Layer 1 blockchain
Arweave ticker AR
Summary
Arweave, live since June 2018, is a blockchain for permanent data storage whose blocks both accept new data and prove access to old data. Mining uses Succinct Proofs of Random Access (SPoRA): a verifiable delay function releases a new mining seed about once a second, the seed picks chunks in the miner's own stored data and across the whole dataset (the weave), and hashes over those chunks are checked against the network difficulty. Each 3.6 TB partition must be packed to the miner's own address (the replica.2.9 format since the February 2025 hard fork), so mining power follows how many unique packed replicas a miner stores and can read. Blocks target two minutes and nodes follow the heaviest fork. Uploaders pay a fee plus an upfront contribution to a storage endowment that pays miners over time. 1571416171821
Design
- System
- Layer 1 blockchainIndependent proof-of-access chain for permanent data storage; it does not settle to another chain. Every block accepts new data and token transfers and also proves that the miner can read old data. Its early papers call the structure a blockweave. AO, a compute system whose messages are stored on Arweave, and gateway networks such as ar.io are separate systems built on top of it, so their features are not assessed as Arweave's own.
- Settlement family
- Its own design
- Scarce resource
- Other
- State model
- Accounts
- Finality
- Probabilistic
- Who makes blocks
- Any miner whose packed data yields a hash that meets the difficulty produces the next block; there is no stake, validator set or permission. Full efficiency needs a complete packed replica (about 373 TB in November 2025), so coordinated multi-node miners dominate and pool miners may hand their mining keys to the pool. Most miners use a VDF server; the docs list public ones, and trusted peers run by the Digital History Association. Rewards are held for about 30 days, and double-signing can forfeit them and revoke the mining address.
- Fork choice
- Heaviest fork: nodes select and advertise the fork with the most accumulated difficulty. A miner whose block is orphaned can rebase the same solution onto the new tip if its VDF step is later than the tip's. The reference node keeps a checkpoint depth of 18 blocks and does not perform fork recovery deeper than that, a node rule rather than a finality vote.
Qualifications
- Settlement family · Partial. Recorded as other. Arweave's own terms: its own Erlang node software (the arweave repository) validates blocks under SPoRA mining with a verifiable delay function, an account tree of RSA and secp256k1 addresses, and a weave of data transactions whose chunks are committed by Merkle roots; it derives from no other chain's client.
- Scarce resource · Partial. Recorded as other. Arweave's own terms: Succinct Proofs of Random Access over packed replicas. Mining power follows how much of the weave a miner stores packed to its own address and can read at the pace the VDF sets; packing needs RandomX work, and the one-second VDF limits how often data can be sampled. AR is not staked for block production.
- Reorg depth · Partial. Settlement is probabilistic, but the reference node sets a checkpoint depth of 18 blocks and does not recover forks deeper than that. The mining docs report a mined block as accepted by the network after 10 confirmations, about 20 minutes, and hold block rewards for about 30 days.
- Staking · Not applicable. AR is not bonded or delegated for consensus. Mining pools share rewards under their own rules outside the protocol, and the docs state that Arweave has no account delegation mechanism.
- Governance · Not applicable. No on-chain voting. Upgrades ship as hard or soft forks in node releases that miners choose to run; recent releases were prepared by the Digital History Association, and the whitepaper describes a social, fork-based framework for evolving the protocol.
- Smart contracts · Partial. The base protocol runs no contracts. AO is a separate compute system whose processes store their message history on Arweave; its own nodes compute the results, and Arweave consensus neither executes nor checks them. AO's docs say its current network, HyperBEAM, launched in February 2025.
- Data permanence · Contested. Uploads prepay for 200 years of replicated storage at present prices, assuming storage costs fall at least 0.5% a year; if the endowment runs low, the protocol halves that assumed rate, raising prices for new data. The node code notes that further token issuance may start if the endowment pool runs empty, so miners' pay would then come from new tokens rather than the prepaid pool. Data stays retrievable only if the uploader seeded it and nodes choose to store and serve it under their own content policies.
- Data access · Partial. Arweave's docs leave high-bandwidth data delivery, indexing and search to gateways and other services, since miners are rewarded for storage, not bandwidth. ar.io says operation of the arweave.net gateway passed from ar.io back to Forward Research on 1 March 2026, with temporary slowdowns; ar.io runs a separate network of gateways. Gateways choose what to index and serve.
- Tps · Contested. The only throughput figures are design limits and design claims without a published method; they are kept in the classified claims list and are not comparable with observed rates.
Tradeoffs
- Emphasizes
- Decentralization and Security Contested
- Gives up
- Base-layer capacity and quick settlement. Blocks come about every two minutes with a capped number of top-level transactions, settlement is probabilistic, and the base protocol runs no contracts; bulk data scales through third-party bundlers that pack many signed data items into one transaction. Mining is open, but full efficiency needs a complete packed replica of a dataset measured in hundreds of terabytes, which pushes most miners into coordinated clusters or pools.The whitepaper presents the proof system as keeping verification light to maximize decentralization and says bundling lets throughput scale without practical limits; reading that as all three corners at once is contested. Permanence also depends on endowment economics, on uploaders seeding their data and on miners choosing to store it.
- Full node at home
- Unknown 567811Every Arweave node needs a data directory on an SSD (at least 256 GB recommended), and a validator that does not mine need not store the weave: it joins by fetching recent blocks and state from trusted peers and validating them. The docs give no other validator sizing and warn that a node falls behind if its CPU cannot keep up with validation and the one-second VDF. Mining at full efficiency needs a full packed replica (about 373 TB in November 2025), which is not home-scale.
- Throughput claims
- Theoretical peak: 1,000 51417Theoretical peak; not comparable across chains or with observed load.Top-level transactions per block, with blocks targeting about two minutes; bundled data items are not counted.A design limit set in the node code and described in the 2023 whitepaper, not a measured rate. One top-level transaction can carry a bundle of many separately signed data items, so this limit does not bound how many uploads the network records. No observed transaction rate, method or hardware assumptions were found in the reviewed sources.
- Theoretical peak 1423Theoretical peak; not comparable across chains or with observed load.Data write throughput through recursive bundling; no figure given.The whitepaper says bundling lets throughput scale without practically applicable limits; this is a design claim with no measurement. Bundling is done by bundling services outside the protocol, so throughput beyond the base layer depends on those services and on trusting them to post the data. The claim covers data writes, not token transfers: bundled data items cannot transfer AR.
Scaling layers
No scaling layer recorded in this profile.
Capabilities
| Capability | How it is provided | Notes and sources |
|---|---|---|
| Account abstraction | Structured assessment pending | Not yet assessed for any chain. |
| Atomic swaps | Unknown | No swap tool for AR was found. A base-layer transaction carries one owner signature, an optional transfer to one target, data and tags, with no script, hash lock or time lock, so the AR side of a swap cannot be locked on Arweave itself; designs that keep every lock on the other chain and link the two sides through signatures were not reviewed for AR. Bundled data items cannot transfer tokens, and trades of wrapped AR inside AO processes belong to that separate system. 21223 |
| Authenticated data publication | Built into the protocolLimited scope | Protocol: the uploader signs a transaction committing to the Merkle root of its data's chunks, and the block's transaction root and block index commit to it (1). Miners, who produce blocks, mine by holding packed chunks of the weave and proving access to them, and nodes and gateways serve data over the HTTP API (2); anyone can check the data or a chunk's Merkle proof against the signed root (3). Limited: updates are new transactions found through tags and gateway indexes outside the protocol, and data stays retrievable only if it was seeded and nodes choose to store and serve it. 5910121427 |
| Light clients | Built into the protocolLimited scope | Protocol, limited: miners embed the Block Index root, a Merkle commitment to every earlier block's hash, weave size and transaction root, in each new block, and the whitepaper says a node that has checked the proofs of the latest blocks can then validate any older block without downloading or validating the headers in between. Limits: no light client or wallet using this was found in the reviewed docs; new nodes fetch recent blocks and state from trusted peers, and apps usually read through gateways. 61114 |
| Native staking or delegation | Structured assessment pending | Not yet assessed for any chain. |
| On-chain governance | Structured assessment pending | Not yet assessed for any chain. |
| Parallel execution | Structured assessment pending | Not yet assessed for any chain. |
| Payment or state channels | None found Earlier definition | Nearest feature: a plain transfer. A base-layer transaction carries one owner signature and no script, lock, multisignature or time condition, and the docs state that Arweave has no account delegation, so the protocol offers no joint or timed spending rule that a channel's dispute or refund path could rely on. 212 |
| Programmable spending | None found Earlier definition | An AR balance moves only with a transaction signed by that account's own RSA or secp256k1 key; there are no scripts, multisignature accounts or account delegation. Nearest feature: AO processes, which are programmable but run in a separate compute system that stores its messages on Arweave and is not executed or checked by Arweave consensus. 21224 |
| Protocol-verified messaging | Structured assessment pending | Not yet assessed for any chain. |
| Reversible transfers or recovery | Unknown | (a) The protocol cannot reclaim or delay a confirmed transfer; the nearest rule, the transaction anchor, only makes an unconfirmed transaction invalid once its anchor block is older than the last 50 blocks, a replay guard rather than a set cancellation window. (b) The docs state that Arweave has no account delegation mechanism, and the wallet FAQ says the team cannot help recover a lost wallet. Wander's site advertises social recovery for its accounts; whether that meets this row's recovery definition was not assessed. 2121338 |
| Rollups | Unknown | AO and other systems store their messages or data on Arweave, but Arweave consensus does not check their execution; whether any system qualifies as a rollup anchored to Arweave was not established. 2425 |
| Shielded transfers | Structured assessment pending | Not yet assessed for any chain. |
| Signed partial offers | Not present | Nearest feature: a plain transfer. Each base-layer transaction has a single signer, its owner, so a maker cannot sign one side for a taker to complete in the same transaction, and bundled data items cannot transfer tokens. Order books or pools in AO processes belong to that separate system. 1223 |
Reading these values
- How this feature is provided
- Built into the protocol, core-team software or independent software says where a feature lives and who can change it. These are implementation layers, not quality: core-team software is not closer to the protocol than independent software.
- Structured assessment pending
- The shared review has not assessed this feature for any chain yet (Account abstraction, Native staking or delegation, On-chain governance, Parallel execution, Protocol-verified messaging and Shielded transfers). That does not mean it is absent, and a chain’s profile may already describe it.
- Unknown
- The review has not established this for this chain under the current definition; the note beside it says why. Unknown is not absent and not a low score. A feature reads “Not present” only when a source shows it is absent.
- Earlier definition
- Payment or state channels, Programmable spending and Rollups keep the earlier definition until the next fact-check. There, “Native” does not separate the protocol from core-team software, “Ecosystem software” does not say who publishes it, and “Partial” does not say which layer.
- Reading across rows
- The list of capabilities is incomplete. Do not read these rows as a chain leading or lagging overall.
Ecosystem & custody
| Product type | Status | Notes and sources |
|---|---|---|
| Automated market makers | Unknown | Arweave's base layer has no contracts or pools, but Arweave's own docs count compute layers such as AO as part of the Arweave ecosystem, and exchange venues run there as AO processes: Botega, for one, is listed by its builder Autonomous Finance among 2024 AO products that are now unmaintained but still online. Which AO venues are live, and whether they trade AR itself or only wrapped AR, was not reviewed. 3122430 |
| Issuer-native stablecoins | None found | Neither Circle's USDC contract list nor Tether's supported-protocol list includes Arweave, and the base layer has no token contracts. 3132 |
| Algorithmic stablecoins | Unknown | None can run on Arweave's base layer, which has no contracts; tokens issued by AO processes were not reviewed. 12 |
| Bridges | Unknown | No bridge is part of the protocol. Wrapped AR used in AO depends on bridges run outside Arweave, and no bridge operator's documentation was readable at review. Risk: Any wrapped AR is only as safe as the custody of the locked AR and the bridge's signers, none of which were verified here. |
| Block explorers | ThinViewBlock (also serving arscan.io) | Arweave's wallet FAQ points users to ViewBlock to check transfers. ArScan's page carries ViewBlock's page description, links to ViewBlock's blog, social and GitHub accounts and loads ViewBlock's assets, so it appears to be a second front end of the same operator rather than an independent explorer. ar.io's docs list turbo-gateway.com and Goldsky for GraphQL search, which are data search services rather than block explorers. Listing is not a quality or completeness check, and explorer data is provider-indexed. 13272829 |
| Hardware wallets | ThinKeystone (through Wander, formerly ArConnect) | Keystone's own guide shows a Keystone 3 Pro on firmware V2.0 or later connecting to the Wander browser extension, named ArConnect before 2025, and signing AR transfers by QR code. Arweave's wallet FAQ still says Ledger and Keystone are working on integrating Arweave, and Trezor says it does not support Arweave. See the custody rows. 13363738 |
Hardware-wallet custody
| Device maker | Status | What users can and cannot do |
|---|---|---|
| Ledger | UnknownSame as native: Unknown | Cannot: Find Arweave in Ledger's published Ledger Wallet currency list (@ledgerhq/cryptoassets 13.56.0) Cannot: Install an approved Arweave device app: the app-arweave repository in Ledger's GitHub organization says it is not yet approved and must not be used in productionLedger's Arweave page (modified 2026-02-20) is a generic template that invites users to manage Arweave in Ledger Live, but Ledger's currency list has no Arweave entry, arweave.app lists its Ledger support as awaiting release on Ledger's app store, and Arweave's wallet FAQ calls full support pending. No working path for ordinary users was confirmed. 1333343539 |
| Trezor | None foundSame as native: No | Cannot: Manage AR in Trezor Suite; Trezor's Arweave page says it does not currently support ArweaveTrezor's page asks users who used this asset with Trezor before to contact its support team. 36 |
Public data
iKnow Blockchain has no public-data lookups for Arweave yet. This is a limit of the service, not a statement about the network.
| Lookup | Status | What it covers and its limits |
|---|---|---|
| Address or account | Not available yet | Public-data lookups for this chain are not built yet. |
| Tokens and assets | Not available yet | Public-data lookups for this chain are not built yet. |
| NFTs | Not available yet | Public-data lookups for this chain are not built yet. |
| Transactions | Not available yet | Public-data lookups for this chain are not built yet. |
Sourced developments
Developments: Reviewed Sep 29, 2026 Next review due Oct 9, 2026, 12:00 UTC.
Publication date · newest first
-
Arweave 2.9.6 ends acceptance of format 1 transactions from height 2,000,000
Arweave node 2.9.6 proposes a soft fork that ends the grace period for format 1 transactions, deprecated since April 2020, from block height 2,000,000, estimated for 13 September 2026 at 12:00 UTC. The notes say format 1 now makes up about 0.007% of base-layer interactions and complicates the node's transaction handling; after the activation height, blocks that include a format 1 transaction are not accepted by upgraded peers.
- Proposal:proposed in the 2.9.6 release notes as a soft fork
- Implementation:shipped in N.2.9.6 and carried in the 2.9.7-alpha1 build
- Release:stable release published 2026-09-07
- Activation:scheduled at height 2,000,000; the network passed that height by 2026-09-29, but enforcement across miners was not verified
Sources: ArweaveTeam (GitHub) — Arweave 2.9.6 release notes (external site) · ArweaveTeam (GitHub) — Arweave 2.9.7-alpha1 release notes (external site) · arweave.net gateway — arweave.net /info network read (external site)
-
Arweave node 2.9.5.1 patches input-validation crashes
Arweave node 2.9.5.1 fixes several input-validation steps that could crash on invalid values and in some cases halt a node, adding graceful validation and defensive deserialization of local binaries. The notes recommend upgrading as soon as possible, state that no funds are at risk, and credit four outside researchers and security teams.
- Implementation:shipped in N.2.9.5.1
- Release:recommended patch release
- Activation:no activation height; takes effect when each node upgrades
Sources: ArweaveTeam (GitHub) — Arweave 2.9.5.1 patch release notes (external site) · ArweaveTeam (GitHub) — Arweave 2.9.7-alpha1 release notes (external site)
-
ar.io handed operation of the arweave.net gateway back to Forward Research
ar.io wrote that on 1 March 2026 it transitioned operation of the arweave.net gateway, the default access point many apps hardcode, back to the Forward Research team. It said the handover brought temporary instability, with slower loads and occasional timeouts, while the underlying Arweave data was unaffected, and it urged apps to route requests across several gateways with its Wayfinder protocol.
- Implementation:operator change completed on 1 March 2026 according to ar.io
- Release:not a software release
- Activation:in effect from 1 March 2026; no statement from Forward Research itself was found
Sources: ar.io — The Dock Isn't the Lake (external site) · Arweave Docs — HTTP API (external site)
-
Arweave node 2.9.5 released as a stability and performance update
Arweave node 2.9.5, prepared by the Digital History Association, rolls up six alpha builds into a bug-fix, stability and performance release. It fixes a mining bug that counted hashes over unpacked entropy, so miners' reported hashrate drops after upgrading; adds an optional faster VDF implementation; supports repacking data out of the replica.2.9 format; and adds metrics for endowment pool flows and the Kryder+ rate multiplier.
- Implementation:shipped in N.2.9.5
- Release:stable release that includes all 2.9.5 alpha changes
- Activation:no activation height; takes effect when each node upgrades
Sources: ArweaveTeam (GitHub) — Arweave 2.9.5 release notes (external site) · ArweaveTeam (GitHub) — arweave releases (external site)
Topics your AI can explain
Your AI can explain these topics for Arweave through the connection, with sources.
Known gaps
What this profile's review did not establish:
- The docs give no hardware sizing for a non-mining validator beyond the SSD data directory, so whether a home machine keeps up with validation and the VDF was not verified.
- How concentrated mining is (share of blocks by miner, pool or coordinated cluster) was not measured.
- Who runs the public VDF servers the docs list, and how many miners depend on them or on the Digital History Association's trusted peers, was not verified beyond the docs' own statements.
- The maximum fork-recovery depth comes from a node-code constant; its effect during a deep fork was not tested, and other clients were not reviewed.
- No light client that checks block index proofs was found; the search covered the Arweave docs and whitepaper only.
- Swap designs that keep every lock on the other chain, and wallet recovery features such as the social recovery Wander advertises, were not reviewed for AR, so the swap and recovery rows stay unknown rather than not present.
- Whether the 2.9.6 soft fork is enforced by most miners was not verified; only the network height passing the activation height was observed.
- The endowment's current size and the protocol's current price per byte were not read.
- The early yellow paper gives a maximum of 66 million AR, and the node code's total supply constant is about 66 million AR, but the same code notes that further issuance may start if the endowment pool runs empty; when or whether that happens was not modelled.
- Wrapped AR on AO and the bridges that issue it were not verified from a bridge operator's primary documentation.
- AMMs, stablecoins and other tokens that live in AO processes were not reviewed; AO is treated as a separate system, though Arweave's docs count it among the ecosystem's compute layers.
- AO and bundling services are described in notes rather than as scaling layers, because neither fits the rollup, channel or sidechain families.
- Ledger support status is unclear: Ledger's marketing page is generic while its currency list lacks Arweave and its device app is unapproved.
- Keystone's Wander guide carries no date, and the Keystone and Wander signing path was not tested hands-on.
- ArScan's page carries ViewBlock's links and assets, but no statement naming its operator was found; no independent Arweave block explorer was confirmed.
- Who operates the turbo-gateway.com GraphQL endpoint was not stated on the page that lists it.
Sources
Oldest dated source check: Source checked Sep 29, 2026 Next check due Oct 29, 2026.
- Protocol (external site)
- ECDSA Keys (external site)
- Motivation (external site)
- Account Tree (external site)
- How Mining Works (external site)
- Node Types (external site)
- Hardware Guide (external site)
- VDF (external site)
- Transaction Blacklist (external site)
- Syncing and Packing (external site)
- Trusted Peers (external site)
- HTTP API (external site)
- Wallet FAQ (external site)
- Arweave: The Permanent Information Storage Protocol (Draft 17, December 2023) (external site)
- Efficient generation and dispersion of entropy during data preparation for Arweave (Draft-5, December 2024) (external site)
- Arweave: A Protocol for Economically Sustainable Information Permanence (yellow paper, Draft-1) (external site)
- apps/arweave/include/ar.hrl (master): protocol and node constants (external site)
- apps/arweave/src/ar_block_cache.erl (master): cumulative difficulty and checkpoint depth (external site)
- Arweave 2.6.0 release notes (VDF, pricing transition, delayed rewards) (external site)
- apps/arweave/include/ar_pricing.hrl (master): replication target, assumed price decay, total supply and endowment latch (external site)
- Arweave 2.9.1 release notes (2.9 hard fork, replica.2.9 format) (external site)
- Arweave 2.9.6 release notes (format 1 soft fork at height 2,000,000) (external site)
- ANS-104: Bundled Data v2.0 - Binary Serialization (external site)
- AO Cookbook: Concepts & Migration (external site)
- AO Cookbook: How ao messaging works (external site)
- The Dock Isn't the Lake (arweave.net operator change, published 2026-03-05) (external site)
- Find Data (via GraphQL) (external site)
- Arweave Explorer (external site)
- ArScan - Arweave Explorer (external site)
- Autonomous Finance (earlier AO products, including Botega, listed as unmaintained) (external site)
- USDC contract addresses (external site)
- Supported protocols (external site)
- Arweave Wallet - Buy, manage & hold your AR (external site)
- @ledgerhq/cryptoassets 13.56.0 currency list (external site)
- app-arweave: Ledger Arweave app (README) (external site)
- Arweave wallet (external site)
- Wander (Extension) | Keystone (external site)
- Wander: wallet for Arweave and AO (formerly ArConnect) (external site)
- ArweaveWebWallet: the arweave.app source files (external site)
- arweave.net /info network read (height 2,011,130 on 2026-09-29) (external site)
Report an error
Found something wrong or out of date? Reporting is free, and every chain goes through the same process. Published corrections appear in the public log.
Funding disclosure
Funding disclosures appear here when an upkeep arrangement exists, in the form “Upkeep funded by [name]; verdicts unaffected.” The funder ledger has not been published yet. Neutrality & funding