ETH vs L2 Tokens

ETH and the tokens issued by Layer 2 networks — ARB, OP, STRK — often get lumped together as "Ethereum ecosystem plays." Mechanically, they could hardly be more different. ETH is the asset the whole system runs on; most L2 tokens are governance instruments with no gas role and no claim on fees. This guide walks through what each token actually does, where the value-accrual arguments hold up, and how to evaluate an L2 token without wishful thinking.

What ETH Actually Does

Ether is not a token that happens to live on Ethereum; it is a structural component of the protocol. Its roles compound each other:

1. Gas on L1 — and on most L2s. Every Ethereum transaction pays gas fees in ETH; no other asset is accepted at the protocol level. Crucially, the major rollups made the same choice: Arbitrum, OP Mainnet, Base, Linea, Scroll, and ZKsync Era all denominate gas in ETH. Moving activity to Layer 2s therefore does not route around ETH — it extends ETH's role as the ecosystem's money.
2. Staking collateral. Ethereum's proof-of-stake consensus is secured by validators who each lock 32 ETH (with larger effective balances possible since the Pectra upgrade). By mid-June 2026, nearly 40 million ETH — about 32% of supply — was staked, with over four million ETH added in the first half of 2026 alone. Staked ETH earns issuance and fee rewards, and an entire economy of liquid staking and restaking is built on top of that collateral.
3. Fee burn. Since EIP-1559 activated in August 2021, the base fee of every transaction is burned — permanently destroyed rather than paid to validators. Cumulative burn passed six million ETH by late 2025, per burn-tracker estimates. Burn scales with usage, which directly ties network activity to ETH's supply schedule.
4. Collateral and settlement money. ETH (and its liquid staking derivatives) is the dominant reserve asset across Ethereum DeFi — the collateral behind loans, the base pair of pools, the backing for portions of major stablecoins.

No L2 token replicates this stack. That asymmetry is the starting point for every honest comparison. You can track ETH itself on our Ethereum asset page.

What L2 Tokens Actually Do

If ETH pays the gas on Arbitrum and Optimism, what are ARB and OP for? The short answer for most L2 tokens: governance, treasury allocation, and incentives. The details differ enough that each is worth examining.

ARB: Governance Only, By Design

ARB launched in March 2023 with an initial supply of 10 billion. The Arbitrum Foundation's own documentation describes it as a special-purpose token whose function is voting in the Arbitrum DAO — which governs protocol upgrades, treasury spending, and the technology of Arbitrum One. That governance power is real: the DAO controls one of crypto's largest treasuries and can even mint new ARB, capped at 2% of supply per year. What ARB does not do is pay gas (that is ETH) or entitle holders to sequencer revenue. Team, advisor, and investor allocations vest monthly over four years, from March 2023 to March 2027 — adding on the order of 92 million ARB to the float each month through early 2027, per unlock trackers.

OP: Governance, With a New Revenue Experiment

OP launched with a genesis supply of 4,294,967,296 tokens and a default inflation rate of 2% per year, adjustable by governance. It powers the Optimism Collective's two-house system: the Token House (OP holders) votes alongside a Citizens' House on protocol matters and the distribution of sequencer revenue from OP Mainnet and other Superchain members — historically funding public goods through retroactive funding rounds rather than paying holders.

That changed partially in January 2026: Optimism governance approved a plan to direct 50% of net Superchain sequencer revenue to recurring OP buybacks in a 12-month pilot starting February 2026. It is the first serious attempt by a major L2 to connect token demand to network revenue — though the scale matters: with roughly 5,900 ETH of revenue collected in the preceding year, the buyback flow is measured in single-digit millions of dollars annually at early-2026 prices, small relative to OP's market cap.

STRK: The Utility Exception

Starknet took a different path. STRK (10 billion max supply) can be used to pay Starknet transaction fees alongside ETH — a portion of STRK-denominated fees is converted to ETH to cover the network's L1 data costs — and it is staked to secure the network, with validators required to lock at least 20,000 STRK and delegation available for smaller holders. New STRK is minted as staking rewards, so the token has real protocol utility but also ongoing issuance. STRK is the template skeptics point to when asked what an L2 token could do beyond governance.

LINEA: The Burn Experiment

The newest design comes from Linea, the Consensys-built zkEVM, whose LINEA token launched in September 2025 with a total supply of 72 billion. Its tokenomics break from every predecessor: per Linea's published tokenomics, 85% of supply is allocated to the ecosystem and 15% to a locked Consensys treasury, with no allocations to team members or investors — and, unusually, no conventional token governance either (the network is stewarded by a consortium of Ethereum-aligned organizations). LINEA is not a gas token; gas on Linea is ETH only.

What LINEA does have is a protocol-level dual-burn mechanism: after covering L1 data costs, 20% of the network's net ETH gas profits are burned outright as ETH, and the remaining 80% is used to buy LINEA on the market and burn it. Every transaction on Linea therefore destroys a little ETH and a little LINEA. It is the first major L2 token whose value mechanism is usage-driven destruction rather than governance rights — effectively importing EIP-1559's burn logic to the L2 token layer. The open question is scale: like Optimism's buybacks, the burn is bounded by network profits, which for all L2s are modest in the post-blob fee era.

Two other majors round out the picture: ZKsync's ZK and Scroll's SCR both launched as governance tokens in the ARB/OP mold, with gas on their networks paid in ETH. The design space, then, now spans four archetypes: pure governance (ARB, ZK, SCR), governance plus revenue buybacks (OP), network utility via gas and staking (STRK), and burn-linked economics with no governance (LINEA).

Base: The Network With No Token

Base — as of August 2026 the largest L2 by value secured, per L2Beat — has no network token at all. Gas is paid in ETH, and sequencer revenue accrues to Coinbase. In late 2025, Base's creator Jesse Pollak said the team was exploring a network token, reversing years of "no token" messaging, but nothing has launched and no airdrop exists. Through mid-2026 the exploration remained just that — no confirmed design, timing, or governance — even as prediction markets priced a 2026 launch as more likely than not and bank analysts sketched double-digit-billion valuations for a hypothetical token, per roundups of the speculation. Base's growth is the cleanest natural experiment in this debate: a network can reach the top of the L2 rankings with no token whatsoever — which tells you the token is not what makes an L2 work.

Utility Matrix: ETH vs ARB vs OP vs STRK vs LINEA

What Changed in 2025–2026: The End of "Governance Only"

For the first two years after the big airdrops, "L2 token" was near-synonymous with "governance token." Between late 2025 and mid-2026, that consensus broke, network by network:

• Optimism's buybacks went live — and kept running. The 12-month pilot approved in January 2026 executed its first purchase on March 5 (95.8 ETH for about 1.57 million OP) and had accumulated more than nine million OP by early August 2026, per the program's public reporting. Repurchased tokens sit in the treasury with their future use decided by governance, per Optimism's announcement. At current activity the flow implies roughly $8 million a year — economically small, but the precedent is the point.
• Arbitrum approved ARB staking. The DAO passed a proposal, with about 91% approval, to let holders stake and delegate ARB in exchange for a liquid stARB token, and to direct future surplus sequencer fees to stakers who delegate to active governance participants, per Crypto Briefing's coverage. Meanwhile the DAO's non-fee revenue experiments grew: Timeboost, an auction for priority transaction ordering, generated about $406,000 in Q1 2026 out of roughly $23.5 million gross DAO revenue in 2025, per mid-2026 progress summaries — though DAO revenue still trails its ecosystem spending, a gap delegates debated openly when the Foundation requested fresh funding in 2026, per The Defiant.
• Linea launched burn-linked tokenomics (previous section) — no governance, no gas role, just usage-driven destruction of ETH and LINEA.
• Starknet doubled down on staking utility. Staked STRK grew roughly elevenfold during 2025 to over 1.1 billion — more than 23% of circulating supply — and the network added trustless Bitcoin staking, making STRK 75% and BTC 25% of consensus weight, per Starknet's year-in-review and The Block.

The honest reading of this shift: the mechanisms are real but the magnitudes are not yet. Buybacks and fee-sharing are bounded by L2 profits, and post-blob L2 profits are thin everywhere. What changed is direction — the era when holding an L2 token bought you nothing but votes is visibly ending, and each network is testing a different route out of it.

The Value-Accrual Debate, Both Sides

The core question for any L2 token: if the network wins, does the token have to win with it? Here are the strongest versions of both arguments.

The Bear Case for L2 Tokens

• Usage does not require the token. Millions of users transact on Arbitrum and Base daily without touching ARB or holding any Base token at all. Demand for blockspace becomes demand for ETH, not for the governance asset.
• Fees flow elsewhere. Sequencer revenue net of L1 costs goes to operators and treasuries. Governance tokens historically had no dividend, no burn, no buyback — a vote is not a cash flow.
• Supply keeps growing. Vesting cliffs, monthly unlocks, DAO incentive programs, and inflation all add float, while nothing mechanically removes it.
• Governance value is capped. Voting power over a treasury has some value, but it is diffuse, and most holders never exercise it.

The Bull Case for L2 Tokens

• Policy can change. Tokens are claims on future governance decisions, and governance can vote itself value — Optimism's 2026 buyback pilot is proof the door is open. A future "fee switch" is always one proposal away, subject to legal and competitive constraints.
• Treasuries are real assets. ARB governs a multi-billion-token treasury; controlling how it is deployed is not worthless.
• Utility can be added. Starknet retrofitted gas and staking roles onto STRK. Others can follow.
• ETH's own accrual is weaker post-blobs. The honest counterpoint to "ETH captures L2 growth": since EIP-4844, L2s pay Ethereum comparatively little for data, so the burn contribution per L2 transaction is tiny. ETH's L2-driven value accrual runs more through its money-and-collateral role than through fee extraction.

Neither side is dishonest; they weigh mechanics against optionality. What is dishonest is pricing an L2 token as if it already had ETH's roles.

A Worked Example: Following the Fee Flow

Abstract arguments become concrete if you trace a single transaction. Suppose you swap tokens on Arbitrum and pay the equivalent of $0.04 in gas:

1. You pay in ETH. The fee is denominated and charged in ETH on Arbitrum, so your action created a sliver of ETH demand — and none for ARB.
2. The sequencer collects it. Your fee, along with thousands of others, accrues to the sequencer operated by Offchain Labs on behalf of the network.
3. Part is spent on Ethereum. The rollup periodically posts compressed transaction data to Ethereum in blobs and pays for it in ETH. A fraction of that L1 payment is burned via EIP-1559; since blobs made data cheap, this fraction is small.
4. The margin goes to the network, not the token. What remains after L1 costs is sequencer profit that accrues to the chain's treasury and operator. ARB holders can vote on how DAO funds are deployed — but nothing in this flow bought, burned, or distributed ARB.

Run the same trace on OP Mainnet in 2026 and one step changes: half of the net margin is now earmarked for OP buybacks under the pilot program. Run it on Starknet with fees paid in STRK and the first step changes too. That is the entire ETH-vs-L2-token debate in miniature: the tokens differ in which steps of the fee flow they touch, and most of them touch none.

How the Airdrop Era Shaped These Tokens

Every major L2 token entered the world the same way — as a retroactive airdrop — and that origin still shapes how they trade. Optimism ran the first big experiment with multiple OP airdrop rounds starting in mid-2022, deliberately spreading distribution across repeated snapshots to blunt farming. Arbitrum followed in March 2023 with a single large drop of ARB to historical users and ecosystem DAOs, per the Arbitrum Foundation's documentation. Starknet distributed STRK "provisions" to users, developers, and even Ethereum stakers in February 2024. Linea closed the era in September 2025 with a launch that put roughly 22% of supply in circulation at genesis, most of it user-facing, per its tokenomics announcement.

The pattern across all four launches was consistent enough to be a lesson in itself: an airdrop converts past usage into present sell pressure. Recipients who farmed the snapshot treat the tokens as income, not investment; professional farming operations industrialized this between 2022 and 2024, spinning up thousands of wallets per expected drop. Each successive launch traded into weaker demand as the market learned the pattern — which is part of why later networks experimented with different designs entirely: Linea attached a burn instead of governance, and Base has so far declined to launch a token at all. If you are evaluating a future L2 token launch (a Base token being the obvious candidate), the airdrop-era history is your base rate: initial float sold hard, and the tokens that stabilized were the ones that eventually attached mechanics — staking, buybacks, burns — beyond the vote.

Supply Pressure: Airdrops, Emissions, Unlocks

L2 tokens typically launch via airdrop with a minority of supply circulating, and the rest — team, investors, foundation, ecosystem funds — vesting over years. This structure has predictable consequences:

• Airdrop overhang. Recipients who farmed the airdrop often sell into launch liquidity, front-loading price pressure.
• Scheduled unlocks. ARB's four-year vesting adds roughly 92 million tokens to the float monthly through March 2027; every unlock is a known future supply event that markets partially pre-price.
• Incentive emissions. DAOs spend treasury tokens on liquidity and growth programs, which recipients frequently sell — rented activity that shows up as supply.
• Low float, high FDV. A small circulating supply against a large total supply means the headline market cap understates what the market is really being asked to absorb over time.

None of this makes a token worthless; it means the demand side must outrun a known supply schedule. Liquidity depth determines how gracefully that absorption happens — our guide on token liquidity explains how to gauge it.

The Regulatory Backdrop: Why Fee Switches Got Easier

For years, the standard answer to "why doesn't the token get a fee share?" was as much legal as economic: routing revenue to token holders looks like a dividend, and a dividend looks like a security. The US environment of 2025-2026 loosened that constraint in stages:

• Staking clarity. In March 2026 the SEC and CFTC issued a joint interpretive release classifying protocol staking rewards as non-securities, and the first staking Ethereum ETFs began trading the same month, per industry trackers of the 2026 staking market.
• Market structure, still pending. The CLARITY Act — the bill that would formally divide token oversight between the SEC and CFTC — passed the House in 2025, but as of August 2026 it had still not cleared the Senate; a cloture motion filed on August 8 set up floor votes for September, per CoinDesk.

The practical effect is visible in behavior: Optimism shipped buybacks, Arbitrum voted fee flows toward stakers, and Coinbase — a US-listed public company — openly discusses a Base token, all things that would have been legally unthinkable in 2022. But "friendlier" is not "settled": until market-structure legislation actually passes, every fee switch and buyback carries residual regulatory risk, which is one reason these programs launch as capped pilots rather than permanent commitments.

ETH's Institutional Lane: A Moat No L2 Token Has

One more asymmetry hardened in 2026. ETH now has a full institutional wrapper stack: spot ETFs since 2024, and from March 2026, ETFs that stake their holdings and pass through yield — BlackRock's iShares Staked Ethereum Trust (ETHB) began trading on Nasdaq on March 12, 2026, joining Grayscale's converted staking product, with more issuer amendments queued behind them, per institutional staking guides. That pipeline — retirement accounts and asset managers buying yield-bearing ETH exposure — simply does not exist for ARB, OP, STRK, or LINEA, none of which has a US spot ETF.

It compounds the earlier point about roles: ETH is simultaneously the ecosystem's gas, collateral, staking bond, and now a regulated yield product. An L2 token competes for a narrower pool of buyers — crypto-native holders who specifically want exposure to one network's governance and policy options. When you compare charts, you are not comparing two versions of the same bet; you are comparing assets with entirely different demand bases.

Case Study: Arbitrum's Network-Token Gap

The cleanest illustration of everything above is ARB itself. Through mid-2026, Arbitrum One remained the second-largest L2, securing about $10 billion in value per L2Beat, with deep DeFi liquidity and growing real-world-asset activity. Over the same period ARB traded around $0.08 with a circulating market cap near $0.5 billion, per CoinGecko — a fraction of its 2024 levels, and a market cap roughly one-twentieth of the value the network secures.

Every mechanism in this guide shows up in that gap: fees paid in ETH rather than ARB, sequencer margin accruing to a treasury rather than holders, roughly 92 million tokens unlocking monthly into 2027, DAO spending running ahead of DAO revenue, and staking rewards approved but not yet delivering meaningful fee flow. None of this says Arbitrum-the-network is failing — the network's own numbers say the opposite. It says the token was priced, for years, on an assumption of value accrual that the mechanics never supported. Whether the stARB era changes that is exactly the kind of question the checklist below is built to interrogate.

How to Evaluate an L2 Token Honestly

A practical checklist, in the order that matters:

1. What does holding it actually entitle you to? Gas utility, staking yield, fee share, or only votes? Read the docs, not the ticker's narrative.
2. Where do network fees go? Trace sequencer revenue: operator, treasury, public goods, buybacks? A token can only capture what its governance routes to it.
3. What is the supply schedule? Circulating vs FDV, unlock calendar, inflation rate, and who holds the locked tokens.
4. Is the governance power real? Can token votes change fees or upgrades, or does a foundation/security council hold the keys that matter?
5. Is the network itself winning? Value secured, activity, and stage of decentralization — the fundamentals covered in our L2 guide — since even perfect tokenomics on a dying network is worth little.
6. How does it trade? Market structure — float, liquidity, momentum — often dominates fundamentals over short horizons; see how ETHBubbles scores momentum for how we quantify it.

You can watch ETH, ARB, OP, and STRK move side by side on our Ethereum tokens map and compare the underlying networks on the L2 dashboard.

Common Mistakes When Comparing ETH and L2 Tokens

The same errors recur in every cycle's version of this debate. Four worth naming:

• Treating the token as an index on the chain. "Arbitrum is winning, so buy ARB" assumes a value link the mechanics may not provide. Always ask what specific flow — gas, burn, buyback, staking yield — connects usage to the token, and how large that flow actually is.
• Comparing market caps across different float regimes. A token with 20% circulating and one with 95% circulating at the same market cap are not similarly priced assets; the former has four times its float scheduled to arrive. Compare FDVs, then ask whether either is justified.
• Assuming announced mechanisms are material mechanisms. A buyback pilot measured in single-digit millions against a multi-hundred-million market cap is a signal about governance direction, not meaningful supply absorption. Scale the mechanism against the market cap before crediting it.
• Forgetting that ETH is on both sides of the trade. Every L2 in this guide except Starknet charges gas exclusively in ETH, and even Starknet's STRK fees are partly converted to ETH. Buying an L2 token instead of ETH is not diversifying away from ETH — it is layering a governance bet on top of an ecosystem whose money is still ETH.

Frequently Asked Questions

Related Guides

• Ethereum Layer 2s Explained — the technology and trust assumptions behind the networks these tokens govern.
• Ethereum Gas Fees Explained — why gas (and its burn) is the engine of ETH's value-accrual argument.
• Token Liquidity Explained — how float, depth, and unlocks shape how these tokens actually trade.
• How ETHBubbles Scores Momentum — the methodology behind the momentum signals on our dashboards.

Sources & Further Reading

• Arbitrum Foundation — The ARB Token: A Conceptual Overview
• DefiLlama — Arbitrum Token Unlocks and Vesting Schedule
• Optimism Community Hub — OP Token Overview
• CoinDesk — Optimism Governance Approves OP Buyback Plan (Jan 2026)
• Starknet Docs — STRK
• CoinDesk (via Yahoo Finance) — Coinbase Considering a Token for Base
• L2Beat — Scaling Summary (value secured and stage ratings)
• ultrasound.money — ETH Supply and Burn Dashboard
• Linea — Tokenomics (official announcement)
• Linea Docs — Tokenomics and the Dual-Burn Mechanism
• Optimism — OP Token Buybacks (official blog)
• crypto.news — Optimism Buyback Program Progress (2026)
• Crypto Briefing — Arbitrum DAO Approves ARB Staking Proposal
• CoinEx Academy — Arbitrum's Mid-2026 Progress (Timeboost and DAO revenue)
• The Defiant — Arbitrum Foundation Funding Request and DAO Revenue Debate
• Starknet — 2025 Year in Review (staking growth)
• The Block — Starknet Enables Bitcoin Staking Alongside STRK
• Bitcoin.com News — Ethereum Staking Nears 40M ETH (2026)
• Everstake — Ethereum Staking ETFs for Institutions (2026)
• CoinDesk — Senate Opens First Stage of CLARITY Act Voting (Aug 2026)
• CoinGecko — Arbitrum (ARB) Price and Market Cap

Disclaimer: Tokenomics can and do change via governance votes, and figures cited here (supplies, unlock schedules, buyback programs, staking totals) reflect sources available as of August 2026. This guide is educational only and is not financial advice.