What growthepie.com Shows About Ethereum L2s That Almost Nobody Is Reading
Most people track price. The real alpha is in the data nobody bothers to pull.
growthepie.com is an open analytics platform tracking every Ethereum L2 across active addresses, throughput, transaction costs, TVL, stablecoin supply, profit, and sequencer revenue. It is one of the most data-rich sources in crypto and one of the least cited in mainstream discourse.
I went through it chain by chain. Here is what I found;
The Dominant Optimistic Rollups — And Why TVL Is Lying to You
Three chains control roughly 90% of all L2 transaction volume:
#Base ,
#ARBİTRUM One, and
#OP Mainnet. But the metrics most people track TVL rankings are telling the wrong story.
Base holds $12.8B in TVL, but that is not its most important number. Its most important number is 87%. That is its onchain profit margin. In 2025, Base generated $80M in onchain profit more than three times what Arbitrum earned on roughly $16.9B in TVL. In February 2026, Base processed $164 billion in a single day of stablecoin transactions, surpassing Ethereum mainnet by more than 3×. Its average transaction cost sits at roughly $0.02 — the lowest of any major L2. Daily active addresses hit a record 3.6 million. Base has no token. Every dollar of sequencer profit flows to Coinbase. The play here is the ecosystem: Aerodrome, Morpho, Aave-on-Base.
Arbitrum One leads TVL at $16.9B, about 40–44% of the entire L2 market. But its 2025 onchain profit was $25M. Base earned $80M. That divergence is the signal. Arbitrum's real moat is not its TVL headline; it is its $4.2B in stablecoin reserves, the deepest stablecoin liquidity pool on any L2 anywhere. GMX, Aave, Uniswap, Camelot, and Curve sit on top of that liquidity. The ARB token currently trades at a P/S ratio of approximately 149×, which suggests significant overvaluation relative to current revenue. Watch fee recovery before sizing in.
OP Mainnet looks weak on its own dashboard: $1.9B standalone TVL, $3.8M in 2025 profit. But this is where most analysts stop and where the real insight begins. Optimism's Superchain strategy means Base, Zora, Mode, and World Chain all built on the OP Stack, pay sequencer rent back into the Optimism treasury. The economic footprint of OP is 5–10× what its standalone dashboard shows. In April 2026, ether.fi migrated $220M in TVL to OP Mainnet with zero downtime across 70,000 active payment cards and 300,000 user accounts. The Superchain flywheel is invisible unless you look at the aggregated revenue flow. The OP token captures it.
The ZK Challenger Tier — Four Chains, Four Very Different Risk Profiles
ZK rollups hold roughly 20% of total L2 TVL. That share has been stable, but the dynamics within it are not.
zkSync Era is the largest ZK rollup by raw TVL, though figures vary significantly by source methodology ($400M to $4.1B depending on whether you count bridge TVL or protocol TVL). Its per-transaction cost is approximately $0.07, and it earned $23M in revenue in the first half of 2024. The structural challenge is Type 4 EVM compatibility, developers have to rewrite contracts, which slows migration from Ethereum. Ongoing token unlock tranches have created persistent sell pressure. The real catalyst to watch is ZK Stack: Matter Labs is enabling third parties to launch their own app-chains using zkSync's proving infrastructure, which could drive a utilisation surge that the token has not yet priced in.
Starknet has the most differentiated narrative in the entire L2 space right now. The strkBTC initiative enabling Bitcoin staking and DeFi on Starknet via a ZK-verified bridge launched in May 2026 and triggered a 50% price surge. Over 1,700
$BTC have already been staked. The Nightfall privacy integration (EY's protocol) and the STRK20 native privacy standard make Starknet the only rollup actively pursuing quantum-resistant, institutional private transactions. This is genuinely distinct. The problem is tokenomics: 38.21% of supply is allocated to early contributors and investors, with monthly unlocks running until March 2027. A 127M
$STRK unlock hit in May 2026 alone. Strong narrative, persistent headwind. You need to hold the thesis through monthly sell pressure.
Linea is the most underreported chain in the ZK tier. Built by Consensys the company behind MetaMask it earned $36.6M in revenue in 2024. That is more than OP Mainnet earned. Almost nobody was talking about it. As a Type 2 zkEVM, it is the most Solidity-compatible ZK chain, meaning existing Ethereum contracts migrate with minimal changes. The distribution moat is structural: Linea is one click away for every MetaMask user over 30 million wallet holders. That default integration is worth more than most people's TVL analysis gives it credit for.
Scroll_ZKP has a more complicated story than it appears. The $SCR token launched in October 2024 via @binance launchpad, two airdrop seasons have already been distributed. Only 19% of the total 1 billion supply is currently circulating, with the next major cliff unlock hitting in October 2026 and the full vesting schedule running to 2028. That means 81% of supply is still locked, significant dilution risk ahead. The revenue picture is also messier than the $35M 2024 figure suggests: by late 2025, on-chain data via @growthepie_eth showed Scroll running negative net revenue fees collected were not covering the costs being paid to Ethereum. Add a governance controversy in April 2026, where the Foundation proposed replacing its security council with a team-controlled multisig, and this is a chain trading primarily on its ZK technology credibility rather than economic fundamentals. The tech is genuinely strong, Type 2 SNARK, audited by Trail of Bits and OpenZeppelin, but the tokenomics and governance trajectory deserve scrutiny before sizing a position.
The Hidden Signals — What growthepie Shows That Price Charts Never Will
This is the batch that matters most. These are the signals sitting in the data that the majority of crypto Twitter is not reading.
Base's profit margin is more important than its TVL. An 87% onchain profit margin at $0.02 per transaction is structurally superior to any L1 on earth. Ethereum mainnet operates at far lower margins with far higher costs. The fact that Base produces this margin while having the lowest fees in the space is not a coincidence it is the outcome of EIP-4844 dramatically reducing data posting costs while Base's Coinbase-driven volume keeps revenue high.
Arbitrum's TVL lead is a liquidity moat, not a growth signal. The $4.2B in stablecoin reserves on Arbitrum is the real barrier to entry for competitors. Liquidity begets liquidity. But the P/S of 149× on ARB means the token is pricing in growth that has not materialised in revenue terms. The divergence between TVL rank and profit rank is the core tension in the ARB thesis.
OP's Superchain revenue is an accounting trick that works in your favour. The standalone OP dashboard makes Optimism look like it is declining. The Superchain view makes it look like a protocol tax on every Base transaction. Both are true. The question is which lens you apply when valuing OP.
Scroll's asymmetry is in the unlock calendar. With 81% of $SCR supply still locked and the next major cliff not until October 2026, near-term sell pressure is structurally lower than peers but the dilution runway to 2028 is real. The tech credibility is there; the governance and revenue trajectory need watching before it becomes a clean thesis.
Linea's MetaMask distribution is a moat nobody is pricing. 30 million wallet users with Linea as a native option is not a small thing. It is the same distribution advantage Coinbase gave Base — except it has not translated into TVL yet. When it does, the move will have already happened.
Polygon PoS is the silent giant. growthepie data shows 82 million organic transactions in 2024 surpassing Ethereum, Arbitrum, and Base. 4.5 million monthly
#stablecoin users, more than Ethereum mainnet, Base, and Solana combined. Polygon dominates small-transaction use cases and developing-market stablecoin flows. This is visible on growthepie. It does not show up in the TVL narratives that dominate X.
TVL Market Share vs Profit Efficiency — The Table That Reframes Everything
When you put TVL share and profit side by side, the story inverts:
TVL variation reflects methodology differences between DeFiLlama, L2Beat, and growthepie (bridge TVL vs protocol TVL vs total value secured).
The takeaway is simple: Base is producing almost $60M more in annual profit than Arbitrum on $4B less in TVL. If Base had a token, this spread would be the loudest conversation in DeFi.
The Rent Paid to Ethereum — The Metric Nobody Talks About
growthepie tracks a metric called "rent paid" the amount each L2 pays to Ethereum mainnet for data availability and security.
In 2025, all L2 networks combined paid approximately $10 million in rent to Ethereum. Their combined revenue was $129 million. They kept $119 million as profit.
This is the direct consequence of EIP-4844 (the Dencun upgrade, March 2024), which cut L2 data posting costs by 80–90% by introducing blob transactions. The upgrade was celebrated as a user fee reduction. It was also the moment L2s became extraordinarily profitable businesses.
Ethereum, meanwhile, transitioned into an inflationary state. It sacrificed over $100 million in guaranteed fee revenue to subsidise L2 growth. The bet is long-term: if L2 activity scales sufficiently, demand for blob space will rise and fees will return to Ethereum. That has not happened yet at scale.
The rent-paid-to-revenue ratio is the cleanest measure of how efficiently a chain is scaling. Lower ratio means the chain is keeping more of what it earns. Base's ratio is the best in the ecosystem.
Bull Run Readiness — How Each Chain Scores on What Actually Matters
Scoring across six factors: TVL depth, profit margin, stablecoin presence, token unlock risk (inverted), narrative strength, and ecosystem moat.
Base — 9.1/10. Highest profit efficiency. Lowest fees. Coinbase distribution. 87% margin. $164B daily stablecoin volume. No token means zero unlock pressure. The bull run play here is ecosystem tokens, not a native chain token.
Arbitrum — 7.8/10. Deepest DeFi stack in L2. $4.2B stablecoin moat. But ARB token is pricing growth that revenue does not yet justify. High-beta when altseason arrives. Watch the P/S compression trade.
OP Mainnet — 7.4/10. Superchain revenue is systematically undervalued by standalone metrics. OP token captures protocol fees across every child chain. Structural undervaluation if the Superchain flywheel accelerates.
Scroll — 7.1/10. Pre-TGE with $35M in demonstrated revenue. The cleanest asymmetric setup in ZK. Zero unlock pressure until TGE. TGE timing is the single catalyst to monitor.
Linea — 6.8/10. $36.6M 2024 revenue. MetaMask distribution moat. Type 2 compatibility. Most underreported ZK chain relative to its real competitive position. No token yet either.
Starknet — 5.9/10. Best narrative differentiation (BTCFi, ZK-verified bridge, quantum-resistant privacy). Monthly token unlocks through March 2027 are the ceiling on near-term price. Strong conviction thesis for patient holders past the unlock schedule.
zkSync Era — 5.4/10. Solid technology, real activity, but Type 4 migration friction and ongoing token unlocks have suppressed both developer migration and token price. ZK Stack app-chain launches are the event to watch.
The One Framework That Changes How You Read L2 Data
Most people open a price chart. A smaller number check TVL on DeFiLlama. Almost nobody pulls up growthepie and reads profit margins, rent paid, stablecoin supply growth, and daily active address trends simultaneously.
When you do, three things become clear:
Profit margin beats TVL as a quality signal. Base proved this in 2025.Token unlock schedules are more important than narrative. Starknet has the best story and the worst unlock calendar. Scroll has no story yet and no unlock pressure. The trade is obvious.Distribution moats are invisible until they are not. MetaMask built Linea's moat quietly. Coinbase built Base's moat the same way. By the time the TVL reflects it, the move has already happened.
The data is public. growthepie publishes all of it for free.
Most people just are not reading it.
Data sourced from growthepie.com, cross-referenced with DeFiLlama, L2Beat, and on-chain reports. May 2026. Not financial advice. Always do your own research.