What Bitcoin Mining Pools Sell: The Business Behind the Block

Every ten minutes or so, somewhere in the world, a Bitcoin miner’s hardware grinds out the winning number and earns the right to publish the next block. At current prices that block is worth about $245,000, a 3.125 BTC subsidy plus transaction fees, with Bitcoin trading near $78,000 (CoinGecko). Here is the odd part: the miner whose machine actually found it almost never keeps the prize. A mining pool does, and it hands that miner a small, steady wage instead.

Mining pools usually get explained as a centralization problem, a story about how a few companies ended up sitting on most of Bitcoin’s compute. That story is true, and we will get to it. But it skips a simpler question that explains almost everything else: what does a pool actually sell, and how does it make money? Follow the cash and a pool stops looking like a piece of neutral infrastructure and starts looking like what it is, a financial business that underwrites other people’s luck. This is the business behind the block.

Variance is the enemy

Start with why pooling exists at all. To add a block, a miner has to grind through an astronomical number of hashes searching for one that lands below a target set by the network’s difficulty, currently around 127 trillion (Hashrate Index). The combined guessing power aimed at that target is enormous: roughly 960 exahashes per second as of mid-September 2026. A single top-tier machine, say an ASIC in the 200 terahash class, is on the order of one part in five million of that total. On any given block it has about a one-in-five-million chance of being the winner.

Stretch that over time and the problem is obvious. A 270 TH/s Antminer S21 XP, one of the fastest single machines you can buy, has roughly a one-in-24,000 chance of finding a block on any given day, which works out to one block about every 67 years on average. A hobbyist can treat that as a lottery ticket and enjoy the tiny chance of a life-changing hit. A business with payroll, power contracts and loan covenants cannot. What kills miners is not a low average return; it is variance, the distance between that average and whatever actually lands in a given week. With hashprice (the daily revenue per unit of hashrate) sitting near $38.70 per petahash per day and Bitcoin down about 38% from its October 2025 record of $126,080 (CoinGecko), margins are thin enough that a few unlucky weeks in a row can push an operation underwater. Miners will pay to make that variance go away. The interplay between hashprice, the Bitcoin price and the cost of money is the whole ballgame for a mining business, a dynamic we traced in our look at how crypto is pricing a live December after the Fed’s hike.

What a pool actually sells: insurance against luck

A pool solves variance by pooling it. Thousands of miners point their machines at one coordinator and agree to search for blocks together. Collectively they find blocks on a steady cadence, and the pool pays each participant according to the work they can prove they did, measured in shares, rather than according to whether their specific machine happened to win.

Under the scheme that now dominates industrial mining, full pay per share (FPPS), the pool pays you every single day as if you had found exactly your statistical share of blocks and fees, whether or not the pool actually got lucky that day. Read that slowly, because it is the entire business in one sentence. The pool promises a smooth, predictable payout and eats the difference itself. That is insurance. The pool is underwriting variance with its own balance sheet, and like any insurer it needs two things: enough capital to survive the unlucky stretches, and pricing that keeps it solvent across them.

Seen this way, the thing a modern pool sells is not coordination software, which is close to free and largely open source. It is settlement certainty. Hashrate Index, the mining-data arm of the pool operator Luxor, describes today’s large pools not as simple aggregators but as coordination and control hubs, liquidity venues and risk managers. Every other feature of the business, the fees, the held balances, the ownership structures, follows from that single fact: somebody has to hold the capital that turns a lumpy, random revenue stream into something that feels like a salary.

The payout menu, and who is carrying the risk

The payout scheme a pool offers is the clearest signal of who bears the variance. It is not a marketing detail; it is the risk-transfer contract, and it tells you exactly how much of a miner’s luck the operator has agreed to absorb.

Full pay per share is the institutional default precisely because it hands the miner the most certainty and keeps the most risk on the pool. Pay per last N shares does the opposite: it pays only when the pool actually finds a block, pushing variance back onto the miner in exchange for lower fees and a structure that discourages pool-hopping. PPS+ splits the difference, paying a steady subsidy but distributing fees only as real blocks land. Ocean’s TIDES pays miners non-custodially, straight out of each block’s coinbase transaction, so the pool never holds the funds. SOLO is the honest extreme: the pool coordinates the work but the miner keeps everything if they win and nothing if they do not. Fees track the risk transfer, running from roughly zero to about 4% depending on the scheme and the pool (D-Central).

Follow the money: the pool revenue stack

If the fee can fall to zero and the biggest pools still compete hard to sign miners, the fee is clearly not where the money is. A pool’s income is a stack, and the headline fee is often the smallest layer in it.

Start with the fee, the percentage skimmed off rewards. Then add the part most outsiders miss: whoever assembles the block chooses which transactions go in and in what order, and that choice is worth money whenever the mempool is busy. A wave of Runes mints, Ordinals inscriptions or a general fee spike can briefly make the fee portion of a block rival the subsidy, and the entity building the template captures the upside of ordering it well. That is the same transaction-ordering value that, in DeFi, can turn an automated market maker’s curve into an oracle to be gamed; on Bitcoin it accrues to whoever holds the template.

Then there is float. Between the moment a miner earns and the moment the pool pays out (typically once a day, above a threshold of around 0.001 BTC), the pool is holding everyone’s coins. Across thousands of miners that unpaid balance is a real treasury, and a pool gets to decide what to do with it while it sits there. Add the funnels: pools owned by exchanges route payouts into trading, custody and lending; pools owned by manufacturers sell the machines and then capture their output; pools owned by financiers treat the aggregated hashrate as collateral. Finally, pools increasingly sell forward hashrate contracts, prepaid settlement and other structured products that turn future, uncertain mining revenue into cash today. Stack all of that up and the skim off the top is the least interesting line in the ledger.

The fee war and the race to zero

Nothing illustrates the loss-leader logic better than Foundry. The largest pool in the world ran completely free from 2019 until April 2023, when, in the middle of the crisis at its parent Digital Currency Group and DCG’s lending unit Genesis, it finally introduced tiered fees pegged to a client’s hashrate over the prior quarter (Bitcoin Magazine). Even now, its effective FPPS fee for large institutional miners sits close to zero.

How does the biggest pool on the network make money charging almost nothing? By not trying to. The pool is a customer-acquisition channel for everything else in the DCG stack: financing, hardware procurement, custody, market data. A near-zero fee is a marketing budget. That posture has dragged the whole market toward the floor. AntPool and F2Pool cluster around 2.5%, ViaBTC lists 4% for PPS+ and 2% for PPLNS, Luxor runs a tiered FPPS in the sub-1% to 2% range, and Ocean’s TIDES can be zero and is 1% when a miner runs the pool’s DATUM template software (D-Central). The direction of travel is unambiguous: fee revenue is being competed away, which forces pools to earn from float, funnels and block construction instead. The percentage on the tin keeps shrinking; the business behind it does not.

Who owns the hashrate, and who owns the pools

The reason a pool can run at cost is that the hashrate it aggregates is worth more to its parent than any fee could be. Look at who actually owns the leaderboard.

Foundry USA belongs to Digital Currency Group, one of the most important balance sheets in North American mining. AntPool is Bitmain’s; the company that sells much of the world’s mining hardware also operates one of its largest pools, so the machine and the payout endpoint come from the same house. ViaBTC grew up alongside the CoinEx exchange under founder Haipo Yang, giving its miners’ payouts a natural place to land. MARA Pool is run by Marathon Digital, a Nasdaq-listed miner that built a pool mainly to serve itself. Luxor is a full-stack operation spanning firmware, a hashrate marketplace and derivatives. Binance Pool is an arm of the exchange. Braiins Pool, at the bottom of that list, is the descendant of Slush Pool, the very first Bitcoin mining pool, launched back in 2010. In almost every case the pool is not the product; it is the mouth of a funnel. That is also why concentration is so sticky. Foundry USA holds roughly 26% of network hashrate and AntPool about 19%, so the two together sit near 45%, and the top four pools control close to 70% of all blocks (Hashrate Index). It takes only three of them to cross half the network, a Nakamoto coefficient of three that has barely moved all year.

How we even know these numbers

A quick caution about every percentage in that table. Nobody measures a pool’s hashrate directly. Pool share is inferred, mostly by reading the coinbase transaction of each mined block, where pools stamp an identifying tag, and by attributing blocks to whoever they appear to have come from over a trailing window. That is an estimate built on self-reported markers; it lags reality, and it can be gamed or simply mislabeled. The same fog surrounds the network hashrate itself, which is not measured but back-calculated from how fast blocks are arriving relative to difficulty, as we explained in our piece on why nobody actually knows Bitcoin’s real hashrate.

Treat single-decimal pool shares as directionally true rather than precise, and be skeptical of any dramatic week-to-week swing. A large miner quietly moving hashrate between two pools it already uses can shift the leaderboard without changing anything real about decentralization, and a pool can flatter itself at the margins. The trend that matters, a small group of pools sitting on most of the blocks, is robust across every methodology and has held for years. So when a pool’s revenue depends on aggregating as much hashrate as possible, remember that the number it advertises is partly a marketing figure and partly a genuine measurement, and outsiders cannot always tell which is which.

The two-tier market: institutions versus everyone else

That concentration is now producing a split market. As the big pools optimize for large, compliant, institutional clients, the service they offer bifurcates. CryptoSlate, surveying the field in 2026, described a two-tier structure in which the four pools controlling more than 70% of hashrate increasingly prioritize institutional miners, with strict know-your-customer requirements, privately negotiated and undisclosed fee schedules, and tiered support that puts big accounts at the front of the line, leaving independent and mid-size miners underserved (CryptoSlate).

That gap is itself a business opportunity. Smaller-focused pools position against the giants precisely on accessibility: EMCD, for instance, advertises FPPS fees starting around 1.5% with the same terms and support regardless of account size, a pitch aimed squarely at the miners the majors have stopped courting (CryptoSlate). It is a familiar shape from traditional finance: once the incumbents chase whales, a long tail of customers becomes a market for somebody else. The uncomfortable implication for Bitcoin is that the very forces pushing pools to serve institutions, compliance, settlement certainty and scale, are the same forces pushing hashrate toward a handful of names.

Diversifying the revenue: beyond Bitcoin

A pool that has competed its Bitcoin fee down to nothing has every reason to find new revenue, and the majors are doing exactly that. In the spring of 2026, Foundry announced its first pool outside Bitcoin, an institutional-grade Zcash mining pool built for United States compliance, with KYC and AML checks, transparent payout math and formal SOC audits (CoinDesk). Foundry chief executive Mike Colyer framed the move as filling an infrastructure gap rather than a pivot. “Zcash has matured into an institutional-grade asset, but the mining infrastructure supporting it hasn’t kept pace,” he said, adding that the firm’s Bitcoin mining business remains strong and its core foundation (CoinDesk).

The strategic logic is the revenue stack again. The same client relationships, compliance apparatus and settlement machinery a pool built for Bitcoin can be resold against another proof-of-work coin, against structured products, or against custody. Luxor’s derivatives desk, Bitmain’s hardware channel and DCG’s balance sheet are all versions of one idea: the pool is the customer-acquisition layer, and the profit lives in whatever else you can attach to a captive base of miners who have already handed you their hashrate and, for part of each day, their coins.

Counterparty risk: when the pool is the risk

That last phrase, their coins, is where the insurance business turns dangerous. The float that makes a pool a treasury also makes it a counterparty, and miners are trusting it with real money that has already been earned but not yet paid out.

The cautionary tale is Poolin. In September 2022 the once-top-tier pool froze Bitcoin and Ether withdrawals, citing a liquidity crunch, after it had been deploying its float into yield strategies; it later switched miners from FPPS to PPLNS to stop underwriting variance it could no longer fund (Bitcoinist). Nothing about Bitcoin’s protocol broke. The pool, acting as an unregulated custodian and insurer at the same time, simply ran out of the capital it needed to keep its promises. There are subtler risks too. A pool can in principle withhold or discard a valid block a miner finds, skim in ways that are hard to detect given that miners cannot see the pool’s full order flow, or quietly become insolvent while still showing green dashboards.

The defense miners have reached for is the one the rest of crypto keeps rediscovering: minimize the trust. Ocean’s TIDES pays straight from the coinbase so the pool never holds the funds, the non-custodial answer to counterparty risk. It is the mining version of a lesson we drew from bridge security, where the safest bridge often turns out to be no bridge at all: the most reliable way to survive a custodian’s failure is not to have a custodian.

The most valuable asset: who builds the block

There is one asset on a pool’s books worth more than its float, and it appears on no balance sheet: control of the block template. Under the original Stratum V1 protocol from 2012, the pool, not the miner, decides which transactions go into each block. Miners simply hash whatever template the pool hands them. That means the operator, not the thousands of machines pointed at it, effectively chooses what Bitcoin confirms, and captures the fee and ordering value that comes with the choice. For a pool that is the crown jewel; for Bitcoin’s neutrality it is the single biggest point of leverage in the whole system. As GoMining chief executive Mark Zalan put it, “for years, mining pools have determined which transactions are included in Bitcoin blocks” (TFTC).

Stratum V2 is the industry’s attempt to hand that power back. Its Job Declaration mode lets an individual miner run a full node, build its own block template locally, and have the pool merely validate the coinbase and smooth the payout; the link between miner and pool is also encrypted. In May 2026 a working group representing seven pools and roughly three-quarters of network hashrate, among them Foundry, AntPool, F2Pool, SpiderPool, MARA, Block and DMND, formally committed to the standard (news.Bitcoin.com). AntPool chief executive Andy Zhou said the company was “proud to support the broader adoption of Stratum V2,” calling it an open standard for efficiency, security and decentralization (news.Bitcoin.com).

The first block whose template was declared by the miner rather than the pool was mined in June 2026 by GoMining through the V2-native pool DMND, block 955,318. DMND chief executive Alejandro De La Torre called it the moment “a miner just mined the first Stratum V2 block to power their own product end to end” (TFTC). Ocean’s DATUM takes a lower-tech route to the same end, letting miners build templates on today’s firmware without a new ASIC image. Both attack the pool’s most valuable asset on purpose. That the pools themselves are backing the change tells you how seriously the industry now takes the concentration risk, and how much reputational value there is in no longer being the single entity that decides what Bitcoin includes.

Censorship as a business liability

The reason template control is not just a theoretical concern is that pools have used it. Researchers have repeatedly caught F2Pool, which runs around 15% of hashrate, filtering transactions tied to wallets on United States sanctions lists, first flagged by the pseudonymous developer 0xB10C in 2023 and recurring since (The Miner Mag). F2Pool co-founder Chun Wang did not hide it. Asked about the filtering, he wrote that he had every right not to confirm transactions from the people he called criminals, dictators and terrorists, then, after backlash, disabled the filter and argued that a censorship-resistant system “must be designed to resist censorship at the protocol level, rather than relying on each participant to act conscientiously” (The Miner Mag).

That episode is the business case for Stratum V2 and DATUM stated in the negative. As long as a single operator picks the transactions for a big share of blocks, that operator is a pressure point: a regulator, a court or a payment-network partner can lean on it, and its miners inherit the censorship whether they like it or not. Move template construction back to the miners and the lever mostly disappears, because there is no longer one desk to lean on. Censorship, in other words, is not only a neutrality problem for Bitcoin; it is a liability for the pool, and a reason its own institutional clients might one day prefer it not hold the template at all.

The regulator’s view, and where it could land

For all that, mining itself has drawn friendlier regulatory treatment in the United States than almost any other crypto activity. In March 2025 the Securities and Exchange Commission’s Division of Corporation Finance issued a staff statement concluding that proof-of-work mining, including participating in a mining pool, is not a securities transaction, because a miner’s reward flows from its own computational contribution, an administrative or ministerial act, not from the entrepreneurial efforts of others (The Block). A broader interpretive release in March 2026, under chair Paul Atkins, extended that no-securities view across mining, staking, wrapping and no-consideration airdrops (SEC). Atkins framed it as the agency doing its job by drawing clear lines in clear terms.

Notice what that covers and what it does not. The act of mining is not a security. The pool, as a business that custodies miners’ funds, promises fixed payouts it must fund from capital, and sells settlement and structured products, looks a lot more like a financial intermediary, and that is a different regulatory surface. Nobody has drawn that line yet, but the float, the FPPS promise and the prepaid-settlement products are exactly the features a supervisor eventually asks questions about. It is worth remembering that in Europe the regulatory conversation has already moved from whether to license crypto businesses to how to supervise them once licensed, a shift we covered in our look at MiCA in 2026, where the license turned out to be the easy part. Mining pools have so far sat outside that perimeter as pure infrastructure. The more they behave like insurers and settlement venues, the harder that position is to hold.

Frequently Asked Questions

How do mining pools make money if fees are near zero?

The headline fee is only one layer. Pools also capture the transaction-fee and ordering value from building blocks, earn on the float of miners’ unpaid balances they hold between payouts, funnel miners into affiliated exchanges, hardware and financing businesses, and sell forward hashrate contracts and prepaid settlement. For the largest pools the fee is closer to a customer-acquisition cost than a profit center.

What is the difference between FPPS and PPLNS mining pool payouts?

Both pay miners for proven work, and they differ in who carries the risk of bad luck. Under FPPS the pool pays you daily as if you found your statistical share of blocks and fees whether or not the pool got lucky, so the pool bears the variance and usually charges a higher fee. Under PPLNS you are paid only when the pool actually finds a block, so you carry the variance in exchange for lower fees and built-in protection against pool-hopping.

Are Bitcoin mining pools a threat to decentralization?

They are the main centralization pressure on Bitcoin today. The top four pools control close to 70% of hashrate and it takes only three of them to cross half the network. The sharpest risk is not a direct attack but transaction selection, because under the older Stratum V1 protocol the pool operator, not the individual miner, chooses what goes into each block. Stratum V2 Job Declaration and Ocean’s DATUM aim to hand that choice back to miners.

Can a Bitcoin mining pool steal my coins?

A pool cannot spend coins already sitting in your own wallet, but most pools do hold your earned rewards as float between payouts, which makes the pool a counterparty you are trusting. Pools have failed before, as when Poolin froze withdrawals in 2022 after deploying its float into risky yield strategies. Non-custodial schemes such as Ocean’s TIDES pay you directly from each block so the pool never holds your funds, which removes that specific risk.

Does the SEC regulate Bitcoin mining pools?

The SEC stated in 2025 that proof-of-work mining, including joining a mining pool, is not a securities transaction, and a 2026 interpretive release extended that view across mining and staking. That covers the act of mining, not necessarily the pool as a financial business. Because pools custody miners’ funds, promise fixed payouts and sell settlement products, they could still draw future scrutiny as financial intermediaries even though mining itself is not treated as a security.

Yuki Tanaka is a senior mining and markets correspondent at HOGE Wire, covering Bitcoin’s industrial base from hashrate to hardware.