Arbitrum One
Executive summary
Arbitrum One is the largest Ethereum L2 by TVL, the most mature optimistic rollup architecturally, and the only major L2 to ship multi-VM execution via Stylus (which lets developers write contracts in Rust, C and C++ alongside Solidity). With ~$18-21B of TVL as of April 2026, $1.5-2B in stablecoin float, and a deeply institutional DeFi stack anchored by GMX, Camelot, Pendle and Aave, Arbitrum has positioned itself as the L2 of choice for sophisticated DeFi rather than consumer onboarding. Its Orbit framework — letting projects launch their own Arbitrum-style L3 chains anchored to Arbitrum One — has become the leading appchain stack with 60+ live deployments and powers brands like XAI, Treasure, Apechain and (in part) Celestia-anchored variants. The Arbitrum DAO controls a $1.6B-equivalent token treasury and has run the most sophisticated grants programme in crypto. Risks include weaker retail UX than Base, slower decentralisation pace than Optimism, and ARB token economics that have struggled to generate sustained accrual against substantial unlocks. Through 2027 Arbitrum's bet is that institutional and developer-grade DeFi remains the highest-value L2 segment, that BoLD fault proofs and Stylus broaden technical leadership, and that Orbit becomes the default appchain framework on Ethereum.
Origin and architecture
Arbitrum was developed by Offchain Labs, a Princeton-academic-spinout founded by Ed Felten, Steven Goldfeder and Harry Kalodner. The original Arbitrum Rollup mainnet beta launched in August 2021. The Nitro upgrade in August 2022 replaced the original AVM with a Geth-based stack compiled to WebAssembly, dramatically improving EVM-equivalence and throughput. Architecturally Arbitrum is an optimistic rollup with multi-round interactive fraud proofs (rather than the single-round model of OP Stack), which produces tighter security guarantees at the cost of more complex challenge logic. The BoLD (Bounded Liquidity Delay) upgrade, deployed to Arbitrum One in early 2025, replaced the legacy permissioned validator set with a permissionless validator system using a bounded-time challenge protocol that limits worst-case settlement delay to ~7 days regardless of attacker resources. Arbitrum settles to Ethereum L1 with batched transactions posted as EIP-4844 blob data, and the Sequencer (operated by Offchain Labs under DAO oversight) currently produces blocks at 250ms intervals. The Stylus VM, live since October 2024, lets contracts compile to WASM from Rust, C and C++; Stylus contracts share state with Solidity contracts and can invoke each other transparently, with WASM execution roughly 10-100x cheaper than equivalent Solidity for compute-heavy operations.
Stylus and multi-VM execution
Stylus is Arbitrum's most differentiated technical asset. Rather than restrict developers to Solidity, Stylus accepts any language that compiles to WebAssembly, with first-class SDK support for Rust (via the arbitrum-stylus-sdk), C and C++. Stylus contracts inherit EVM compatibility — they live in EVM-compatible addresses, fire EVM-style events, and read/write the same state trie — but execute via the WASM runtime. The performance differential is substantial: cryptographic primitives, complex math, image processing and ML inference can run at 10-100x lower gas cost than Solidity equivalents. By April 2026 Stylus had attracted 230+ deployed contracts including Renegade (dark-pool DEX with zero-knowledge proofs), Lit Protocol's threshold cryptography deployments, several gaming engines, and infrastructure tooling like account-abstraction primitives and bridge verification logic. The strategic implication is that Arbitrum can attract development categories — high-performance compute, ZK applications, gaming engines — that Solidity-only chains cannot economically host. Stylus's main limitation today is tooling maturity (debuggers, IDE integration, security-audit firms) compared to Solidity's well-developed ecosystem, but this is closing rapidly with foundations from Foundry, Hardhat-equivalent tooling and Halmos-style symbolic execution coming online.
Orbit and the L3 appchain framework
Arbitrum Orbit is the licence-free framework letting projects launch their own Arbitrum-style chains that settle to Arbitrum One (as L3s) or directly to Ethereum (as alternative L2s). Orbit chains can be configured as rollups (full security inheritance) or AnyTrust chains (using a Data Availability Committee for cheaper but weaker guarantees), can choose between EVM+Stylus or pure EVM, and can customise gas tokens. By April 2026 Orbit has 60+ live deployments, with the most prominent being XAI Games (a gaming-focused L3 with $250M+ TVL), Treasure (a gaming/NFT economy L3), Apechain (the official Yuga Labs Apecoin chain), Sanko, Cygnus and several institutional permissioned chains operated by traditional finance partners. Orbit chains generate fee revenue for their operators and contribute to the broader Arbitrum economic zone — the Arbitrum DAO has framed Orbit as 'horizontal scaling' that grows the platform without competing for Arbitrum One's settlement bandwidth. The Orbit programme has been more successful than Optimism's analogous Superchain in attracting traditional gaming brands, though smaller in stablecoin and consumer integration. Orbit's interoperability features — cross-chain message-passing via the Universal Cross-chain Messaging Protocol — are still maturing relative to OP Stack's Superchain interop.
Arbitrum DAO and ARB token
ARB launched in March 2023 with the largest L2 airdrop in history (1.275B tokens distributed to 600k+ addresses) and immediately empowered the Arbitrum DAO with a 3.55B-token treasury (~35.5% of the 10B total supply). The DAO has run the most sophisticated grants programme in DeFi, including the Short-Term Incentive Programme (STIP and STIP-Bridge), the LTIPP long-term programme, multiple Gaming Catalyst Programme grants, Stylus Sprint technical grants and several specialised committees. Cumulative DAO disbursements exceeded $300M-equivalent by early 2026. ARB itself has struggled with token economics: it is a pure governance token with no direct fee accrual, the unlock schedule has consistently created sell pressure (50% of supply unlocks linearly through March 2027), and proposals to introduce sequencer fee revenue sharing or ARB-staking based fee discounts have stalled. The 2024-25 'TimeBoost' system — an opt-in priority lane where bidders pay ARB for guaranteed inclusion at the head of blocks — became the first direct fee accrual mechanism, generating ~$8-12M annualised by Q1 2026. The DAO's governance throughput is high but contentious: major decisions like the Treasury Management Initiative (delegating $35M to Karpatkey, Avantgarde and others for yield generation) and the 'Atlas' constitutional-amendment package have produced strong debate but generally constructive outcomes.
Ecosystem and TVL
Arbitrum One TVL of ~$18-21B is the largest among L2s, composed of GMX (~$650M; the on-chain perp DEX that pioneered the GLP model), Aave v3 (~$5.5B), Pendle (~$2.1B for yield-trading), Camelot DEX (~$300M), Uniswap v3 (~$1.5B), Curve (~$400M), Stargate (~$200M cross-chain), Radiant Capital (~$150M after the 2024 exploit recovery), Balancer, Trader Joe (after merging assets from Avalanche), and a long tail of yield, derivatives and structured-product protocols. Arbitrum hosts ~$1.5-2B of stablecoin float (USDC ~$1.0B, USDT ~$700M, DAI/Sky ~$200M, smaller minted stablecoins). Daily transactions average 2.5-3M, materially below Base but with higher per-transaction value. Arbitrum is the leading L2 for derivatives volume, with GMX, Vertex, Aevo and Hyperliquid (which started on Arbitrum before launching its own chain) collectively driving 30-40% of all L2 derivatives volume. Institutional activity — RWA tokenisation by entities like Securitize, Backed and Maple Finance — has trended toward Arbitrum due to its EVM equivalence, mature security model and DAO-governed upgrades.
Sequencer and decentralisation roadmap
Arbitrum One's sequencer is operated by Offchain Labs under formal DAO oversight, with a Security Council (12 members, requires 9/12 multisig for emergency upgrades) elected through DAO governance. The decentralisation roadmap centres on three axes: (1) BoLD permissionless validators, deployed in 2025, allowing any party to challenge fraudulent assertions without permissioned whitelist — this addressed the 'Stage 1' decentralisation criterion; (2) decentralised sequencing, with the planned migration to a multi-sequencer model using either Espresso shared sequencing or a custom FairSequencer protocol, targeted for late 2026; (3) Stage 2 trust minimisation, requiring no Security Council override of routine operations and full BoLD reliance, mapped for 2027. Critics note that Arbitrum has moved more slowly toward sequencer decentralisation than Optimism in some respects, but more substantively on fraud-proof maturity. The Arbitrum DAO's ability to vote on protocol upgrades, treasury deployments and Security Council composition makes it the most politically autonomous L2 — a quality that institutional users sometimes value (predictability, due-process governance) but that occasionally produces slower consensus than centralised competitors.
Competitive position
Arbitrum's competitive set is led by Optimism (similar OP Stack architecture, smaller TVL but stronger Superchain coordination) and Base (Coinbase distribution, larger transaction count). Among ZK rollups, zkSync Era, Scroll, Linea and Polygon zkEVM compete on technical superiority claims but trail materially on TVL, developer mindshare and DeFi integrations. Arbitrum's defensive position rests on: (1) the largest L2 TVL and deepest DeFi integrations, with most major DeFi protocols deployed on Arbitrum first or simultaneously with mainnet; (2) Stylus multi-VM execution, which is unique among major L2s and unlikely to be matched in the next 18 months; (3) Orbit appchain framework with strong gaming and institutional traction; (4) DAO-governed protocol with autonomous treasury and grants programme; (5) BoLD permissionless fault proofs and mature security record. Disadvantages include weaker retail UX than Base, slower consumer-app traction (Arbitrum has no equivalent of Farcaster or Friend.tech), ARB token without direct fee accrual, and slower sequencer decentralisation than Optimism's shared sequencer ambitions.
Risks and outlook through 2027
Acute risks: ARB unlock schedule continues through March 2027, creating ~25% of supply remaining to unlock, sustaining sell pressure unless fee-accrual mechanisms scale faster. Sequencer centralisation: a regulatory action against Offchain Labs or Arbitrum Foundation could compromise sequencer operation; the multi-sequencer migration is the structural mitigation but introduces its own risks. Competitive risk: Base's distribution advantage and Solana's monolithic consumer success could compress Arbitrum's category leadership in retail-facing categories, leaving it as the institutional/developer L2 rather than the dominant L2 overall. Through 2027 the base case is continued category leadership in TVL and DeFi, with TVL growing to $30-40B as Pendle, Aave RWA expansion and Stylus-native applications mature. Orbit appchains expand to 200+ deployments and become the leading L3 framework, with several major gaming and traditional-finance brands choosing Orbit over Optimism Stack or independent chains. ARB fee accrual mechanisms — TimeBoost expansion, sequencer revenue share, possibly ARB staking against off-chain revenue — generate annualised $50-100M of value capture by 2027. Stage 2 decentralisation completes in 2027 with multi-sequencer and full BoLD reliance. Bear-case scenarios feature ARB token failure to find a sustainable demand source, a competitive shift that pushes major DeFi protocols toward Solana or alternative L2s, or a Stylus security incident that damages the multi-VM thesis. The breakthrough scenario is Stylus enabling a new application category — high-performance gaming, ZK-native finance, ML inference markets — that gives Arbitrum a structural advantage no EVM-only L2 can replicate.
Watch points
- Stylus contract deployment growth and tooling maturity
- Orbit chain count and major brand activations
- ARB fee-accrual mechanisms (TimeBoost expansion, sequencer revenue share)
- Stage 2 decentralisation timeline and multi-sequencer activation
TL;DR
Arbitrum One is the largest L2 by TVL ($18-21B), uniquely supports Rust/C/C++ contracts via Stylus, hosts the leading L3 appchain framework via Orbit, and is governed by the most sophisticated DAO in DeFi, with BoLD permissionless fault proofs and Stage 2 decentralisation as the 2027 milestones.
Get DeFi Intel research in your inbox
Weekly long-form coverage of papers, incidents, jurisdictions, chains, tokens and the people building them. Free tier covers headlines; Pro adds the analyst-grade breakdowns.