Every four years the Bitcoin block reward is cut in half — and every four years, miners relearn the same lessons. Here is what the halving actually does to mining economics, beyond the headlines.
Bitcoin's issuance schedule is fixed in code: every 210,000 blocks (roughly four years), the reward paid to miners per block halves. After the 2024 halving the reward stands at 3.125 BTC per block. The next halving, expected in 2028, will cut it to 1.5625 BTC. Nothing else about mining changes on halving day — the same machines hash at the same speed and burn the same electricity. Only the revenue side of the equation is cut in half overnight.
Miner revenue per terahash — known as hashprice — drops by roughly 50% at the moment of halving, assuming price and network hashrate stay constant. In practice they never do. Historically, BTC price appreciation around halving cycles has partially or fully compensated miners, while inefficient hashrate leaving the network reduces difficulty and redistributes rewards to the machines that remain. Both effects take months to play out, and neither is guaranteed.
The halving acts as an efficiency filter. A machine's survival depends on one number: joules per terahash. At post-2024 hashprice levels, hardware above roughly 30 J/TH struggles at typical hosting rates, while modern units in the 13–18 J/TH range (Antminer S21 series, Whatsminer M60 family) retain healthy margins. This is why fleet refresh cycles cluster around halvings, and why used-market prices for older generations collapse in the months after one.
For anyone buying hardware or mining shares, the halving calendar should shape the decision. Buying efficient, current-generation machines mid-cycle means the equipment earns through one full halving with margin to spare. Power rate matters more than ever: at $0.055/kWh a machine survives conditions that would bankrupt the same machine at $0.09. This is precisely why HiveHash deploys in low-cost hydro and gas regions — the cheapest electricity is the best halving insurance there is.