Four Pools Build 70% of Bitcoin's Blocks. The Fix Is Running on 3%

Four companies decide the contents of roughly 70% of all Bitcoin blocks. In May 2026, seven pools representing nearly 75% of global hashrate publicly agreed to fix it. Four months later, the fix is running on about 3–5% of the network. That gap between announcement and deployment is the most under-reported story in mining.

Here is the state of play. Foundry USA commands roughly 37% of network hashrate. AntPool sits near 14%. Together those two produce more than half of all Bitcoin blocks. Add F2Pool and SpiderPool and you have four operators deciding what goes into about 70% of blocks. Research has also indicated that AntPool functions as a “pool of pools” — several ostensibly independent pools sharing identical block templates and transaction prioritisation, which suggests they are proxies rather than genuinely separate operators.

None of this is a secret and none of it is new. What is new is that the industry has agreed on the remedy and has not yet shipped it.

First, the stakes: what a pool actually does with your hashrate

Most miners think of a pool as a payout mechanism — you point your machine at it, you get paid smoothly instead of waiting years for a block. That is true and it is the smaller half of the story.

Under the long-dominant Stratum V1 protocol, the pool does not just distribute rewards. It builds the block. The pool decides which transactions are included, in what order, and which are excluded. Your machine is handed a template and told to hash it. You supply the electricity and the silicon; someone else exercises the judgement.

For most miners most of the time, this is invisible and harmless. It becomes visible in exactly the situations where it matters most: if a pool operator came under legal pressure to censor certain transactions, or chose to prioritise its own order flow, the hashrate pointed at that pool would enforce the decision without its owners ever being consulted.

Stratum V2, and why everyone agreed to it

Stratum V2 is the protocol upgrade that addresses this directly. Its critical component — Job Declaration — lets the individual miner, rather than the pool operator, choose which transactions go into the blocks their hashrate is working on. The pool still handles payouts and variance smoothing. It stops being the sole author of the network’s contents.

It also brings real operational improvements: encrypted, authenticated connections that close off hashrate-hijacking attacks, and a more efficient binary protocol that reduces bandwidth and stale shares. Those are the reasons a pool operator would adopt it even if they were indifferent to decentralisation.

In May 2026, seven of the largest pools — nearly 75% of global hashrate — joined the Stratum V2 Working Group and backed the open standard for block construction. On paper, that settles the debate.

The gap between agreement and deployment

As of mid-2026, only Braiins Pool and DMND run Stratum V2 with Job Declaration in production, representing roughly 3–5% of hashrate. Everything else is commitment rather than deployment.

There are honest reasons for the lag. Job Declaration requires miners to run their own Bitcoin node to build templates, which is a meaningful operational change for a large fleet. Firmware support has to reach the machines. Pools have to run both protocols in parallel through a long transition. None of that is trivial and none of it is fast.

There is also a less charitable reading available, and it would be dishonest not to state it: the entities being asked to give up template control are the entities that benefit from holding it. We are not accusing anyone of bad faith. We are saying that “joined a working group” and “shipped it to production” are different claims, and only one of them changes the concentration figure.

What you can actually do about it

This is a structural problem, but it is not one where an individual miner is powerless. Three things are genuinely in your control.

1. Choose your pool deliberately, and know you can change it

Pool choice is portable. Pointing your hashrate somewhere else is a configuration change, not a contract renegotiation. If decentralisation matters to you, a pool running Stratum V2 with Job Declaration in production today is a vote you can cast this afternoon — and it costs you nothing but a few minutes.

2. Keep the pool decision separate from the hosting decision

These get conflated constantly, and the conflation always favours the provider. Your host’s job is to deliver cheap, reliable power and keep machines running. Your pool’s job is to build templates and distribute rewards. A hosting arrangement that permanently ties you to one pool has taken a decision away from you that it did not need to take.

3. Judge the network by deployment figures, not announcements

Watch the share of hashrate actually running Job Declaration in production. That number — not the working-group membership list — is the honest measure of whether this problem is being solved.

Where hosting fits, honestly

We will be straightforward about this, because the temptation to overclaim is obvious and we would rather you trusted the rest of the article.

Hosting does not solve pool centralisation. No hosting provider can. What a host determines is whether your machine runs at all, and what each hour of running costs you — and those remain the variables that decide whether you are still mining when this debate resolves.

What OneMiners does provide is relevant on the control side. Remote machine access means you can see and restart your units from the dashboard rather than filing a ticket and waiting. Our AI Smart Mining feature is, mechanically, dynamic mining-pool switching driven by real-time market data — a reminder that pool assignment is a live, changeable parameter rather than something fixed at signup.

And the economics stay what they always were:

• Published rates, every site, every tier. Nigeria at $0.0480/kWh premium and $0.0364/kWh on a seven-year prepay. The prepaid ladder is a formula off the online-ordered rate — 4% at one year, 12% at three, 30% at seven — reproducible on a calculator. Full card at hosting centers.
• Seven-year fixed electricity contracts, in a year when utilities filed for $18.6 billion in rate increases.
• Price includes all management services — no performance fee taken out of your output.
• 95% guaranteed minimum uptime with compensation, and a seven-year warranty on hosted machines covering defects, excluding normal wear and consumable items.
• Fifteen locations across grid, hydro, gas, wind and solar.

Typical industrial hosting in 2026 runs $0.065–$0.08/kWh all-in, and a number of well-known providers publish no rate at all or take a revenue share instead — which, incidentally, is its own form of handing someone else a claim on your output.

Final thoughts

Bitcoin’s security rests on the assumption that no small group decides what the network does. Today four companies build most of the blocks, the protocol upgrade that would fix it has near-universal stated support, and it is running on a twentieth of the network.

That is not a crisis. It is a deadline that keeps getting extended, and the only thing that moves it is miners treating pool choice as a decision rather than a default. It takes minutes, and it is one of the very few things in mining where an individual operator’s choice registers directly in a network-level statistic.

Then go back to the part that decides whether you are still here to vote next year: efficiency, and the price of a kilowatt-hour.

Pools are a choice. Protocols are a negotiation. Electricity is physics.