How to Choose a Mining Pool for a Bitcoin ASIC

For a Bitcoin ASIC miner, the pool decision comes down to one thing most guides skip: whether the pool pays you a share of transaction fees or only the block subsidy. That difference is usually worth more than the gap between a 1% and a 2% pool fee, and it is the first thing to check. Everything else — pool size, server location, reputation — matters less than it did in the GPU era.

This page is about pools for SHA-256 machines specifically. If you have not bought hardware yet, start with what you actually need to start mining.

Why you need a pool at all

The network runs at roughly 866 EH/s as of 27 July 2026. A single Antminer S19 at 95 TH/s is about one nine-millionth of that, which works out to finding a block roughly once every 170 years on average. You would receive 3.125 BTC in one lump, eventually — or, far more likely, nothing at all before the machine is scrap.

A pool aggregates thousands of machines, collects whatever blocks the combined hashrate finds, and pays each participant in proportion to the work they submitted. You give up the lottery ticket and take a wage instead. For anything short of a very large fleet, that is not really a choice.

Payout models, in the order that matters

The acronyms hide a simple question: what exactly are you being paid for?

FPPS — full pay per share

You are paid a fixed amount per valid share, covering both the block subsidy and an averaged share of transaction fees. The pool absorbs all variance. This is the default worth choosing for most miners, because transaction fees are real money that a plain PPS pool keeps.

PPS — pay per share

Same steady payment per share, but calculated on the block subsidy only. Transaction fees go to the pool. Payouts are perfectly smooth and slightly smaller than FPPS on the same hashrate. Compare the total expected payout, not the headline fee — a 1% FPPS pool commonly beats a 0% PPS pool.

PPLNS — pay per last N shares

You are paid out of blocks the pool actually finds, weighted by your recent share history. Over a long enough period it can pay slightly more than PPS because you are not paying the pool to absorb variance. In exchange you carry that variance yourself, and you are penalised for switching pools — leaving forfeits the shares still in the window. Reasonable for a stable, long-running operation; poor for someone experimenting.

Solo

The pool coordinates work but you keep the entire block if your hardware finds one. This is a lottery, correctly priced. It is a legitimate choice if you understand you are buying a small chance of a large payout, and a bad one if you need predictable income.

Fees, and the numbers that hide behind them

Advertised fees generally run between 0% and 3%. Treat that figure as one input rather than the answer:

• Fee share vs subsidy share. Covered above, and usually larger than the fee difference.
• Minimum payout threshold. A high threshold on a small operation means your earnings sit with the pool for weeks. If you run one or two machines, this is a real consideration — it is your money held by someone else, and pools have failed.
• Withdrawal fees and network fees. Frequent small payouts can be eaten by on-chain costs. Some pools batch payments to manage this.
• Firmware dev fees. Braiins OS+ carries a 2–2.5% development fee that is waived when you mine on Braiins Pool. If you run that firmware, the arithmetic changes noticeably in one direction.

What matters less than people think

Pool size. Beyond a certain scale, extra size stops smoothing your payouts in any way you can feel — and on FPPS or PPS your payment is fixed per share regardless. Very small pools do have meaningfully higher variance on PPLNS. Between the large established pools, size is close to irrelevant to your income.

Server latency. This mattered a great deal in GPU mining, where share times were short. On SHA-256 at ASIC hashrates, stale share rates on a reasonably connected machine are a fraction of a percent. Pick a regional endpoint if one exists, then stop thinking about it. A flaky local network connection will cost you far more than a distant server.

Dashboards. Nice to have. They do not pay you.

The two settings people forget

A backup pool at a different provider. Every miner supports at least two or three pool entries, tried in order. Most people fill the second slot with another endpoint from the same pool, which protects against exactly nothing when that operator has an outage. Use a genuinely different company. A machine with no reachable pool draws full power and earns nothing, and it can do that overnight without anyone noticing.

Worker names that map to physical positions. Name each machine for where it actually sits — rack, shelf, position. When the pool dashboard shows one worker degraded, that name is the difference between walking straight to it and checking every unit in the room. Configuration details are in our Antminer setup guide.

A note on pool centralisation

A handful of pools direct a large fraction of Bitcoin’s hashrate, and the pool — not you — has historically chosen which transactions go into the blocks your machine helps find. Stratum V2 changes that by letting miners declare their own block templates, and it is worth preferring a pool that supports it. This is not going to change your monthly revenue. It is a reason to think twice before adding your hashrate to whichever pool is already the largest.

How to choose, in practice

1. Shortlist pools that pay FPPS and have operated for years without a payout failure.
2. Compare total expected payout per unit of hashrate, not advertised fee.
3. Check the minimum payout threshold against how much hashrate you actually run.
4. If you run Braiins OS+, price the waived dev fee into the comparison.
5. Configure a backup at a different provider before you walk away.
6. Prefer Stratum V2 support, and prefer not to feed the largest pool.

Then leave it alone. Pool hopping costs more in forfeited PPLNS shares and reconfiguration mistakes than it gains, and the machine you are pointing at the pool matters far more than the pool does. What decides whether you earn anything at all is efficiency and your power rate — the arithmetic for that is in the ASIC miner buying guide, and the levers that move margin are set out in what actually moves mining margin.

Hardware to point at whichever pool you choose is in used ASIC miners and the full ASIC miner inventory.

Frequently Asked Questions

What is the difference between PPS and FPPS?

PPS pays a fixed amount per share based on the block subsidy alone, and the pool keeps the transaction fees. FPPS pays for the subsidy and an averaged share of transaction fees as well. Because fees are genuine revenue, an FPPS pool charging one percent commonly pays more in total than a PPS pool charging nothing, so compare expected payout rather than the advertised fee.

Does pool size affect how much I earn?

Very little on PPS or FPPS, where you are paid per share regardless of when the pool finds blocks. Size matters on PPLNS, where a very small pool produces noticeably choppier payouts. Between the large established pools the difference in income is close to nothing, so reputation and payout terms are better selection criteria.

Why do I need a backup mining pool?

Because a machine with no reachable pool keeps drawing full power and earns nothing, potentially for hours before you notice. Miners accept several pool entries and fall through them in order. The important detail is that the backup should be a different company, not another server belonging to the same pool, since an operator outage takes all of its endpoints down together.

Can I mine Bitcoin without joining a pool?

You can, but with a single machine it is a lottery. One 95 TH/s miner against a network of roughly 866 EH/s would expect to find a block about once every 170 years. Solo mining is a reasonable choice if you understand you are buying a small chance of a whole block reward, and a poor one if you need predictable income.

Does server location matter for ASIC mining?

Much less than it did for GPU mining. Stale share rates on a well-connected ASIC are a fraction of a percent, so choosing a regional endpoint where one exists is sufficient. An unreliable local network or a failing switch will cost you considerably more revenue than a geographically distant pool server.

Should I switch pools to chase better rates?

Rarely. On PPLNS, leaving forfeits the shares still inside the payout window, and every reconfiguration is a chance to introduce a mistake. The choice of machine and the price of your electricity determine your margin far more than the choice of pool does, so it is usually better to pick a sound pool once and leave it running.