Xandeum vs Filecoin, Arweave & Walrus
Decentralized storage is not one market. These four networks solve different problems. Here is an honest 2026 comparison, including where Xandeum is genuinely differentiated and where it is still the smallest and least proven.
| Filecoin | Arweave | Walrus | Xandeum | |
|---|---|---|---|---|
| Base chain | Own L1 | Own chain (Blockweave) | Storage nodes + Sui | Storage layer for Solana |
| Storage model | Deal-based sealed sectors | Permanent replicated copies | Erasure-coded "slivers" (Red Stuff) | Sharded + erasure-coded, file-system-like |
| Proof | PoRep + PoSt (24h audits) | Proof of Access | On-chain availability (Sui objects) | Merkle roots anchored to Solana, then Bitcoin |
| Data access | Whole-file; contracts can't read bytes | Whole-file via HTTP | Whole-blob; Sui can manage availability | Random access, contract-native (peek, poke, prove) |
| Pricing | Rental (market deals) | Pay once, store forever | Rental per epoch | Usage-based; operators paid via STOINC |
| Market cap (Jun 2026) | ~$590M | ~$120M | ~$75-88M | ~$1-2M |
| Live scale | Multi-exabyte committed | Multi-petabyte permaweb | ~4 PB, 100 nodes | Early; exabyte is the goal |
| Maturity | Since 2020 | Since 2018 | Mainnet ~2025 | Mainnet early 2026 |
Market caps and scale are June 2026 snapshots and move daily.
The short version
Filecoin and Arweave are the mature incumbents for archival storage: huge capacity, multi-year track records, but their stored data is not something a smart contract reads directly. Walrus is a newer, fast-growing option for programmable blobs and data availability on Sui. Xandeum is the newcomer with a narrower, sharper aim: storage that Solana programs can read and write at fine granularity, like a second tier of accounts.
Per-network detail
Filecoin (FIL)
A storage marketplace matching clients to providers via on-chain deals; data is sealed into 32 or 64 GiB sectors. It proves storage with PoRep (a unique replica exists) and PoSt (it still exists, audited daily). The FVM works on chain state, not the stored bytes, and retrieval can take seconds or hours depending on whether a provider keeps an unsealed copy. Best for archival, datasets and backups at massive scale; its known challenge is converting committed capacity into paid demand.
Arweave (AR)
The "permaweb": pay once and data is stored for 200+ years, funded by an endowment. Miners prove they can access random historical blocks (Proof of Access). Data is served as whole files over HTTP, which is why Solana's Metaplex uses it for NFT media. The flip side of permanence is that it is immutable and not designed for mutable, random-access app state.
Walrus (WAL)
Stores erasure-coded slivers across nodes with high durability at about 4.5× replication, coordinated by Sui. Move contracts on Sui can check availability and manage blob lifecycles, so it is "programmable" at the blob level, though still whole-blob retrieval rather than byte-level random reads. Newer and smaller than Filecoin and Arweave, with a relatively concentrated node set.
Xandeum (XAND)
A storage layer that behaves like a second tier of Solana accounts. Provider nodes (pNodes) store data; the network reaches consensus only on Merkle roots each ~15-second pulse, then anchors them every Yuga (~2.8 days) to Solana and then Bitcoin for tamper-evidence. Its headline feature is random-access, smart-contract-native storage so Solana programs read and write data directly, keeping execution lean. Operators earn usage-based STOINC income.
Where Xandeum stands out
- Contract-native random access for Solana. The thing the others do not target: programs reading and writing stored data at fine granularity, not whole-file retrieval.
- A complementary tier, not a competing chain. It extends Solana rather than asking apps to leave it.
- Bitcoin-anchored proofs for externally verifiable integrity.
- STOINC ties operator rewards to real storage usage.
The honest caveat
Xandeum is roughly 40 to 500 times smaller than these incumbents and only reached mainnet in early 2026. "Exabyte-ready" is an architectural goal, not demonstrated scale. Its pulse-and-anchor design is novel and far less battle-tested than Filecoin's PoRep/PoSt or Arweave's years of Proof of Access. It is a genuinely different tradeoff for a specific job, not a proven "better" across the board. If you need permanent archival, Arweave fits; for cheap cold capacity, Filecoin; for Sui-native blobs, Walrus.
An independent comparison by Pulsar Network. Figures are June 2026 snapshots and not financial advice.