What is Arbitrum (ARB)?

Arbitrum is an optimistic rollup: it executes transactions off Ethereum, posts the results and the underlying data back to it, and assumes those results are correct unless someone proves otherwise within a challenge window. Ethereum stays the settlement and data layer, so Arbitrum inherits its security rather than building its own — and users get fees an order of magnitude lower than the base layer.

"Optimistic" is the important word. Nothing is verified up front; instead, anyone can submit a fraud proof during the challenge period. That makes execution cheap and makes withdrawals to Ethereum slow — about a week in the canonical path, though bridges will front you the funds for a fee. It is the opposite trade from a zero-knowledge rollup, which proves every batch mathematically and costs more to run.

Stylus

Arbitrum runs Solidity like any EVM chain, and Stylus adds a second path: contracts written in Rust, C and C++ compiled to WebAssembly, running alongside Solidity contracts and able to call them. Computation-heavy work is dramatically cheaper this way. It has moved from announcement to production adoption, which is more than most alternative execution environments manage.

The Expansion Program is the actual business

Arbitrum licenses its technology to teams launching their own chains, which pay a share of revenue back. This is where the growth is: Robinhood Chain contributed $360,000 in licensing fees in July 2026, about 35% of the DAO's revenue that month. Arbitrum is becoming a chain vendor as much as a chain — and that is a different, more durable business than competing for users one transaction at a time.

• Governance of the Arbitrum DAO, which controls the treasury, the protocol parameters and the Expansion Program terms.
• Voting on how DAO revenue is spent — the decision that now has real money behind it.
• Incentive programmes and grants funded from the treasury.
• Not gas. Arbitrum transactions are paid in ETH, which is the crux of the section below.

The value capture question, with numbers for once

ARB is a governance token. It does not pay for gas, it is not staked to secure the chain, and holding it entitles you to nothing directly. For three years the reasonable criticism was that Arbitrum could succeed completely while ARB captured none of it.

2026 changed the terms of that argument rather than settling it. The DAO now has real, growing, high-margin income — $6.19m in six months at over 97% gross margin, plus $125m of non-ARB treasury assets — and ARB holders vote on it. That is not a dividend and nobody should describe it as one. It is control over a revenue-generating treasury, which is worth something and is worth less than ownership. Whether the DAO ever converts revenue into direct holder value, as Optimism voted to do with buybacks, is a governance decision that has not been made.

10 billion ARB total, distributed in the March 2023 airdrop with allocations to the DAO treasury, the team and investors on multi-year vesting. Those vesting schedules ran through 2024–2027 and were the dominant supply pressure on the token in its first years.

The treasury holds a very large share of the total, spendable by DAO vote. That is both the strongest asset here — a funded organisation that can pay for growth — and a persistent overhang, because every grant is eventually sold by its recipient.

Where the revenue comes from

Four streams, and the mix is shifting in an interesting direction. Transaction fees on Arbitrum One; priority auctions for sequencer access; Expansion Program licensing from other chains; and returns on the treasury itself. The licensing line is the one growing fastest, and it is revenue Arbitrum earns without needing its own chain to be busier.

In September 2026 Arbitrum replaced Timeboost — a priority-access auction that had run since April 2025 — with per-transaction priority gas auctions, routing 97% of proceeds to the DAO treasury and 3% to the Arbitrum Developer Guild. Iterating on how to monetise sequencer ordering, in public, with the revenue split published, is a more honest approach to MEV than most chains manage.

ARB cannot be staked. Arbitrum's security comes from Ethereum, not from bonded ARB, so there is no validator set to join and no protocol yield.

You can delegate your ARB's voting power to a representative, which is worth doing if you hold any — DAO turnout is low and the decisions now involve real money. It pays nothing. Anything advertising ARB staking is a lending product or a liquidity position, and the distinction matters because the risks are entirely different.

The sequencer is centralised

One sequencer, run by Offchain Labs, orders every transaction. It cannot steal funds — Ethereum settlement prevents that — but it can censor, reorder or go down, and it has gone down. Decentralising it has been on the roadmap for years. Until it ships, the chain's liveness depends on one operator.

Fraud proofs have never been used in anger

The security model rests on someone detecting an invalid state root and submitting a challenge inside the window. The mechanism exists and is permissionless. It has not been tested by a real adversarial event on a chain holding billions, and "it should work" is a different statement from "it has worked".

Governance risk is now financial risk

The DAO controls a large treasury and growing revenue, with historically low voter turnout and influence concentrated among large delegates. The thing that makes ARB interesting in 2026 — the money — is the same thing that makes its governance worth attacking.

Competition on price is a race to zero

Arbitrum competes with Base, Optimism's Superchain and every new rollup, largely on cost. Fusaka made blob space cheaper for everyone in December 2025, which is good for users and compresses the margin that funds the DAO. High margins in a commoditising market attract exactly the competition that ends them.

Token and network are only loosely joined

Gas is paid in ETH. Arbitrum's usage could double without creating a single unit of ARB demand. The link runs through the DAO treasury and governance, and it is indirect by design.