What does it actually cost to produce one bitcoin?

Traditional mining calculators often present breakeven as a single date. But difficulty can change every 2,016 blocks, so a long-term projection can move materially after a single retarget. PM Countdown models breakeven as a range instead.

Book depreciation is a policy choice that hits earnings and tax. Market value is what somebody will actually pay you. An ASIC does not decay on a straight line; it is worth whatever its remaining earning power is worth to a buyer, which is why old rigs go to zero the week hashprice drops below their power cost rather than gliding gently to salvage.

Five of the eight published mining valuation models are the same trade wearing different hats, they all resolve to (hashrate × efficiency × power price) ÷ issuance. Averaging eight raw signals silently hands 70% of the weight to one factor. So they are grouped into families first, and only families are weighted.

No. Bitcoin spent 26,325 of 29,093 windows under the network’s cost to produce one bitcoin and closed Down 49.9% of the time, against 52.1% over it. Buying Down five minutes in made −0.9¢ a contract under the line and +0.9¢ over it; buying Up made −0.9¢ and −2.4¢. Month by month, Down beat Up under the line in 6 of 11 months. Mining cost is a floor that works over weeks and months, as miners switch machines off and sell less. A 15‑minute window does not feel it.

Should Agent LoLo stop buying Down when Bitcoin trades under its mining cost? A little worse, and still worth taking: under the line his Down tickets made +2.3¢ a contract (619 tickets) against +3.5¢ over it (361). Skipping them would have given up $9,175 on his practice bankroll, so there is no rule. The bar for a rule, set on Oct 20, 2026, is 3¢ a contract worse under the line, in both halves of the days. The test reruns every week; the full review is Nov 17, 2026.

Buying hashrate is a portfolio decision, not a dropdown. The question is whether a machine sits on the efficiency frontier or whether you are paying a premium for terahashes you cannot power.

Almost all public ASIC price data is asking price, not sale price. A listing is an offer nobody has agreed to pay. The only free, dated, lot-sized record of what machines actually traded for is Luxor's RFQ Insights series, which ran April to November 2023 and stopped; their ongoing index is itself ask-derived and sits behind an API key and a paid tier. So nothing here is called a sale unless a source published a fill.

How much does it cost to mine one bitcoin?

It depends on the machine and your power price. Counting power, pool fees, maintenance, the machine net of its resale value and a 10% required return over a three-year life through the halving, a 50-machine Antminer S21 Pro fleet on 5¢ power produced one bitcoin for about $92,700 on 20 October 2026. The most efficient machine on sale, the SealMiner A4 Pro Air, did it for about $83,500, at ASIC prices from 2 October 2026. The calculator above recomputes both as you change the inputs.

What is bitcoin’s fair value based on mining costs?

Our model blends four mining valuation models: production cost, hashprice payback, the Puell multiple and thermocap. It samples the uncertain inputs 20,000 times. On 20 October 2026 its fair value for bitcoin was about $79,000, with a P10 to P90 range of $69,300 to $89,600, against a price of $81,500. The network's cost to produce one bitcoin, all in, was about $88,600. It is a reading of mining economics, not a price forecast.

Which ASIC miner has the lowest cost per bitcoin?

On 20 October 2026, at 5¢ per kWh and with ASIC market prices from 2 October 2026, the SealMiner A4 Pro Air ranked first at about $83,500 all in per bitcoin, followed by the Antminer S21e XP Hyd 3U at about $88,500. The ranking moves when you change the power rate or the hurdle rate.

How does the 2028 halving change the cost of mining?

At block 1,050,000, expected around April 2028, the block subsidy halves from 3.125 to 1.5625 BTC. The same machines then earn half as many coins for the same power bill, so the power cost of each bitcoin roughly doubles: for an S21 Pro fleet on 5¢ power, from about $46,300 to about $100,000 at the difficulty of 20 October 2026. Any machine bought now spends a large share of its life after the halving, which is why this calculator prices it in.

What does all-in cost per BTC include?

Electricity at the wall including cooling, the pool fee, maintenance, the hardware cost net of what the machine resells for later, and the cost of the money tied up in it at your hurdle rate. Each is discounted over the machine’s working life and divided by the coins it is expected to mine as difficulty rises and the halving arrives.

Why does this calculator show ranges instead of one break-even date?

Difficulty changes every 2,016 blocks, about every two weeks, and nobody knows where it goes next. A single break-even date can move by months after one retarget, so the projections show a P10 to P90 range rather than one number.

Does bitcoin trading under its mining cost predict 15-minute price moves?

Not in our test. Across 29,093 Kalshi Bitcoin 15-minute windows from Dec 11, 2025 to Oct 20, 2026, bitcoin closed Down 49.9% of the time when it traded under the network's cost to produce one bitcoin and 52.1% when it traded over it, and buying Down five minutes in did no better under the line (−0.9¢ a contract) than over it (+0.9¢). Mining cost works as a floor over weeks and months, not minutes. The test reruns every week.

This page spends a lot of words arguing that an undated number is not evidence. It would be embarrassing to then quote "JPMorgan says $78,000" with no date and no link, so here is the whole bibliography with both.