Share
The Ethereum L2 narrative is one of the few stories in crypto where the headline claims are largely true. In this case, the claims are about mass adoption, dramatically lower fees, and billions in locked value.
But first, let’s look at the background details. Layer 2 consolidation accelerated sharply in 2025. Base emerged as the undisputed leader in TVL, daily users, and overall activity, while most newer L2s saw usage collapse once their incentive programs ended. The ecosystem is clearly working.
Yet simply declaring that “L2s are winning” misses the more important question: what exactly are they winning, and why does that distinction matter for anyone trying to understand where real value is being created on Ethereum?
In 2025, Ethereum solidified itself as the secure foundation for our growing digital civilization. From industry-leading adoption to new technology that reinforces protocol resilience, here are 12 themes that defined the past year:
1/ DeFi reinforced Ethereum's role as the…
— Ethereum (@ethereum) January 6, 2026
The truth is that Ethereum’s major Layer 2s are not competing in the same race. They are operating in different markets, with different capital structures, different user bases, and very different levels of genuine organic activity hidden beneath the headline TVL numbers.
As things stand, Base leads all Ethereum L2s with $4.2 billion in TVL as of early 2026, commanding roughly 41% of the total L2 market share. Arbitrum follows with approximately $1.2 billion, while Optimism holds around $358 million on its main chain (its wider OP Stack ecosystem is larger). zkSync Era trails far behind at under $22 million.
These headline numbers are real, but they are inadequate as a measure of genuine adoption. Here’s why.
TVL is a stock metric. It simply tells you how much capital is sitting in a protocol at a given moment. It does not reveal whether that capital is actively generating fees, serving real users, or merely parked there waiting for token incentives to run out. This was why Sergej Kunz argued for an entirely different metric called the Total Value Unlocked (TVU) in his keynote address delivered at EthCC Cannes.
𝐋𝐢𝐪𝐮𝐢𝐝𝐢𝐭𝐲 𝐢𝐬 𝐧𝐨𝐭 𝐓𝐕𝐋 — 𝐢𝐭 𝐢𝐬 𝐡𝐨𝐰 𝐜𝐚𝐩𝐢𝐭𝐚𝐥 𝐚𝐜𝐭𝐮𝐚𝐥𝐥𝐲 𝐦𝐨𝐯𝐞𝐬
Most people look at DeFi and stop at one metric:
👉 𝐓𝐕𝐋 (𝐓𝐨𝐭𝐚𝐥 𝐕𝐚𝐥𝐮𝐞 𝐋𝐨𝐜𝐤𝐞𝐝)
But TVL is misleading.
Because it describes what is sitting in a system, not… pic.twitter.com/kRYUsodFAF
— GLITCH (@Rukkssss__) April 15, 2026
A far more useful metric is the ratio of fee revenue to TVL, a rough proxy for capital efficiency. For instance, a protocol with $1 billion in TVL that generates $50 million in annual fees is creating more genuine value than one with $5 billion in TVL that only produces $10 million.
Arbitrum, for example, saw fees grow 13% in February 2025 while TVL grew only 1%, suggesting its activity is becoming more efficient and less dependent on constant new capital inflows.
Blast offers the clearest cautionary tale of what happens when capital efficiency is essentially zero. Its TVL collapsed by 97% from $2.2 billion in June 2024 to roughly $55 million by December 2025. This followed a disappointing airdrop, founder silence, and a mass user exodus to Base and Arbitrum.
The capital was never really there for DeFi. It was there for the airdrop. Once the incentives faded and the airdrop resolved unfavorably, the liquidity vanished almost overnight, turning the chain into a virtual ghost town.
Arbitrum makes the most defensible claim to DeFi dominance among all Ethereum L2s.
Its TVL mix is notably balanced: derivatives account for approximately 30% and decentralized exchanges roughly 22%. This diversification means Arbitrum’s fee revenue is not overly dependent on any single use case rising or falling.
GMX stands out as the clearest example of genuine capital efficiency on the chain. With over $244.58 million in TVL, GMX represents nearly 21% of Arbitrum’s total locked value. The protocol eliminates slippage through synthetic token pricing and distributes 30% of protocol fees directly to GMX stakers. This is real yield generated from actual trading activity, not artificial token emissions.
Aave’s lending market on Arbitrum follows a similar pattern. Aave v3 commands roughly $612.37 million in lending TVL, accounting for roughly three-quarters of all lending activity on the chain. Its risk parameters effectively set the baseline borrow rates for much of the Arbitrum ecosystem.
As a result, a chain reaction is formed where capital tends to stick around. Arbitrum hit a record 5.7 million daily transactions in February 2026.
Arbitrum’s main vulnerability lies in developer momentum. Optimism currently leads with 3,044 active developers and 172,954 commits, while Arbitrum has 2,374 developers and 189,957 commits.
Arbitrum holds a slight edge in raw commit volume but trails in active developer count. For a chain of its size, this suggests the ecosystem is deepening rather than broadening. Existing protocols are growing, but fewer new categories and applications are emerging.
Base stands out as the most interesting competitive story in the Ethereum L2 space, not because of superior technical architecture, but because of distribution.
Throughout 2025, Base consistently captured around half of all DEX volume among Layer 2s. Much of that dominance stems from Coinbase’s mainstream user funnel, combined with a growing mix of consumer-facing applications and real usage from protocols like Aerodrome, Echo, and Morpho. Morpho, for example, grew from $354 million in deposits in January 2025 to more than $2 billion by December 2025.
Applications on Base generated a combined $369.9 million in revenue in 2025. Aerodrome alone contributed $160.5 million, which is 43% of the total, while the AI agent launch platform Virtuals added $43.2 million and the sports prediction app Football.Fun contributed $4.7 million.
— AJC (@AvgJoesCrypto) December 23, 2025
The ecosystem shows real revenue diversification. Base is not a one-protocol chain. However, Aerodrome’s heavy contribution remains an honest caveat. A single DEX generating 43% of all ecosystem revenue means Base’s fee profile is still heavily dependent on one application continuing to perform well. If Aerodrome loses meaningful market share to competitors like Hyperliquid or cross-chain alternatives, the ecosystem revenue picture could shift significantly. This is because Aerodrome currently drives the majority of DEX volume and fee generation on Base, making it a key pillar of the chain’s DeFi revenue distribution.
A quieter but more structural issue is the trend in retail activity. Daily filtered user numbers on Uniswap and Aerodrome dropped 74% and 49% respectively, even as on-Base DEX trading volume reached an all-time high. Activity is becoming increasingly concentrated among traders with larger capital and higher trading volumes.
This creates a quiet irony at the heart of Base’s success story. The chain built explicitly for mainstream consumers is seeing its retail participation decline while institutional-sized positions grow. That change is not necessarily negative. It signals improving capital efficiency. But it does suggest that the consumer adoption narrative is currently running ahead of the actual consumer adoption data.
Optimism’s position is structurally different from both Arbitrum and Base. Instead of competing directly for DeFi TVL, Optimism has focused on becoming the infrastructure layer that other L2s build upon.
Optimism’s main chain holds around $4.8 billion in TVL, but the full OP Stack ecosystem, which includes Base, Worldchain, and others, is significantly larger. Optimism recently announced plans to shift to its own integrated tech foundation, effectively turning many of its competitors into customers.
The downside is that Optimism’s standalone metrics look relatively weak next to the broader ecosystem, with Base consistently outperforming Optimism mainnet on most activity measures. This raises a key question: Do OP Stack sequencer fees and governance rights deliver meaningful value to OP token holders, or does most of the value accrue to the individual chains instead?
RetroPGF is an economic flywheel that supports positive impact to the Optimism Collective.
It rewards the individuals and projects that build and create infrastructure, tooling, and content that enable the ecosystem to thrive.https://t.co/y8rYXqFjIg
— Optimism (@Optimism) January 11, 2024
Optimism’s answer is its RetroPGF public goods funding mechanism. Whether RetroPGF creates sustainable developer incentives or simply acts as a permanent subsidy masking weak underlying economics remains an open debate within the ecosystem.
zkSync Era occupies a distinctly different niche. Its TVL remains modest at under $22 million, but its technical differentiation is genuine. The zero-knowledge proof architecture delivers near-instant finality and privacy-preserving features that optimistic rollups cannot match.
This has made zkSync the one of the go-to institutional-grade L2 for enterprises needing compliance-driven privacy. Deutsche Bank’s asset tokenization project on zkSync is proof of this positioning, with the ZK-proof system enabling selective disclosure of transaction details.
However, strong institutional architecture has not yet translated into broad retail adoption. zkSync currently supports 291+ dApps, roughly 162,000 daily transactions, and just over $22 million in DeFi TVL. The capital efficiency gap is clear. $22 million in DeFi TVL on an architecture built for enterprise-scale asset tokenization.
zkSync’s path is clearly institutional-first, retail-second. The protocol allocated $60 million in ZK tokens through its Ignite incentive program to bootstrap DEX activity on Uniswap and SyncSwap. This is the same incentive playbook that failed for Blast, now applied to a chain with more defensible long-term architecture. The central question for zkSync in 2026 is whether it can convert incentive-driven liquidity into sticky institutional usage before the token budget runs out.
Evaluating the four major Ethereum L2s by sector performance rather than raw aggregate TVL produces a much clearer picture.
In DeFi, Arbitrum is the clear winner on depth, breadth, and capital efficiency. Its superior fee-to-TVL ratio, strong derivatives market share, and diverse lending ecosystem are unmatched among L2s. Arbitrum consistently generates more genuine yield from more genuine users than any competitor.
base is growing.
2M+ daily transactions, up 340% YoY
100K+ active users, up 10% YoY
if you’re not on base yet, you’re on the sidelines 🟦 pic.twitter.com/dtQhyYL7sh
— bro.base.eth (@basebro_) April 18, 2026
In consumer and retail applications, Base dominates through distribution. Coinbase’s 100-million-plus user base gives Base a mainstream funnel that no other L2 can replicate. The risk, however, is that the consumer adoption narrative is currently running ahead of the actual consumer data, retail user numbers have declined even as aggregate trading volumes have risen.
In institutional infrastructure, zkSync holds the most defensible long-term position. Its zero-knowledge proof architecture enables privacy-preserving features and selective disclosure that optimistic rollups cannot match. Enterprise partnerships, such as Deutsche Bank’s asset tokenization project, pinpoints use cases that are difficult for other L2s to serve. The adoption curve is slower, but the addressable market is significantly larger.
In ecosystem strategy, Optimism wins by refusing to play the pure TVL game. The OP Stack has become the default rollup framework, powering Base, Worldchain, and a growing number of appchains. This structural moat is systematically undercounted by aggregate TVL metrics, as much of the value flows through the broader ecosystem rather than Optimism’s mainnet alone.
The gaming L2 narrative deserves its own evaluation because it is the most frequently cited growth sector and the one where the gap between promise and data is widest.
Arbitrum Nova was built specifically for lower-cost gaming and social activity, while Arbitrum Orbit lets developers build custom chains for specific apps and games.
The Arbitrum Orbit Expansion Program opens the doors for anyone to deploy an Orbit chain that settles to any Ethereum chain while enjoying future upgrades to the Arbitrum technology stack such as:
🔵 BOLD’s next-gen fraud proofs
🖊 EVM+ compatibility via Stylus
— Arbitrum (@arbitrum) January 25, 2024
The Treasure ecosystem of gaming protocols built on Nova represents the most concentrated gaming DeFi on any Ethereum L2. But Treasure’s TVL and daily user counts remain modest relative to the ambition of the narrative.
Immutable is the clearest gaming-specific thesis in the L2 space. Its zkEVM runs in validium mode with off-chain data availability, optimizing specifically for high-frequency in-game transactions without the settlement costs that would make micro-transactions uneconomical. Its roadmap includes Passport dashboard enhancements for user onboarding and zkEVM prover integration to progressively open the chain more broadly.
The pipeline of game titles being built on Immutable’s infrastructure is real. The question is whether blockchain gaming broadly finds product-market fit before the institutional runway for these projects runs out.
Ronin, built for Axie Infinity and now expanding to third-party developers, has arguably done more for actual gaming user retention than any Ethereum L2, but it is not an Ethereum L2 in the conventional sense. Games originally on Arbitrum have migrated to Ronin specifically to benefit from its lower fees and gaming-first ecosystem.
That is not a sign of Arbitrum winning the gaming sector. It is a sign that generic-purpose L2s may not be the right architecture for gaming applications regardless of their DeFi performance.
Translating the sector analysis into a simple capital efficiency framework produces the most useful comparative picture. Here’s an attempt:
Arbitrum generates roughly $150 million in annual app fee revenue against $20 billion in overall TVL, a ratio of approximately 0.9% annually. Base generated $369.9 million in ecosystem app revenue in 2025 against a peak TVL of approximately $5.6 billion, a ratio closer to 6.6%.
These ratios reveal something important. Base’s capital efficiency is roughly seven times higher than Arbitrum’s on a revenue-to-TVL basis. This is partially because Base’s TVL is lower, making the denominator smaller, but it also reflects genuine activity density. Base is doing more economic work per dollar of locked capital than any other major L2.
The counterargument, well established in DeFi analysis, is that TVL in DeFi serves a different function than revenue-generating capital in traditional finance. Deep liquidity on Arbitrum’s lending and derivatives markets functions as infrastructure that makes other activity possible. A lending market with $1 billion in deposits generating $30 million in annual fees is not inefficient by any sense of the word. Rather, it’s providing the depth of liquidity that makes $500 million in leveraged trading positions safe to execute.
Both frameworks are valid and neither alone is sufficient. The right analytical approach is to hold both simultaneously: Base wins on capital efficiency and consumer activity density, while Arbitrum wins on capital depth and DeFi infrastructure quality. They are not competing for the same users or capital in any meaningful sense.
The L2 space is contracting into two categories. First, there are commodity rollups competing on fees and throughput. Second are the specialised L2s with fundamentally different execution models. Asides from those two, everything in between, generic EVM rollups with no distribution, no unique features and no reason to exist beyond being a Layer 2, faces extinction.
That consolidation creates clarity for anyone trying to evaluate the sector. The chains worth watching in 2026 are the ones that can answer a simple question: what do you enable that your competitors cannot? Arbitrum’s answer is DeFi depth. Base’s answer is Coinbase distribution. zkSync’s answer is institutional privacy. Immutable’s answer is gaming-specific architecture.
Everything else is an incentive program waiting to expire.
21Shares has predicted that most L2s will not survive past 2026. The ecosystem will consolidate around ETH-aligned designs, high-performance entrants and exchange-backed networks. L2s without strong distribution or sustainable economics face zombie chain status or shutdown.
The consolidation is already happening. The question for investors is not whether the L2 sector is growing, it is. The question is which chains are growing because users need them and which are growing because their incentive programs have not yet expired. On that measure, Arbitrum’s fee revenue growth, Base’s Coinbase distribution advantage and zkSync’s institutional architecture are the three most defensible positions in the space. Everything else requires a closer look at the incentive calendar before drawing conclusions about genuine adoption.
Share
preferred on
