Stratum V2 explained: what changes for small miners
Stratum is the language a miner and a pool use to share work and report results, and for over a decade that language has been Stratum V1: the pool builds a block template, hands the miner a puzzle, and the miner hashes without ever seeing which transactions went into the block. Stratum V2 is the successor protocol, and the coverage of it tends to arrive as a single headline, miners take back control of block construction. For a small miner that headline hides the useful structure. V2 is two separate upgrades with two separate price tags. The first, an encrypted and more efficient connection, is available today on hardware already owned. The second, building your own block templates through Job Declaration, costs a Bitcoin node of your own and a pool that accepts declared jobs. Nothing here is financial advice; it is the protocol mechanics, sorted by tier, plus what the upgrade pointedly does not change.
What V1 does, and its three weaknesses
The original Stratum shipped in 2012, was built for a hobbyist era that mining outgrew, and still carries most of the network. Its pooled-mining loop is simple and it works: the pool sends each miner a job, a block template to hash against, the miner submits shares, near-miss solutions that prove work, and when a solution clears the network target the pool splits the reward by contribution. Three weaknesses are structural, not incidental.
Plaintext. V1 sends its messages as unencrypted JSON text. Anyone on the network path can read pool URLs, worker names, and hashrate data. That exposure enables two attacks: a man-in-the-middle can swap pool credentials and steal hashrate, redirecting work to another pool without the miner noticing, and an internet provider can watch the traffic and estimate what a miner earns.
Pool-built templates. Under V1 the pool builds every block template and decides which transactions go in. The miner has no say in that selection and no way to object if the pool drops certain transactions. With a handful of pools directing most blocks, the power to choose what enters the chain concentrates in a few hands, the censorship concern that motivates most V2 advocacy.
Weight. A text-based JSON format wastes bandwidth and adds latency on every message. Every wasted millisecond risks stale work: shares computed against a block the network has already moved past are worth nothing. Remote sites and bandwidth-capped operations feel this cost first, but it is paid by every connected machine.
Tier one: encryption and efficiency on the hardware already owned
Stratum V2 is a binary protocol that succeeds the V1 JSON-RPC interface between miners (or their proxies) and pools. Every V2 connection encrypts and authenticates its channel, the specification uses the Noise framework, and the encryption negotiates on its own, with no manual certificate setup for the miner. Credential swapping and hashrate theft close off, because an intermediary can no longer read or rewrite the traffic, and share data is hidden from providers and routing attackers. The binary format cuts miner bandwidth by about 70% and pool bandwidth by about 60% relative to V1, and lighter messages travel faster: miners receive new jobs sooner, waste less work on stale shares, and remote operations pay less for data.
The part the headlines skip: none of this requires new hardware, and none of it requires declaring anything. A translation proxy lets a V1 ASIC speak to a V2 pool, the machine talks V1 on the local network, the proxy talks encrypted V2 to the pool, so adoption can be staged instead of forced through a firmware upgrade. Pools that do not offer Job Declaration can still run encrypted V2 with pool-chosen jobs, and that mode is a complete, legitimate use of the protocol: the connection is protected and cheaper, while the pool keeps building templates exactly as before. For most small miners, tier one is the whole realistic upgrade today, and it is worth having on its own terms.
Tier two: Job Declaration and building a template
Job Declaration is the mechanism that moves transaction selection. A miner running the declaration client builds its own block template and proposes it to the pool; the pool validates the template, and the miner's version goes forward while the pool continues to coordinate hashrate and pay rewards. The template has to come from somewhere: in practice the miner runs a Bitcoin node and picks transactions from its own mempool. V2's sub-protocols cover exactly this, separating template distribution from a Bitcoin node and job declaration between the miner-side client and the pool-side server. That node is the real cost of tier two. It is a machine to buy or dedicate, keep synced, and keep online, in exchange for a say in which transactions the miner's hashrate works on.
The milestone that proved the tier works in production arrived on June 25, 2026: block 955,318, carried by the DMND pool for GoMining, is the first known Bitcoin block produced using Stratum V2, with the miner using Job Declaration to select its own transactions and build its own template, including transactions from its own payments product. One block is a proof, not a prevalence statistic. It demonstrates that a miner can keep the revenue-smoothing of pooled mining while taking back template construction; it does not demonstrate that most pools, firmware, or miners are there yet.
The two tiers, side by side
| Stratum V1 (today's default) | V2 via proxy or supporting firmware, pool-chosen jobs | V2 with Job Declaration | |
|---|---|---|---|
| Channel | Plaintext JSON | Encrypted, authenticated, binary | Encrypted, authenticated, binary |
| Bandwidth vs V1 | Baseline | About 70% lower (miner side) | About 70% lower (miner side) |
| Who builds the template | The pool | The pool | The miner, from a local node's mempool |
| Hashrate hijacking | Possible via man-in-the-middle | Blocked by encryption | Blocked by encryption |
| Extra kit required | None | A translation proxy, or firmware with V2 support | All of tier one, plus a synced Bitcoin node and a pool that accepts declared jobs |
| Payout trust | Pool must be trusted to pay | Unchanged | Unchanged |
What Stratum V2 does not change
Three non-changes deserve the same prominence as the features. First, payout trust: job declaration does not remove the need to trust a pool for payouts. The pool still coordinates the split of rewards; V2 redistributes transaction selection, not the money path. Second, economics: V2 does not raise block rewards, and it changes neither pool fees nor payout schemes. PPS, FPPS, PPLNS and the rest are pool contracts that sit above the wire protocol, untouched by it. Third, the profit figure: deployment tests are quoted as showing V2 can lift profitability by up to 7.4%, from faster template delivery, fewer stale shares, and better fee capture. Read that number the way its own source presents it, as a ceiling that varies by setup and pool, from specific tests, not a protocol guarantee and not a raise every adopter banks. A miner whose connection was never hijacked and whose stale rate is already low has less to recover.
Adoption, honestly dated
As documented in mid-2026 sources: V1 still works and still carries most of the network, and V2 is an optional upgrade, not a requirement. Braiins Pool and DMND run V2 in production. On May 7, 2026, seven large operators. ANTPOOL, Block, DMND, F2Pool, Foundry, MARA Foundation, and SpiderPool, joined the Stratum V2 Working Group, which signals industry direction; working-group membership is not the same claim as production support on a given pool's public endpoint, and network-wide V2 hashrate share remains a minority, growing. Firmware is the other gate: stock firmware on the dominant Antminer line does not support Stratum V2; third-party firmware (Braiins OS+ is documented as bringing native V2 to most S19 variants plus the S21 and S21 XP) is the main native route, one vendor's own firmware supports V2 on its own ASICs, and open-source hobby hardware supports it natively. Any stock-firmware machine can still reach a V2 pool through the translation proxy, which is precisely why tier one is the practical question for most small miners and tier two is a deliberate project.
Should a small miner switch? A decision procedure
- Name the goal. Cheaper, protected traffic, tier one. A say in template contents, tier two. If the goal is neither, V1 is not broken; stay and spend the effort elsewhere.
- Check the pool first. Does the current pool run a V2 endpoint in production, and does it accept declared jobs or only pool-chosen jobs? Read the pool's own connection documentation, not a comparison article, support claims date quickly.
- Check the firmware path. Native V2 firmware available for the machines in question, or is a translation proxy the route? A proxy is one small always-on device on the local network; budget for it as infrastructure, not an afterthought.
- For tier two only: price the node. A synced Bitcoin node, kept online, plus the time to run it. That ongoing cost is the subscription fee for template control, decide whether the principle or the fee-selection control is worth it at one-machine scale.
- Switch one machine, measure, then decide. Move a single miner, and compare accepted-share rate and stale-share rate against the same machine's V1 baseline over at least a full difficulty epoch. Protocol benefits that do not show up in the share log are not benefits yet.
What success looks like: after the switch, the pool dashboard shows the expected hashrate with a stale-share rate no worse than the V1 baseline, the connection is to the pool's documented V2 endpoint, and, for tier two, the pool confirms declared templates are being accepted rather than silently replaced with pool-chosen jobs.
FAQ
Does Stratum V2 change how much a miner earns? It does not raise block rewards, change pool fees, or change payout schemes. Documented test figures of up to 7.4% higher profit are ceilings from specific deployments, fewer stale shares, no hijacked hashrate, better fee capture, and a setup without those losses has proportionally less to gain.
Is a Bitcoin node required to use Stratum V2? No. A node is required to build templates for Job Declaration. Encrypted V2 with pool-chosen jobs, including through a translation proxy on existing hardware, needs no node at all.
Does Job Declaration make pooled mining trustless? No. It moves transaction selection toward miners who run nodes; the pool still coordinates payouts and must be trusted for them. Template control and payout trust are separate questions, and V2 answers only the first.
Will V1 stop working? Nothing in the current record says so: V1 still carries most of the network and V2 adoption is staged and optional, with proxies existing precisely so V1 hardware is not stranded.