Sia v2 explained — the full guide.

Sia v2 is the most significant protocol upgrade Sia has had since launch in 2015. Its headline change is that the network adopted Utreexo-style accumulator proofs, meaning that full consensus state no longer needs to be stored by every client — light clients can verify balances with cryptographic proofs against a compact accumulator, without needing the full UTXO set. This made lite-mode wallets viable for the first time (Sia Wallet's default 60-second setup is a direct downstream benefit). It also broke Sia-UI permanently — Sia-UI's transaction format isn't valid under v2 rules — and it introduced a new 12-word BIP39 seed format that replaced the legacy 28-word seeds. The fork activated at block 526,000 on June 6, 2025. A cleanup fork, v2 Final Cut, finalised the rules at block 552,100 on December 2, 2025. Everything since has run on clean v2 consensus.

By 2023, the Sia UTXO set was large enough that running a full node was no longer a weekend project on a laptop. Initial block download could take most of a day on a consumer internet connection, and full validation required hundreds of GB of disk. For a proof-of-work network whose value proposition includes end-user verifiability, that's a problem — the more painful it is to run a node, the fewer people run them, and the more the network trusts a shrinking set of operators. At the same time, the Sia Foundation's work on renterd, hostd, and walletd (the 2024 modular split) was running into friction with legacy consensus assumptions. Transaction construction logic that made sense for a combined wallet-renter-host program didn't make sense for three independent applications. v2 was the coordinated protocol response to both pressures.

Utreexo is a proof system that compresses the state a blockchain needs to store. In a traditional UTXO-based blockchain (Bitcoin, pre-v2 Sia), every full node keeps a copy of every unspent transaction output — the set of spendable coins. That set grows forever. Utreexo replaces 'keep the full set' with 'keep a small accumulator that can verify membership.' When you want to spend a coin, you provide both the coin's details and a small cryptographic proof that the coin is in the accumulator. The network verifies the proof and updates the accumulator. Net effect: full nodes need to store only the accumulator (a few kilobytes), not the full UTXO set (hundreds of GB). Light clients can download the accumulator and verify balances against it with similarly small proofs. Sia's v2 implementation adopts the Utreexo approach with Sia-specific modifications for storage contracts, which have more structure than plain UTXOs.

Four big changes. First, the accumulator — the protocol now maintains a state commitment structure that replaces the need for full UTXO storage. Second, the transaction format — v2 transactions carry Utreexo proofs for each input they spend, proving the input's inclusion without referencing the full state. Third, the seed format — the legacy 28-word (and rare 29-word) Sia seeds were replaced with standard 12-word BIP39 seeds, aligning Sia with the broader cryptocurrency ecosystem and making BIP39-compatible hardware wallets (Ledger, Trezor) first-class citizens. Fourth, the storage contract format — Sia's core value proposition is on-chain enforceable storage contracts, and v2 cleaned up their encoding to fit the new proof system. Each of these is backward-incompatible with v1, which is why v2 required a coordinated hardfork.

The v2 fork activated at block 526,000, which the Sia network reached on June 6, 2025. The activation block was fixed well in advance to give ecosystem tools (exchanges, block explorers, wallets) time to upgrade. The transition was clean: miners upgraded their nodes, The Sia Foundation's core infrastructure was v2-ready, most exchanges paused Siacoin deposits and withdrawals around the fork window and resumed after confirming their own integration. There was no chain split — every node that followed the upgrade was on v2, and legacy v1 nodes simply stopped receiving valid blocks. Wallets running old Sia-UI code started reporting 'synchronised' status to a network that had moved on. That's the moment when Sia-UI effectively stopped working — not because it crashed, but because the rules under its feet had changed.

The June activation switched the network to v2 rules, but the codebase still carried v1 backward-compatibility scaffolding — code paths that handled both v1-style and v2-style data to ease the transition. Six months of running pure v2 in production was enough to confirm the upgrade worked as de. The v2 Final Cut cleanup fork, activated at block 552,100 on December 2, 2025, removed the v1 scaffolding. This was a smaller and less disruptive fork than v2 itself — it affected internal code organisation more than observable network behaviour — but it finalised the v2 consensus rules. After Final Cut, the Sia network is running clean v2 consensus without legacy fallbacks. This is the state of the network as Sia Wallet v2.12.0 launches in April 2026.

If you hold Siacoin, v2 affects you in two ways. First, lite-mode wallets are now real — you no longer need a full-node sync to hold or transact Siacoin, which is why Sia Wallet starts in under 60 seconds rather than over a weekend. Second, if your SC was sitting in a Sia-UI wallet, you need to migrate. Your 28-word legacy seed doesn't work directly in post-v2 wallets, but it still derives your legacy address. Use Sia Wallet's built-in migration assistant (or walletd's command-line equivalent) to generate a new 12-word BIP39 seed, transfer your balance with a single on-chain transaction, and continue. Your coins didn't go anywhere — the address they're at didn't change until you moved them. This isn't a loss event; it's a housekeeping event.

Hosts felt the biggest operational impact. Storage contracts — the core on-chain primitive that pays hosts for storing renter data — had their encoding updated in v2. hostd (the foundation's host software) was v2-ready at activation; third-party host operators running older siad-based host software had to migrate. Hosts who missed the migration window risked their existing contracts either failing to pay out or being marked invalid by the new consensus. By December's Final Cut, the host population was effectively fully migrated. Ongoing operations after v2 are more efficient — proofs of storage use the new accumulator structure, which reduces state requirements for hosts storing many contracts simultaneously.

Renters interact with Sia by forming storage contracts with hosts and uploading file data. v2 mostly invisibly improved this experience. renterd handled the v2 transition; existing contracts continued through their original terms, and new contracts use v2 formats. For a renter storing data on Sia, v2 means slightly more efficient contract negotiation, slightly lower data structure overhead, and a simpler integration path for third-party tools because the consensus layer is more modular. Practically, if you use Sia for storage, v2 was a behind-the-scenes improvement rather than a user-visible change.

v2 reshuffled the Sia wallet ecosystem. Sia-UI, the long-standing consumer desktop wallet, was retired. walletd became the foundation's reference wallet but is developer-first. SiaCentral adapted its web wallet for v2 but kept the browser-flag Ledger situation. Atomic Wallet followed the v2 migration but remained multi-coin and closed-source. Third-party mobile apps either stopped working (for those running pre-v2 consensus) or never made the transition. Sia Wallet was launched to fill the gap: a consumer-grade, native-desktop, open-source, MIT-licensed fork of walletd with the consumer features v2 made feasible (lite mode), the migration tools the transition required (the 28-word to 12-word assistant), and the hardware-wallet polish the post-v2 network deserves (native Ledger over USB and Bluetooth). That's the state of Sia wallets in April 2026. The protocol is stable, the fork is finalised, the tools are caught up. If you've been waiting to self-custody your Siacoin, the waiting part is over.