Solo Mining Bitcoin at Home: a 34 TH/s Miner's Daily Odds Are 1 in 205,534
Plug in 34 TH/s, point it at a solo pool, and the network gives you a 1-in-205,534 shot each day at a full bitcoin block. That is what today's difficulty, read from our own node, works out to. Two small miners beat worse odds on CKPool this year. The other side of that ledger is thousands of machines that found nothing.
Heatbit opened preorders on October 22 for Bitwater, a heating element for electric water tanks that warms the water with the waste heat of bitcoin mining chips [3][4]. The spec sheet lists up to 34 TH/s at up to 600 W, with first deliveries planned for the first quarter of 2027 [3][4]. The line that traveled furthest on X was a different one: out of the box, the company says, the device mines solo on its own pool, "so a block found is yours" [3].
This is not a product review. The question that line raises applies to anyone with a miner in the garage: how likely is a machine that size to find a block on its own? Heatbit declines to put a number on earnings, and says anyone quoting a yearly figure "is guessing" [3]. The block odds, at least, don't require guesswork. Difficulty is set by the chain, and our node reads it.
The arithmetic
At block 972,732, mined at 07:04 UTC on October 23, 2026, our node (a full, unpruned Bitcoin Core) reported a difficulty of 140,577,680,051,920 [1]. That value has applied since block 971,712 on October 16 and implies a network running at just over 1 ZH/s [1]. Here is what it means at four hashrates:
| Hashrate | Average wait per block | Chance per day | Chance per year | Pooled average (subsidy only) |
|---|---|---|---|---|
| 1 TH/s | 19,133 years | 1 in 6,988,154 | 0.005% | 45 sats/day |
| 34 TH/s | 563 years | 1 in 205,534 | 0.18% | 1,520 sats/day |
| 230 TH/s | 83 years | 1 in 30,383 | 1.19% | 10,285 sats/day |
| 1 PH/s (1,000 TH/s) | 19 years | 1 in 6,988 | 5.09% | 44,719 sats/day |
Read the 34 TH/s row as a lottery ticket that never expires but almost never pays: over a full decade, the chance of at least one block is 1.76%. The same machine in a pool would earn about 1,520 sats a day from the subsidy alone, before pool fees, if it hashed nonstop. Heatbit says its unit is designed to run more than 80% of the time [3], so a real figure would land lower.
These are our own calculations from the node's difficulty and the formula above [1][2]. The yearly columns assume both the machine's hashrate and the network difficulty hold steady for twelve months. Difficulty never does.
Same average, very different ride
Solo and pooled mining have the same expected value over a long enough horizon. A pool combines thousands of machines and splits each reward by contribution, so payouts arrive small and often. Solo strips that away: you keep the whole block, today 3.125 BTC in subsidy plus transaction fees, minus whatever the service charges. solo.ckpool.org, the best-known option, takes 2% [5].
What changes is variance, and at home-miner scale it is extreme. A 34 TH/s miner choosing solo is trading a steady trickle of sats for a tiny chance at a lump sum. Neither choice earns more on average. Only one of them can go a lifetime without paying.
563 years. That is the average wait for a 34 TH/s miner to find a block alone at today's difficulty.
A water heater that mines: what the math says
Bitwater is less a new idea than the latest turn in a trend: putting the heat a mining chip throws off anyway to use. Heatbit already sells space heaters that mine, and other companies, Superheat among them, are working on similar products [3][4]. The pitch, repeated on the company's site and in its launch thread, is that the electricity "was going to be spent on hot water anyway," so whatever it mines comes on top [3][9].
The published numbers are these. The element carries 20 BM1370 chips rated at up to 34 TH/s and up to 600 W, built to mine more than 80% of the time, according to the spec sheet and the thread Heatbit posted on X on October 22 [3][9]. The company also says that as the water nears its set temperature, autotune underclocks the chips and hashrate drops [9]. In other words, 34 TH/s is a ceiling, not an average. Preorders are $899 for the first batch and up to $1,299 later, with shipping from the first quarter of 2027, to the US and Canada only [3][9].
Run through the same formula at today's difficulty, 34 TH/s mining solo around the clock averages one block every 563 years, a 1-in-205,534 chance per day [1][2]. At the 80% utilization Heatbit cites, about 27.2 TH/s on average, the wait stretches to 703 years and the daily odds fall to 1 in 256,917 [1][2][3]. Pointed at a pool instead, the same hashrate would average between 1,216 and 1,520 sats a day in subsidy, before fees [1][2].
Three caveats no marketing can make go away:
- It only adds up if you already heat water with a resistive electric element. Bitwater replaces the lower element in an electric tank heater. It does not work with gas tanks or tankless units, the company said in replies on X [9]. Against any setup that uses less electricity for the same heat, the "free" part stops being free.
- Your power price still matters. Using the same electricity as before doesn't make it free. What the miner earns has to be weighed against what the device costs and what you already pay to heat water, and that balance shifts with every tariff, with difficulty and with the bitcoin price. Heatbit says as much: anyone quoting a yearly figure "is guessing" [3]. For what a petahash actually earns right now and how much of that goes to electricity, see our breakdown of hashprice and mining costs.
- Pooled, it's not a lottery ticket. Out of the box it mines solo on Heatbit's pool; switch it to a pool such as Luxor, which the company says takes a minute in the app, and it stops being a shot at 3.125 BTC and becomes a trickle of sats [3][9]. Either way, the average is the same, as with any miner.
The miners who beat the odds
Improbable is not impossible, and the chain has two recent examples anyone can check.
Block 943,411. A solo.ckpool.org user found it at 21:02 UTC on April 2, 2026, with 3,961 transactions and a 3.13904392 BTC reward [6]. CKPool developer Con Kolivas posted on X that the winner had roughly 230 TH/s and that a miner that size has "a 1 in ~28k chance per day" [5]. It was the 312th block solved solo on his service [5]. Our formula agrees: at the difficulty back then, 133.79 trillion, it gives 1 in 28,904 per day [2][6]. At today's difficulty, the same rig would be at 1 in 30,383.
Block 960,804. Another CKPool miner found it at 02:11 UTC on August 3, 2026, with 4,243 transactions and 3.15689830 BTC [6]. Kolivas described the winner's hashrate as "wildly variable, presumably rental, peaking at 100 PH," according to Bitcoin.com, which counted it as the service's 317th solo block [7]. Bitcoin.com dates the block August 2; the chain's timestamp says August 3, and that is the date we use [6][7].
CoinDesk also cited earlier wins on the same service: a miner with about 6 TH/s in November, at 1-in-180-million odds by Kolivas's estimate, and another who rented roughly $75 worth of hashpower in late February and pointed 1 PH/s at the pool for a few hours [5]. We could not match those two to specific blocks, so they stand on that attribution.
The wider count is the sobering part. Citing a solo-block tracker, The Block reported that solo pools found 20 blocks in the twelve months to April, one every 18.7 days on average [8]. The network produced more than 50,000 blocks over the same stretch.
Next week's retarget makes it harder
Difficulty resets every 2,016 blocks. As of this writing, our node had counted 1,020 blocks into the current period at an average of 573 seconds each, faster than the protocol's 600-second target, and projected a 4.7% increase at block 973,728, around 21:41 UTC on October 29 [1]. Other public trackers show somewhat different figures because they look back over different windows; ours updates with every block on our Network page.
If it lands near that estimate, difficulty would reach about 147.19 trillion [1]. The 34 TH/s wait would stretch to 589 years, or 1 in 215,204 per day, and the 230 TH/s rig to 87 years [1][2]. Whether the trend keeps climbing is visible on our bitcoin hashrate chart.
Before you point a miner at a solo pool
- You are not choosing a higher payout. Over time, solo and pool pay the same. You are choosing how the payout arrives.
- The pool builds the block, not you. On a service like CKPool, the block template comes from the service. Picking your own transactions requires your own node; what changed in the latest release is in our Bitcoin Core 32 checklist.
- Every win is public. A winning block sits on the chain for good, header and signal bits included, the same data we used to track miner support for BIP-110. Anyone can verify who found it and what it paid.
- The subsidy halves at block 1,050,000. It drops to 1.5625 BTC, which cuts the pooled column in the table in half if difficulty stays put.
Difficulty, the running retarget estimate and network hashrate refresh block by block on our Network page, straight from our own node. If you're running your own numbers, start there.
Sources
Source type in parentheses. Times in UTC.
- NodeWitness, reading from our node (full Bitcoin Core, unpruned) at 07:09 UTC on October 23, 2026, block 972,732: difficulty 140,577,680,051,920, in force since block 971,712; 1,020 blocks into the period, 573.3-second average, estimated +4.7% adjustment (estimated difficulty 147,191,955,966,640) at block 973,728. Available on our Network page and hashrate chart. Table calculations are our own. (Primary, own data)
- Bitcoin Wiki, Difficulty, accessed October 23, 2026. Average time per block: difficulty × 2³² / hashrate. (Secondary, collaborative technical documentation)
- Heatbit, Bitwater, accessed October 23, 2026. Up to 34 TH/s and 600 W, solo mining on its pool by default, designed to run more than 80% of the time, caveat on earnings estimates. (Primary, manufacturer)
- Crypto Briefing, Heatbit unveils Bitwater, a water heater that mines Bitcoin, October 22, 2026. Preorder opening and Q1 2027 shipping target. (Secondary, trade press)
- CoinDesk, Solo bitcoin miner overcomes 1-in-28,000 odds to secure $210,000 block reward, April 6, 2026. Con Kolivas's X post, ~230 TH/s, 312th solo block, 2% fee, the November and February cases. (Secondary, trade press)
- mempool.space, blocks 943,411 and 960,804, accessed October 23, 2026. Timestamp, miner (Solo CK), reward, transaction count and difficulty. (Secondary, open-source on-chain explorer)
- Bitcoin.com News, Solo Bitcoin Miner Defies the Odds, Lands $200K Block Reward Jackpot, August 6, 2026. CKPool's 317th block and Kolivas's comments on rented hashrate. (Secondary, trade press)
- The Block, Lucky solo bitcoin miner beats 1-in-28,000 daily odds to win $210,000 block reward, April 5, 2026. Count of 20 solo blocks in twelve months, one every 18.7 days. (Secondary, trade press)
- Heatbit, Bitwater launch thread on X and the account's replies in the same thread, October 22–23, 2026 (first post at 16:01 UTC). Specs, "Solo mining or pool – your choice," Q1 2027 shipping, temperature-based hashrate throttling, electric tank heaters only. (Primary, manufacturer)
More in Network and mining
Last updated: October 23, 2026