Rollup-as-a-Service (RaaS) is a managed deployment platform that lets teams launch custom Layer 2 or Layer 3 rollup chains without building or running the underlying infrastructure themselves. A RaaS provider handles the sequencer, data availability posting, bridge UI, and monitoring, so developers can ship an app-specific chain in days rather than months. Caldera, Conduit, AltLayer, Gelato, and Ankr RaaS are among the most active platforms as of Q1 2026.
​

Before RaaS, teams wanting a dedicated rollup had to fork OP Stack or Arbitrum Orbit repos, run their own sequencer nodes, write custom bridge contracts, and maintain all of it indefinitely. That work consumed engineering cycles better spent on the application layer. RaaS removes that burden by treating rollup deployment as a product, not an infrastructure project.
​

Rollup-as-a-Service is a managed deployment model where a third-party provider launches and operates a custom rollup chain on behalf of a project. The team specifies the chain framework, data availability layer, gas token, and throughput settings; the provider handles provisioning and operations. Projects get a dedicated chain without hiring DevOps engineers to run sequencer infrastructure.
​

The term borrows the "-as-a-service" framing from cloud computing, where managed providers abstract infrastructure complexity. AWS does this for servers; RaaS providers do it for rollup chains. The analogy is close: projects pay for uptime, scalability, and maintenance rather than owning the hardware or node software.
​

RaaS is distinct from the rollup frameworks it deploys. OP Stack, Arbitrum Orbit, Polygon CDK, and zkSync ZK Stack are the open-source toolkits that define how a rollup processes transactions and settles to an L1. RaaS platforms wrap those frameworks with deployment tooling, managed sequencers, bridge UIs, and observability layers. One RaaS provider may support multiple frameworks — Caldera, for instance, supports both OP Stack and Arbitrum Orbit deployments from the same dashboard.
​

For a deeper look at how rollups work at the protocol level, see What Is a Rollup: Optimistic vs ZK Rollups.
​

A RaaS deployment follows a configuration-and-provision model. The team chooses a rollup framework (OP Stack, Arbitrum Orbit, Polygon CDK, or zkSync ZK Stack), selects a data availability layer, and sets chain parameters including gas token and block time. The provider then provisions the sequencer, deploys bridge contracts to L1, and configures DA posting from a dashboard.
​

The sequencer is the most operationally sensitive component. It orders transactions, batches them, and posts compressed data to the DA layer at regular intervals. In a self-hosted rollup, a sequencer outage means transaction ordering stops until the node is recovered. RaaS providers run managed sequencers with redundancy, automatic failover, and SLA-backed uptime. Conduit, for example, publishes infrastructure status pages for every chain it operates.
​

Data availability (DA) is where batched transaction data lives before it is proven final on L1. Ethereum calldata has historically been the default; EIP-4844 blobs (shipped March 2024) cut DA costs by roughly 10x at similar throughput. Many RaaS providers now route new deployments to blob storage by default. Alternative DA layers — Celestia, EigenDA, and Avail — offer even lower marginal costs at the tradeoff of different trust assumptions. Caldera, Conduit, and AltLayer all support Celestia as a DA option alongside Ethereum.
​

Bridge infrastructure is usually co-deployed with the chain. The standard pattern uses a canonical bridge built into the rollup framework (the OP Standard Bridge for OP Stack chains, for example) plus a bridge UI that lets users move assets between L1 and the new chain. RaaS providers package that UI and host it so the project team does not need to build one from scratch.
​

Monitoring and analytics close the loop. Sequencer health, DA confirmation latency, block production rates, and bridge event streams are surfaced through dashboards that the project team can access without running their own indexers.
​

For a detailed look at OP Stack specifically, see What Is the OP Stack: Architecture and Superchain Explained.
​

Several platforms compete in the RaaS market as of 2026. Caldera, Conduit, AltLayer, Gelato, and Ankr RaaS are the most active, and they differ on supported frameworks, DA options, pricing structures, and customization depth. No single provider supports all four major rollup frameworks; framework choice narrows which RaaS platform is available.
​

Caldera focuses on gaming and DeFi chains with deeper customization hooks (custom precompiles, whitelisting). Conduit targets production-grade OP Stack deployments with strong DevOps tooling. AltLayer is the broadest by framework and DA support, including restaked security options via AltLayer MACH AVSes. Gelato adds automation services (relayers, oracle feeds) natively alongside chain deployment. Ankr bundles RaaS with its existing RPC and node infrastructure network.
​

For a focused look at Caldera specifically, see What Is Caldera: Rollup-as-a-Service Explained.
​

RaaS platforms do not build their own execution environments. They deploy chains using one of four established open-source rollup frameworks, each with a distinct architecture and ecosystem. The framework choice determines how the chain settles to L1, which cryptographic proof system it uses, and which existing tooling is compatible.
​

OP Stack is the modular framework maintained by OP Labs behind Optimism and Base. It uses an optimistic fraud-proof model where transactions are assumed valid unless challenged within a 7-day dispute window. OP Stack chains settle to Ethereum and can opt into the Superchain architecture, which shares a bridging and governance layer across chains including Optimism, Base, Zora, and over 30 others as of Q1 2026. OP Stack is the most widely deployed RaaS framework by chain count.
​

Arbitrum Orbit is Offchain Labs' framework for deploying L2 chains settling to Ethereum or L3 chains settling to Arbitrum One or Arbitrum Nova. Orbit uses Arbitrum's BOLD dispute protocol for fraud proofs. Xai — a gaming chain purpose-built for Xai Games titles — is an Orbit chain that settled over 200M transactions in 2024 before migrating to Caldera's managed infrastructure. For more on Arbitrum, see What Is Arbitrum: Ethereum's Largest L2 Explained.
​

Polygon CDK (Chain Development Kit) is a ZK-based framework that produces validity proofs rather than fraud proofs, enabling faster finality. CDK chains can participate in Polygon's AggLayer, which aggregates ZK proofs from multiple chains into a single Ethereum settlement. This gives CDK chains shared liquidity and near-instant bridge exits compared to the 7-day optimistic window. Astar zkEVM, deployed via Gelato RaaS, is a live CDK chain.
​

ZK Stack is Matter Labs' framework for deploying ZK chains using the same ZK circuits that power zkSync Era. ZK Stack chains can be either "sovereign" (self-contained finality) or "connected" (sharing liquidity with the Elastic Chain network via ZK bridges). Gelato RaaS is the primary managed provider supporting ZK Stack deployments as of Q1 2026.
​

RaaS is most valuable when an application needs dedicated throughput, a custom gas token, or transaction pricing that general-purpose L2s cannot offer. The three categories that have produced the most RaaS deployments to date are gaming, DeFi protocols seeking isolated execution environments, and enterprise applications requiring permissioned access control.
​

Blockchain games generate high transaction volumes from low-value actions (item crafts, match results, tournament scores) that are impractical on shared L2s where gas competition from DeFi activity can spike fees unpredictably. A dedicated game chain fixes gas at near-zero, lets the game studio use a native game token as gas, and keeps non-game transactions off the chain entirely. Xai (Arbitrum Orbit via Caldera) and Ronin (Axie Infinity's dedicated chain, self-hosted) are the reference deployments. Ronin processed over 560,000 daily active addresses in Q4 2023 according to DappRadar.
​

Some DeFi protocols deploy app-chains to capture MEV more aggressively, set custom precompiles that lower compute costs for their specific operations (e.g., AMM curve math), or prevent front-running by controlling sequencing order. Reya Network (AltLayer) is a perpetuals exchange built on a custom chain that removes the mempool entirely, eliminating front-running at the infrastructure level. Hychain (Caldera) targets onchain gaming economies with its own TOPIA gas token.
​

Enterprises deploying tokenized assets or running regulated workflows often need a chain where only whitelisted addresses can transact. RaaS providers support allowlist configuration at the sequencer level, so a chain operator can gate participation without modifying core protocol code. This model appears in tokenized fund administration, supply-chain provenance, and identity-gated applications where public mempool access would create compliance issues.
​

Deploying with a RaaS provider trades operational control for speed and reduced engineering overhead. RaaS cuts time to a live chain from months to days, but the team depends on the provider's sequencer uptime and upgrade roadmap. Self-hosting keeps full control but requires two to four engineers for operations before the chain serves its first transaction.
​

The cost comparison is not straightforward. Self-hosted rollup infrastructure requires dedicated engineering to set up sequencer high-availability, monitor DA posting, manage upgrades when framework versions change, and respond to outages 24/7. That engineering time has a real dollar cost before the first transaction is processed. RaaS subscription fees (typically a monthly flat rate plus DA cost pass-through) replace that overhead with a predictable line item.
​

The key differences the table below captures:
​

For most teams launching an app-chain, the relevant question is not "can we self-host" but "should we spend the engineering cycles to do so." Projects with more than a handful of engineers and a clear performance constraint that RaaS configuration options cannot address may find self-hosting justified. For early-stage deployments and games, RaaS is the practical default.
​

A Layer 3 (L3) is a rollup that settles to an L2 rather than directly to Ethereum. This reduces DA and proof costs because L2 block space is cheaper than Ethereum L1 calldata. RaaS providers support L3 deployments: Arbitrum Orbit explicitly enables L3s settling to Arbitrum One, and several Caldera-managed chains use Arbitrum One as their settlement layer.
​

L3s trade off decentralization for cost. Settlement on an L2 adds a trust layer: the L3 inherits the L2's security model in addition to Ethereum's. For gaming or high-frequency trading use cases where near-zero transaction costs matter more than maximum decentralization, L3 is often the right architecture. For DeFi protocols holding significant TVL, settling to Ethereum L1 directly (via a standard L2) is generally preferred. See What Is a Layer 3 (L3) Blockchain for a full comparison.
​

RaaS providers handle L3 deployments with the same configuration model as L2s. The primary difference from the operator's perspective is the selection of settlement chain, which affects gas costs, finality time, and which bridge infrastructure is available for moving assets between the L3 and the broader ecosystem.
​

RaaS providers expose configuration options beyond what a standard rollup framework ships with by default. These features matter for teams building application-specific chains where the default settings of OP Stack or Arbitrum Orbit do not fit the use case.
​

Custom gas tokens let a project designate its own ERC-20 token (or native token) as the chain's gas currency instead of ETH. This is common in gaming, where the game studio wants players to pay transaction fees in the game's native token rather than ETH. Caldera and Conduit both support custom gas tokens via an additional precompile layer.
​

Custom precompiles are built-in functions at the EVM level that run at lower gas cost than equivalent Solidity code. An AMM can precompile its core curve math; a ZK application can precompile its proof verifier. Precompile support requires the RaaS provider to ship custom sequencer builds, which not all providers offer — Caldera and AltLayer have more flexibility here than consumer-tier platforms.
​

Transaction whitelisting restricts which addresses can submit transactions to the mempool. This is the primary mechanism for permissioned chain deployments. The whitelist is configurable at the sequencer level and does not require changes to the rollup framework's core contracts.
​

Throughput tuning adjusts target block gas limit, block time, and batch posting frequency to match the application's transaction profile. A gaming chain expecting millions of low-value transactions per day needs different settings than a DeFi chain with fewer, larger transactions. RaaS providers expose these parameters in their dashboards; self-hosted deployments require editing node config files and restarting infrastructure.
​

RaaS is not a rollup — it is the deployment and operations layer for rollups. Ethereum provides settlement, rollup frameworks (OP Stack, Arbitrum Orbit, Polygon CDK, ZK Stack) provide execution logic, DA layers (Ethereum blobs, Celestia, EigenDA, Avail) provide data availability, and RaaS providers sit at the top, making the full combination accessible without infrastructure overhead.
​

This layering mirrors how cloud computing abstracted server hardware. Just as AWS runs on physical servers but teams interact with EC2 API calls, RaaS teams interact with configuration dashboards while Caldera or Conduit runs actual sequencer nodes. The RaaS market is still early: most deployments are under two years old, and the largest providers (Caldera, Conduit) have each deployed over 50 production chains as of Q1 2026.
​

As rollup frameworks mature toward interoperability — OP Stack's Superchain and Polygon's AggLayer both attempt cross-chain composability between chains sharing a framework — RaaS providers are adding native support for those cross-chain bridges at the deployment layer. A team deploying on Conduit today gets automatic Superchain bridge configuration if they choose OP Stack, reducing the work of connecting their chain to the broader ecosystem.
​

Eco's cross-chain stablecoin routing operates across 15 chains including several that run on OP Stack, supporting USDC transfers between chains like Base, Optimism, Arbitrum, and Unichain. As app-chain deployments multiply, routing infrastructure that spans those chains becomes increasingly important for users who hold assets on one chain and want to spend or deploy them on another.
​

A rollup framework (OP Stack, Arbitrum Orbit, Polygon CDK, ZK Stack) is the open-source software that defines how a chain processes transactions and settles to L1. RaaS is a managed service that deploys and operates chains using those frameworks. One is the blueprint; the other is the builder and property manager. See What Is a Rollup for the framework mechanics.
​

Most RaaS platforms advertise testnet deployment in under an hour via their configuration dashboards. Production deployment — including custom gas token setup, bridge configuration, and onboarding to monitoring tools — typically takes one to five business days. Compare that to the eight to twenty weeks required for a self-hosted rollup with custom infrastructure.
​

Yes, though the migration requires care. The chain's state, sequencer keys, and bridge contract addresses are all assets that need to be transferred or re-deployed. Most RaaS contracts are standard framework contracts (OP Stack or Orbit), so the underlying code is open-source and portable. Teams migrating away from a RaaS provider typically need two to four weeks of engineering work to stand up equivalent infrastructure.
​

Ethereum blobs (EIP-4844) are the safest default for production chains holding real user assets, since they inherit Ethereum's security for data availability. Celestia and EigenDA offer lower costs but different trust assumptions. For gaming chains or low-stakes applications where cost efficiency matters more than maximum security, Celestia or Avail are reasonable. High-value DeFi chains should default to Ethereum DA.
​

Several DeFi protocols run on RaaS-managed chains with material TVL. The key variables are the provider's sequencer SLA, the framework's fraud or validity proof maturity, and the DA layer choice. OP Stack and Arbitrum Orbit both have formal audit histories and active bug bounty programs. A DeFi team should audit the RaaS provider's security model as rigorously as they would audit their own contracts before launching on a managed chain. See What Is Arbitrum for context on Orbit's security assumptions.
​

Sources and methodology. Provider feature comparisons verified against official documentation from Caldera, Conduit, AltLayer, Gelato, and Ankr RaaS as of Q1 2026. Chain TVL figures from DeFiLlama, April 29, 2026. DappRadar Ronin activity data sourced from DappRadar public rankings. Framework details verified against official docs: OP Stack, Arbitrum Orbit, Polygon CDK, zkSync ZK Stack. Figures refresh quarterly.
​