Hashprice Is Up 22% and Hashrate Has Not Set a Record in 320 Days

Hashprice just ran 22% in thirty days. Network hashrate has not made a new high in roughly 320 days — the longest drought in a decade. Those two facts together describe a window that does not open often, and the only thing that decides whether it is open for you is the price you pay per kilowatt-hour.

Let us start with the numbers, all of them checkable. As of the 6 September 2026 adjustment, Bitcoin difficulty rose 1.31% to 127.45T at block 965,664 — the eighth increase of the year. Network hashrate was roughly 934 EH/s at that adjustment and has been tracking near 906–940 EH/s since, per CoinWarz. Over the preceding thirty days, hashprice climbed from about $32.42 to $39.63 per PH/s per day. Transaction fees contributed just 0.43% of miner revenue in a recent 24-hour window.

Read those together and the conclusion is uncomfortable but clear: in 2026, hashprice is almost entirely a function of Bitcoin’s price divided by how much competition showed up. Fees are a rounding error. And competition, for the first time in years, is not showing up.

First, the stakes: a window is only a window if you can afford to be in it

When difficulty falls and price recovers at the same time, the same machine does more of the network’s work for the same electricity. That is the entire mechanism, and it is temporary by design — the difficulty adjustment exists precisely to close it.

The catch is that the benefit is not distributed evenly. It accrues in proportion to how far your electricity cost sits below the breakeven line. A miner at $0.17/kWh running a 9.5 J/TH machine is at breakeven today and has captured nothing. A miner at $0.0364/kWh running the same machine has a cost base roughly a fifth of the breakeven rate, and captures the whole move.

This is the only formula in mining worth memorising. It tells you the electricity price at which a given machine produces exactly nothing. Above it you lose money on every kilowatt-hour; below it, the gap is the margin. Here is what it produces at today’s $39.63 hashprice, across the live OneMiners catalogue.

	Miner	Efficiency	Breakeven $/kWh	Headroom at $0.0364
	Sealminer A4 Ultra Hyd 886 TH/s	9.45 J/TH	$0.1748	4.8×
	Antminer S23 Hyd 3U 1.16 PH/s	9.50 J/TH	$0.1738	4.8×
	Antminer S23 Hyd 580 TH/s	9.50 J/TH	$0.1738	4.8×
	Antminer S21 XP Hyd 473 TH/s	12.00 J/TH	$0.1376	3.8×
	Antminer S21+ 235 TH/s	16.50 J/TH	$0.1001	2.8×
	Antminer S19k Pro 110 TH/s	23.00 J/TH	$0.0718	2.0×

Two things jump out of that table, and both are worth saying plainly.

First, the most efficient machine in our catalogue is not the most expensive one. The Sealminer A4 Ultra Hyd runs 9.45 J/TH at $9,999. The Antminer S23 Hyd 3U runs 9.50 J/TH at $28,399. On efficiency alone, the cheaper unit edges it. The S23 Hyd 3U earns its price on density — 1.16 PH/s in a single 3U chassis, which matters enormously when you are paying for rack space and hydro plumbing — but if the question is strictly joules per terahash, we are not going to pretend the flagship wins a contest it does not win.

Second, even the oldest machine on that list clears breakeven at our rate — the S19k Pro breaks even at $0.0718/kWh, roughly double a $0.0364 prepay. At a US residential marginal rate of $0.17, it is deeply underwater. Same machine, same hashprice, opposite outcome. The variable is the address.

Why the hashrate did not come back

In every previous cycle, a 22% hashprice recovery would have been met within weeks by machines switching back on and new capacity energising. This time the response has been close to flat. There are three reasons, and they compound.

1. The best-capitalised operators are chasing AI leases instead. Listed miners have disclosed over $70 billion in AI and HPC contracts, and a megawatt under a fixed AI lease does not come back to Bitcoin when hashprice ticks up.
2. Power itself has been repriced. Utilities filed for $18.6 billion in rate increases in the first half of 2026. Marginal capacity that was economic at last year’s power price is not economic at this year’s.
3. Older hardware never came back from the drawdown. The largest contraction since 2021 — roughly 12% off the November 2025 peak — retired a lot of high-J/TH fleet permanently. Machines that break even at $0.07/kWh do not get re-racked in a $0.09 market.

How to actually use a window like this

The discipline here is boring and it works.

• Buy efficiency, then buy the rate. J/TH determines whether a machine can survive a bad market. $/kWh determines how badly a bad market hurts. Get both right and the middle takes care of itself.
• Lock the term long. A rate that floats has no defensive value in exactly the moments you need it. OneMiners fixes electricity for up to seven years, with the prepaid ladder running 4% off at one year, 12% at three and 30% at seven — a formula you can reproduce yourself off the published online-ordered rate.
• Check the all-in, not the headline. A rate quoted before management and performance fees is not a rate. OneMiners pricing includes all management services.
• Verify independently. Run any machine through asicprofit.com and check difficulty yourself at CoinWarz. Every figure in this article is designed to be checkable.

The cheapest published rate in our network is $0.0364/kWh on a seven-year Nigeria prepay, with the premium tier at $0.0480. Against typical 2026 industrial hosting around $0.065–$0.08/kWh — and against several well-known providers who publish no rate at all or take a percentage of your output instead — the comparison is not close. The full card for all fifteen sites is published at hosting centers.

Final thoughts

Difficulty is 13% off its peak. Hashprice is up 22% in a month. Hashrate has not set a record in the better part of a year because the industry’s biggest builders are pointing their megawatts at GPUs instead. Every one of those conditions favours a miner whose cost base is fixed and low.

None of them will last forever. Difficulty adjusts every two weeks precisely so that windows close. What does last is a contract — and the miners who come out of this period ahead will be the ones who used a good market to lock a good rate, rather than waiting for a better one.

Revenue is weather. Fees are a rounding error. Electricity is physics.