Which mining pool fits your farm size: five miner profiles

"Pick the pool with the lowest fee" is equally useless advice for someone running a single S21 in a garage and for a site running tens of petahashes. Their bottlenecks are nothing alike. The first one loses money to payout thresholds and the power bill. The second one watches a fee percentage turn into a five-figure annual number.

Five profiles by fleet size follow. For each one: what actually decides the outcome, which of the pools we verified fit, and why.

What changes as the fleet grows

Expected income per terahash is the same in any pool. It comes from your share of network hashrate, and the pool keeps its cut of that same number. What changes is which mistake costs you more. On a small fleet a wrong threshold costs months of waiting and a wrong fee costs pennies. Above a petahash it inverts.

Per the arithmetic in our August 2026 pool comparison, the spread between a 1% pool and a 4% pool at 100 TH/s is 9 cents a day, roughly $33 a year. At 1 PH/s it is $0.94 a day, roughly $343 a year. Scale from there.

Profile 1. One ASIC at home: fee or payout threshold?

The threshold. On a single 100 TH/s machine, a 0.005 BTC threshold at AntPool or F2Pool fills in about 107 days, a 0.001 BTC threshold in about 21 days, and Foundry USA's 0.01 BTC in about 215 days. Over those same three months, the fee difference between the cheapest and priciest pool amounts to a few dollars.

What to check:

1. Payout threshold. It decides whether you see money in three weeks or in half a year.
2. Stratum stability and readable worker stats. One hour of downtime costs more than the whole annual fee spread.
3. Support in a language you speak, because you are the one debugging a stalled worker.
4. Fee. Last, honestly.

Who fits, from the verified set:

NiceHash needs its own line. Its 0.00001 BTC threshold pays out roughly every three weeks even at 1 TH/s. It is a hashpower marketplace though, not a classic pool: a buyer pays you at an RTPPS rate tied to demand, so over time the actual payout drifts from mining math more than an FPPS pool does.

Bitaxe class devices at 1 TH/s are a separate story. A 0.005 BTC threshold there takes about 10,700 days, close to 29 years. The pool technically works. You will never see a payout.

Profile 2. Five to ten ASICs: when the payout scheme starts to matter

At this size the threshold stops being a problem. Even 0.005 BTC clears in a few days. What replaces it is the payout scheme and reliability. FPPS and PPS+ give you a flat daily number. PPLNS hands you the pool's luck variance, and that shows up on the monthly total.

A hashrate reference point: on current hardware like the Antminer S21 XP Hyd at 473 TH/s, five machines run around 2.4 PH/s. The 1 PH/s mark that most advice is built around sits at roughly two or three of them. Older hardware, recalculate from your own model's rated hashrate.

What decides:

1. Scheme family. FPPS and PPS+ pass through the transaction fee share, worth an extra 0.70% over the base subsidy on the 2026-08-14 network snapshot.
2. Uptime and reject rate. A couple of percent of rejects costs more than the entire fee spread across the market.
3. Latency to the stratum server, especially on routes crossing several countries.
4. Fee, which at 1 PH/s already means about $343 a year between 1% and 4%.

Who fits: Trustpool (1%, PPS+) and Luxor (2.5%, FPPS) as the cheapest publicly disclosed rates. ViaBTC (4%, PPS+ or PPLNS, 0.001 BTC threshold) and F2Pool (4%, FPPS or PPLNS, 0.005 BTC threshold) if pool age and size matter more to you than three percent. AntPool lets you pick between FPPS, PPS and PPLNS at a 0.005 BTC threshold.

Run the scheme difference on your own numbers in the mining profitability calculator; the mechanics of each scheme are laid out in our payout scheme reference.

Profile 3. Fifty to a hundred ASICs: is switching pools worth the percentage?

Yes. A hundred S21 XP Hyd class machines run around 47 PH/s, and three percent of that turns into real money over a year. This is the scale where firmware edits, worker renaming and new payout addresses pay for themselves.

Before you move, price the move itself. Leaving a PPLNS pool costs you the tail of shares still sitting in the window unpaid: the new pool credits nothing for them, and the old one closes the window without you. Add the downtime while workers get reconfigured, plus the ramp before hashrate settles on the new stratum. What that adds up to in satoshis per 100 TH is a calculation we will publish as its own article. We are not putting a number here without measuring it ourselves.

What decides:

1. Fee. Here it finally takes first place.
2. Uptime and any SLA. An hour of downtime on 47 PH/s is a visible number.
3. Reject rate on your specific route. Only your own hardware can measure it; no ranking does that for you.
4. Reporting and exports for accounting, since every payout is an income event.

Who fits: Trustpool and Luxor on published rates, F2Pool, ViaBTC and AntPool as large pools with long histories. This is also the point where writing to Foundry USA and Neopool makes sense. Neither publishes a percentage, Foundry states tiered terms, and at this volume a negotiated rate can beat any public tariff.

Foundry USA's 0.01 BTC threshold, the one that rules out home scale, clears in under a day at 47 PH/s and stops being an argument.

Profile 4. Farms at 1 PH/s and above: why public tariffs stop applying

Past a certain volume the fee table becomes an opening position, not a price list. Foundry USA and Neopool publish no single number at all. That is ordinary practice for large operations, and it also means you cannot compare them against the rest on paper.

There is no published minimum volume at which either pool starts the conversation, either. Terms go through an application, and Foundry USA adds KYC on top of it. The only way to learn your rate is to write to them with your hashrate.

What to ask for in negotiation:

1. The rate, and the base it applies to: block subsidy only, or subsidy plus transaction fees.
2. Uptime guarantees and what compensation looks like during an outage.
3. Withdrawal terms: threshold, frequency, and who pays the on-chain fee.
4. Access to raw share data, so you can reconcile credited amounts against expectation instead of trusting a dashboard.
5. Termination terms and how fast the remaining balance leaves the account.

For a small miner an undisclosed fee is a drawback, because there is nothing to verify against. For a megawatt site it works the other way: the public tariff is usually the worse of the two.

Profile 5. The solo enthusiast: what a whole block is worth waiting for

Solo is not an income strategy. It is a lottery ticket bought by giving up steady payouts. You either take the full block reward or you take nothing for months and years. Expectation is the same as in a pool, minus the fee.

What to understand before starting:

1. Time to find a block scales off your share of network hashrate. Against 933.99 EH/s and a home-scale farm, that horizon is very long.
2. There are no interim payouts, so you need reserves to cover electricity from another pocket.
3. The payout threshold is irrelevant in solo, since payment arrives as one lump when a block lands.
4. The solo fee percentage barely enters the decision. Willingness to see nothing arrive for years does.

Public solo services as of 2026-08-20:

Ocean does not offer a solo mode in the usual sense. What it offers is DATUM Gateway, a protocol for genuine solo mining through your own full node, with no pool in the middle at all.

We covered the solo versus pool fork for a small farm in mining payback: solo versus pool, and the SOLO scheme itself, alongside TIDES, in the payout scheme reference.

Profiles at a glance

What this breakdown does not cover

No ranking by network share or by "luck": those numbers move daily and are better read live on each pool's own stats page. No altcoin pools. No referral discounts, which get revised often. And no electricity tariff, which at home scale decides the whole outcome by itself: at $0.12 per kWh a home 100 TH/s loses money in every pool listed here, $3.01 earned against $3.46 spent.

Full terms for all twelve pools sit in our breakdown of twelve Bitcoin pools, and the live ranking recalculates net income for your hashrate and power price in the pool ranking.

FAQ

Which pool should I choose with a single ASIC?

The one whose payout threshold you can clear in a sane amount of time. At 100 TH/s a 0.001 BTC threshold fills in about 21 days and 0.005 BTC in about 107. From the verified set: EMCD and Promminer at 0.0001 BTC, Trustpool and Luxor at 0.001 BTC.

At what farm size does the pool fee start to matter?

Around a petahash. At 100 TH/s the gap between 1% and 4% is about $33 a year; at 1 PH/s about $343, growing linearly from there. Below a petahash the electricity bill outweighs the pool choice entirely.

Is it worth moving to a cheaper pool?

At home scale, no: switching for 9 cents a day does not repay an evening of reconfiguration. From several petahashes up, yes, but price in the downtime and the unpaid tail of shares left in the PPLNS window.

Where can I mine solo without registering?

As of 2026-08-20 two services accept a connection where the username is simply your BTC address: solo.ckpool.org, charging 2% taken only out of a block you find, and Braiins Solo at 0.5%. Neither requires an account.

Why do Foundry USA and Neopool show no fee?

Neither publishes a single rate, and Foundry USA's is tiered and negotiated through an application with KYC. For a large site a negotiated rate may well beat the public one. For a home miner it means there is nothing to check your earnings against.

Does NiceHash suit a home miner?

Its 0.00001 BTC threshold is the lowest in the verified set, and payouts arrive regularly even on weak hardware. It is a hashpower marketplace though: a buyer pays you at a demand-linked rate, so real results drift from the calculation more than an FPPS pool would.

---

Pool fees, schemes and thresholds verified on 2026-08-18 against official pages; solo services on 2026-08-20. Network parameters are from the 2026-08-14 snapshot. POOL BTC compares the published terms of other people's pools and does not accept hashrate. Figures are indicative and not financial advice.