The Bitcoin Beacon
Hashrate is down 15% from January and miners are chasing AI money. The difficulty adjustment quietly makes that a non-event.
Read the mining charts and you might think something is wrong. Bitcoin's network hashrate has fallen to roughly 870–900 exahashes per second, down about 15% from the January high near 1,065 EH/s and more than 20% below the peak of around 1,150 EH/s set last October. Difficulty has stepped down through more than a dozen adjustments this year. Machines are going dark. This is not a problem. It is the system working.
Two things are happening at once. First, a growing share of public miners is diverting energy and capital toward artificial intelligence and high-performance computing, which pay better than block rewards at today's prices — the sector's already-permitted, energized sites are more valuable rented to AI than run as pure mining. Second, the profitability squeeze from a subdued bitcoin price has pushed the least efficient rigs offline.
So hashrate falls. And every roughly two weeks, the network notices and responds.
Bitcoin retargets its mining difficulty every 2,016 blocks — about a fortnight — to keep blocks arriving on average every ten minutes. When miners leave and blocks start coming slower, the next adjustment lowers difficulty, making it cheaper to find a block for whoever remains. When miners flood in, difficulty rises to soak up the extra power. It is a closed-loop controller, and it has run without interruption since 2009.
The consequence is the part people miss: the amount of new bitcoin issued does not depend on how many miners there are. Blocks pay the same subsidy whether the network runs at 500 EH/s or 1,500. Falling hashrate does not slow issuance, does not threaten the 21-million cap, and does not, on any normal timescale, stop the chain. It lowers the cost of participating until participation stabilizes.
The security budget floats to meet whoever shows up. The issuance schedule ignores them entirely. That asymmetry is the whole design.
None of this means hashrate is meaningless. A very large, very sudden exodus — a coordinated shutdown, a regional grid failure — could stretch block times uncomfortably in the days before the next retarget, as happened briefly after China's 2021 mining ban. And a lower total hashrate is, at the margin, a lower physical cost to attack the network, even if that cost remains astronomically high.
The subtler worry is concentration. If the miners who leave are the small and independent ones, and the survivors are a handful of large public companies straddling mining and AI, the network's security starts leaning on fewer shoulders. The United States, China, and Russia already account for something like two-thirds of global hashrate. That is worth watching — not because the network is fragile today, but because decentralization is the property that makes all the rest of it matter.
But those are second-order concerns, not the headline. The headline — “hashrate down 15%” — describes a network doing precisely what it was built to do: absorb the departure of miners without a flicker in the thing that actually matters, the steady, unbribable arrival of the next block.
Hashrate is a price signal, not a heartbeat. It rises and falls with the economics of mining, and the difficulty adjustment exists precisely so those swings never touch the monetary schedule underneath. Miners chasing AI money is a story about miners. Bitcoin, meanwhile, keeps printing a block every ten minutes and doesn't ask who's still in the room.
Editor's note: hashrate and difficulty figures are drawn from public trackers and move continuously; the values cited reflect mid-July 2026 readings. This is opinion and analysis, not financial advice.
Free. Five minutes. No hype.
Subscribe free