, so any tooling that handles Ethereum state proofs now handles Scroll state proofs too. Most developers I've talked to don't know this happened. It's April 2026 and it's still not common knowledge.
The TVL question
Ether.fi exited to Optimism in early 2026. They took 300,000 user accounts, 70,000 active cards, and roughly 85% of Scroll's TVL with them, dropping the number from over $1 billion to around $27 million. That's the number. Ether.fi's reason: Optimism has larger TVL and a broader app suite, which matters more than ZK rollup maturity for a payments product. That's a real signal about where liquidity goes. But it's a liquidity routing decision, not a judgment on the developer infrastructure. The tooling is solid. The proving stack is active. The GitHub has ongoing commits across 152 repositories as of April 2026. TVL and developer experience are different charts.
What the pick says
Picking the second-ranked chain means the rubric is a starting point, not a verdict. A 1-point gap is noise. What mattered was whether the friction points I care about actually lined up: faucet auth on day one, a fresh technical story, proven tooling. They did.
Scroll was open-source from the first commit. Not retroactively. A governance vote happened before the biggest upgrade shipped. W1 was the establishment chain. This is the contrast.
→ The live app is at https://proof-of-support.pages.dev
→ Scoring methodology for the series: How I'm Scoring the Chains
Top comments (0)