Bitcoin Difficulty Rises 4.16%, Short of the 4.70% Modelled

Bitcoin's mining difficulty rose 4.16% at the 19 September retarget, lifting it to 132.76 trillion from 127.45T. The change landed at block 967,680 at 07:09:05 UTC on Friday, on mempool.space's adjustment endpoint, and CoinWarz gives the same percentage and the same two difficulty figures. It is the ninth increase of 2026. Network hashrate sat near 914 exahashes a second when both dashboards were read on Sunday afternoon, under the 934 EH/s recorded at the previous retarget on 6 September. Difficulty is the number that decides how much work a block costs, and the protocol resets it every 2,016 blocks to hold the average block near ten minutes.

The forecast overshot

CryptoSlate modelled this retarget five days out and put it at 4.6976%, noting that some sources showed 5.26%. The network delivered 4.16%. The same 15 September analysis, by editor Liam Akiba Wright, set out what an increase of that size would cost: hashprice of $39.25 per petahash a day falling 4.49% to $37.49, and a bitcoin price of about $82,900 needed to hold miner revenue where it was. That $82,900 is modelled against 4.70%, not against the 4.16% that arrived, so the figure does not transfer. No outlet read here has published a breakeven recalculated on the real number, and this desk will not invent one. Bitcoin traded around $81,330 on Friday morning on crypto.news' reading, up roughly 6% in a day off the $75,000 area.

Fees are still a rounding error

Across the 1,008 blocks to Sunday, miners were paid 3,168.96 bitcoin, of which 18.96 came from transaction fees. That is a share of 0.60%. CryptoSlate's mid-month figure, 0.0183 BTC a block and 0.59% of rewards, sits in the same place, and news.bitcoin.com counted 0.43% over a single day on 6 September. Three windows, three readings, no contradiction between them. What they say together is that the fee market does almost nothing for miners, so a difficulty increase lands on the block subsidy and on the price of the coin.

What the estimators cannot agree on

The next retarget is where the dashboards part company. CoinWarz expects an increase of 1.85% to 135.22T on 3 October at 05:33 UTC. mempool.space gives a decrease of 5.66% and a date of 4 October. Newhedge shows a decrease of 8.07%. All three were read inside six minutes of each other, at about 7.5% through the epoch, with more than 1,800 blocks still to mine. Early estimates move, and these have not settled. mempool.space puts the running epoch's average block at 11 minutes 50 seconds, which is what a downward revision looks like while it is happening.

Who feels 4% first

An increase of 4.16% takes about that much off the bitcoin each machine earns, with nothing given back unless the price moves. This desk covered a 0.99% rise to 127.48T on 8 August, and the weeks since have run both ways: eight increases and ten decreases had left difficulty 11.93% under its January level as of 6 September. Power cost separates the operators who absorb this from the ones who cannot. CryptoSlate modelled electricity at $48 a megawatt-hour and found machines under roughly 30.5 joules per terahash covering their power in every scenario it ran. Miners on worse rates have less room, and Ethiopia cut its miners to about a quarter of contracted power this month. Canaan disclosed selling 54 bitcoin near $79,000 and 3,952 ether for $13.9 million in the same stretch. CryptoSlate, which reported the sale, declined to read intent into it.

The disposal "does not establish distress or prove that the projected adjustment drove the transaction," the analysis said.

Nothing published since Friday says how the 4.16% has landed on any operator's output. The figure to watch is the October retarget, and the three estimators aimed at it currently point three ways.