Sui Ecosystem's Walrus vs. Irys Data Dispute
TechFlow Selected TechFlow Selected
Sui Ecosystem's Walrus vs. Irys Data Dispute
This article presents an objective comparison between Walrus and Irys from six technical dimensions.
Navigating Web3 tides with focused insightsAuthor: Ponyo
Translated by: Sui Network
Key Takeaways
🔧 Architecture: Irys is a full-stack, integrated Layer 1 "data chain" that provides native blob (data block) access to contracts but requires a new set of validators. Walrus is an erasure-coded storage layer built on top of Sui, easier to integrate but requiring cross-layer coordination.
💰 Economic Model: Irys uses a single token, IRYS, to unify payment and rewards—simple for users but exposed to higher price volatility risk. Walrus splits functionality across two tokens: WAL (for storage) and SUI (for gas), effectively isolating costs but requiring maintenance of two incentive systems.
📦 Durability & Compute: Irys maintains 10 full replicas and streams data directly into its VM. Walrus uses approximately 5x redundancy via erasure coding with hash verification, achieving lower per-GB storage cost but with greater protocol complexity.
💾 Adaptability: Irys offers a donation-based "pay once, store forever" model ideal for immutable data, though with high upfront cost. Walrus uses a subscription-style "pay-as-you-go, auto-renew" mechanism, enabling better cost control and faster integration with Sui.
📈 Adoption: Walrus, despite being early-stage, is growing rapidly—already storing PB-scale data, powered by 100+ node operators, and adopted by multiple NFT and gaming brands. In contrast, Irys remains in pre-scaling phase, with sub-PB data volume and a still-growing validator network.
Both Walrus and Irys aim to solve the same core problem: reliable, incentivized on-chain data storage. However, their design philosophies diverge significantly. Irys is a purpose-built Layer 1 blockchain for data, integrating storage, execution, and consensus into a vertically integrated stack. Walrus, by contrast, is a modular storage network relying on Sui for coordination and settlement, while operating an independent off-chain storage layer.
While the Irys team initially framed their solution as the superior “built-in” approach, positioning Walrus as a limited “bolt-on” alternative, reality reveals trade-offs on both sides. This article presents a technical, six-dimensional comparison between Walrus and Irys, challenging one-sided narratives and offering developers a clear decision framework based on cost, complexity, and developer experience.
1. Protocol Architecture
1.1 Irys: Vertically Integrated L1
Irys embodies the classic "self-contained" philosophy. It features its own consensus, staking model, and execution virtual machine (IrysVM), all tightly integrated with its storage subsystem.
Validators play three roles simultaneously:
-
Store user data as full replicas;
-
Execute smart contract logic within IrysVM;
-
Secure the network via a hybrid PoW + staking mechanism.
Because these functions coexist within a single protocol, every layer—from block headers to data retrieval rules—can be optimized for large-volume data processing. Smart contracts can directly reference on-chain files, and storage proofs follow the same consensus path used for regular transactions. The advantage lies in architectural coherence: developers interact with a single trust boundary, a single fee asset (IRYS), and enjoy native-like data access within contract code.
The trade-off is high bootstrapping cost. A new Layer 1 must recruit hardware operators from scratch, build indexers, launch block explorers, harden clients, and cultivate developer tooling. In its early stages, before validators are well-established, block finality guarantees and economic security lag behind mature chains. Thus, Irys sacrifices ecosystem launch speed for deeper data integration.
1.2 Walrus: Modular Overlay Layer
Walrus takes a fundamentally different approach. Its storage nodes operate off-chain, while Sui’s high-throughput L1 handles ordering, payments, and metadata via Move smart contracts. When a user uploads a blob, Walrus shards it and distributes fragments across nodes, then records an on-chain object on Sui containing content hashes, shard assignments, and lease terms. Renewals, slashing, and rewards are executed as standard Sui transactions—paid in SUI gas, but settled in WAL tokens for storage economics.
By leveraging Sui, Walrus immediately gains several advantages:
-
Proven Byzantine fault-tolerant consensus;
-
Mature development infrastructure;
-
Strong programmability;
-
A liquid base token economy;
-
Access to a large pool of existing Move developers who can integrate without protocol migration.
However, this comes at the cost of cross-layer coordination. Each lifecycle event (upload, renewal, deletion) must be synchronized across two semi-independent networks. Storage nodes must trust Sui’s finality while maintaining performance during Sui congestion; meanwhile, Sui validators do not verify whether data is actually stored on disk, so accountability relies entirely on Walrus’ cryptographic proof system. Compared to monolithic designs, this architecture inevitably introduces higher latency and routes part of transaction fees (SUI gas) to actors who don’t store data.
1.3 Design Summary
Irys adopts a vertical, monolithic architecture, while Walrus follows a horizontal, modular integration strategy. Irys enjoys greater architectural freedom and a unified trust boundary, but faces the challenges of cold-start ecosystem building. Walrus leverages Sui’s mature consensus to drastically lower the barrier for developers already in the ecosystem, but must manage the complexity of coordinating two economic domains and operational systems. Neither approach is inherently superior—each optimizes differently: one for coherence, the other for composability.
When protocol choice depends on developer familiarity, ecosystem appeal, or time-to-market, Walrus’s layered model may offer more practical relevance. Conversely, when the bottleneck lies in deep data-compute coupling or custom consensus logic, Irys’s data-native chain justifies its heavier architectural burden.
2. Token Economics & Incentives
2.1 Irys: One Token Powers the Entire Stack
Irys’s native token, IRYS, underpins the entire platform economy:
-
Storage fees: Users prepay in IRYS to store data;
-
Execution gas: All smart contract calls are priced in IRYS;
-
Miner rewards: Block subsidies, storage proofs, and transaction fees are all paid in IRYS.
Since miners handle both data storage and contract execution, computational revenue can offset lower storage yields. In theory, strong DeFi activity on Irys could subsidize data storage via compute income, enabling near-cost pricing. If contract traffic drops, the subsidy adjusts inversely. This cross-subsidy helps balance miner incentives and aligns stakeholder interests. For developers, a single asset simplifies custody and improves UX—especially valuable in applications where multi-token exposure is undesirable.
But the downside is single-asset risk correlation: if IRYS price falls, both compute and storage rewards decline simultaneously, squeezing miners from both sides. Thus, the protocol’s economic security and data durability are tied to the same volatile price curve.
2.2 Walrus: Dual-Token Economic Model
Walrus separates responsibilities across two tokens:
-
$WAL: The economic unit of the storage layer. Users pay WAL for leasing space; node operators earn WAL rewards through staking and storing data fragments, scaled by their delegated stake weight.
-
$SUI: The gas token for on-chain coordination. Any transaction on Sui—upload, renew, slash—consumes SUI, which goes to Sui validators, not Walrus storage nodes.
This separation keeps storage economics clean: WAL’s value depends only on data demand and lease duration, insulated from DEX trading surges or NFT manias on Sui. At the same time, Walrus inherits Sui’s liquidity, bridges, and fiat on-ramps—most Sui builders already hold SUI, so adopting WAL has low marginal cost.
Yet dual tokens introduce incentive fragmentation. Walrus nodes cannot earn SUI gas fees, so WAL’s price must independently cover hardware, bandwidth, and return expectations. If WAL stagnates while SUI gas spikes, user costs rise without direct benefit to storage providers. Conversely, a DeFi boom on Sui boosts validator earnings but leaves Walrus nodes unaffected. Maintaining long-term balance thus requires active economic tuning: storage prices must flexibly adjust based on hardware costs, demand cycles, and WAL market depth.
2.3 Design Summary
In short, Irys offers a unified, seamless UX but concentrates risk. Walrus draws clear boundaries at the token level, enabling finer-grained economic accounting, but must manage two markets and split fee flows. Builders must decide: do they prioritize frictionless integration or prefer isolated risk management aligned with their product roadmap and funding strategy?
3. Data Durability & Redundancy Strategy
3.1 Walrus: Lightweight High Reliability via Erasure Coding
Walrus splits each blob into k data shards and adds m redundant parity shards (using RedStuff encoding). This technique resembles RAID or Reed-Solomon coding but is optimized for decentralized, highly dynamic environments. As long as any k out of k+m shards are available, the original file can be reconstructed—offering two key benefits:
-
High space efficiency: With typical parameters (~5x expansion), Walrus cuts storage overhead nearly in half compared to traditional 10x replication. Storing 1GB of data consumes about 5GB of total network capacity (distributed across nodes), whereas a full-replica system might need 10GB for similar safety.
-
On-demand repair capability: Walrus’s encoding saves not just space but also bandwidth. When a node goes offline, only the missing shard—not the entire file—is regenerated, significantly reducing bandwidth usage. This self-healing mechanism downloads only O(blob_size / number_of_shards), versus O(blob_size) in replica-based systems.
Each shard-node mapping is recorded as an object on Sui. Walrus rotates its staking committee every epoch, cryptographically challenges node availability, and automatically re-encodes when node loss exceeds thresholds. Though complex—spanning two networks, many shards, and frequent proofs—this mechanism achieves maximum durability with minimal capacity.
3.2 Irys: Conservative But Robust Multi-Replication
Irys deliberately chooses a simpler, more direct durability method: each 16TB data partition is fully replicated by 10 staked miners. The protocol prevents double-counting on the same drive using miner-specific “salts” (Matrix Packing). Nodes undergo continuous “proof-of-useful-work” checks—random reads verifying every byte exists—or face slashing.
Data availability hinges on whether at least one of the 10 miners responds to queries. If a miner fails validation, the system immediately initiates re-replication to restore the 10-copy standard. This strategy incurs 10x storage redundancy but is simple and transparent—all state resides on one chain.
3.3 Design Summary
Walrus focuses on using efficient encoding and Sui’s object model to handle frequent node churn, ensuring durability without raising costs. Irys bets that with rapidly declining hardware prices, a heavier but simpler multi-replica approach is more reliable and operationally easier.
If you’re storing PB-scale archival data and accept higher protocol complexity, Walrus’s erasure coding wins on per-byte economics. But if you prioritize operational simplicity—one chain, one proof, abundant redundancy—and view hardware costs as negligible compared to delivery speed, Irys’s 10-copy model offers durable peace of mind.
4. Programmable Data & On-Chain Computation
4.1 Irys: Native Data-Aware Smart Contracts
Because storage, consensus, and the IrysVM share the same ledger, contracts can easily call read_blob(id, offset, length) like accessing internal state. During block execution, miners stream requested data chunks directly into the VM, perform deterministic checks, and continue processing—all within a single transaction. No oracles, no manual input, no off-chain relay.
This enables powerful use cases:
-
Media NFTs: Metadata, high-res images, and royalty logic fully on-chain, enforced at the byte level.
-
On-chain AI: Run inference directly on model weights stored in partitions.
-
Big data analytics: Contracts scan logs, genomic files, or large datasets without external bridges.
Gas cost scales with bytes read, but the UX remains a single IRYS-denominated transaction.
4.2 Walrus: “Verify Then Compute” Pattern
Since Walrus cannot stream large files directly into the Move VM, it uses a “hash commitment + witness” pattern:
-
When a blob is stored, Walrus records its content hash on Sui;
-
Later, callers submit specific data fragments along with lightweight proofs (e.g., Merkle paths or full hashes);
-
Sui contracts recompute the hash and compare it against Walrus metadata. If valid, the data is trusted and further logic proceeds.
Advantages:
-
Immediate usability without modifying the L1 protocol;
-
Sui validators never process GB-scale raw data.
Limitations:
-
Manual data fetching required: Callers must retrieve data from Walrus gateways or nodes and include limited-size chunks in transactions (constrained by Sui’s size limits);
-
Overhead from fragmented processing: Large computations require multiple micro-transactions or off-chain preprocessing + on-chain verification;
-
Dual gas cost: Users pay SUI gas (for verification) and indirectly fund WAL (for underlying storage).
4.3 Design Summary
If your app needs contracts to process several MB of data per block—such as on-chain AI, immersive media dApps, or verifiable scientific workflows—Irys’s embedded data API is more compelling.
If your use case emphasizes data integrity proofs, small media display, or computation primarily off-chain with only results verified on-chain, Walrus is already sufficient.
Ultimately, the choice isn’t about “whether it’s possible,” but where you want to place complexity: at the protocol layer (Irys) or the middleware/application layer (Walrus)?
5. Storage Duration & Permanence
5.1 Walrus: Pay-As-You-Go Rental Model
Walrus uses fixed-term leases. Upon upload, users pay $WAL to purchase storage for a defined period (billed in 14-day epochs, up to ~2 years max). After expiry, nodes may delete the data unless renewed. Apps can write auto-renewal scripts via Sui smart contracts, turning rentals into de facto permanent storage—but renewal responsibility always rests with the uploader.
Benefits: No prepayment for potentially abandoned capacity; pricing tracks real-time hardware costs. Lease expiration allows garbage collection, preventing accumulation of “forever junk.” Drawbacks: Missed renewals or drained funds result in data loss; long-running dApps must maintain “keep-alive” bots.
5.2 Irys: Protocol-Guaranteed Permanent Storage
Irys offers Arweave-like “permanent storage.” A one-time $IRYS payment funds a chain-based endowment to cover miners’ storage costs for centuries (assuming continued cost declines, covering ~200 years). Post-payment, renewal responsibility shifts to the protocol—users no longer need to manage it.
The result is a “store once, available forever” UX, ideal for NFTs, digital archives, and tamper-proof datasets (e.g., AI models). Downside: high initial cost; long-term viability depends heavily on $IRYS’s price stability over decades. Unsuitable for frequently updated or temporary files.
5.3 Design Summary
Choose Walrus if you want control over data lifecycle and pay-per-use billing. Choose Irys if you need unshakeable long-term data permanence and are willing to pay a premium.
6. Network Maturity & Adoption
6.1 Walrus: Production-Grade Scale
Walrus mainnet has been live for only 7 epochs, yet already supports 103 storage operators, 121 storage nodes, and 1.01 billion WAL staked. The network stores 14.5 million blobs, has triggered 31.5 million blob events, averages 2.16MB per object, and holds 1.11PB of data (26% of its 4.16PB physical capacity). Upload throughput is ~1.75KB/s, with a sharding graph spanning 1,000 parallel shards.
Economic momentum is strong:
-
Market cap ~$600M, FDV ~$2.23B;
-
Storage price: ~55K Frost per MB (~0.055 WAL);
-
Write price: ~20K Frost per MB;
-
Current subsidy rate: 80%, accelerating early growth.
Multiple high-traffic brands—including Pudgy Penguins, Unchained, and Claynosaurs—have adopted Walrus for asset pipelines or archival backends. The network serves 105,000 accounts, with 67 projects in integration, supporting PB-scale data flows for real-world NFT and gaming applications.
6.2 Irys: Still in Early Stage
According to Irys public dashboards (as of June 2025):
-
Contract execution TPS ≈ 13.9, storage TPS ≈ 0
-
Total stored data ≈ 199GB (officially claims 280TB space)
-
Data transactions: 53.7 million (13 million in June alone)
-
Active addresses: 1.64 million
-
Storage cost: $2.50 / TB / month (temporary), or $2.50 / GB (permanent)
-
Mining system “coming soon” (uPoW mechanism not yet live)
Programmable data access costs $0.02 per chunk, but since the permanent storage endowment is not yet funded, actual data writes remain extremely limited. While contract execution throughput is decent, bulk storage capacity is effectively zero—indicating current focus on VM functionality and dev tools rather than data scalability.
6.3 What the Numbers Mean
Walrus has reached PB-scale, generates revenue, and has passed rigorous testing by consumer NFT brands. Irys remains in early bootstrap mode—feature-rich, but awaiting miner participation and meaningful data volume.
For evaluating production readiness, Walrus currently demonstrates:
-
Higher real-world usage: Over 14 million blobs uploaded, PB-scale storage;
-
Broad operational scale: 100+ operators, 1,000 shards, over $100M staked;
-
Stronger ecosystem pull: Top Web3 projects actively integrating;
-
Clearer pricing: Transparent WAL/Frost fees and visible on-chain subsidy mechanisms.
While Irys’s integrated vision may unlock advantages in the future (miner launch, endowment activation, TPS scaling), based on current measurable throughput, capacity, and user adoption, Walrus holds a tangible lead.
7. Looking Ahead
Walrus and Irys represent opposite ends of the on-chain storage design spectrum:
-
Irys consolidates storage, execution, and economics into a single token (IRYS) and a purpose-built L1 blockchain for data, offering frictionless on-chain big data access and protocol-level “permanent storage” guarantees. The trade-off is migrating to a younger ecosystem and accepting higher resource consumption.
-
Walrus builds an erasure-coded storage layer atop Sui, reusing mature consensus, liquid infrastructure, and robust dev tooling to achieve highly cost-effective per-byte storage. However, its modularity brings added coordination complexity, dual-token UX, and ongoing attention to lease renewals.
The choice isn’t about right or wrong—it’s about identifying your critical bottleneck:
-
If you need deep data-compute composability or protocol-level “forever storage” guarantees, Irys’s integrated design is better suited.
-
If you prioritize capital efficiency, fast time-to-market on Sui, or fine-grained control over data lifecycle, Walrus’s modular approach is the more pragmatic choice.
Going forward, both are likely to coexist as the on-chain data economy expands, serving different types of developers and applications.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
Sui Network
