What pays miners after the last bitcoin subsidy?

The subsidy ends. The job of processing blocks does not.

A block reward has two parts: newly issued bitcoin, called the subsidy, and transaction fees. Under the current mainnet rules, the subsidy halves every 210,000 blocks. That is a block-height rule, not a calendar appointment. The next reduction from 3.125 to 1.5625 BTC happens at block 1,050,000; it is not the last halving.

Bitcoin Core counts whole satoshis. Starting with 5,000,000,000 satoshis and repeatedly halving with integer rounding leaves one satoshi after the thirty-second reduction and zero after the thirty-third, at block 6,930,000. The familiar date around 2140 assumes roughly ten minutes per block. Actual timing depends on blocks arriving, and any discussion this far out assumes the consensus rules remain unchanged.

With zero subsidy, miners can still collect fees for transactions included in their blocks. That is the intended long-run incentive described in section 6 of the white paper. It does not guarantee that fees will support any particular electricity bill, mining company or level of network security.

Hold the other inputs still and measure what disappears.

Assume a miner contributes 1 percent of network hashrate, the network averages 144 blocks per day, fees average 0.1 BTC per block and bitcoin sells for 100,000 dollars. These are deliberately round teaching inputs, not readings from the current market. Expected production is network share multiplied by blocks multiplied by subsidy plus fees.

The next halving removes 225,000 dollars per day in this frozen example. Revenue does not quite halve because fees remain. In the fee-only case, keeping the old 4.644 BTC/day would require 3.225 BTC of average fees per block at the same share and block rate. That is 32.25 times the assumed fee level. This is the replacement requirement, not a forecast that users will pay it.

Difficulty adjusts work, not a company's profit.

If unprofitable machines stop, blocks can arrive more slowly until difficulty adjusts. At a lower difficulty, surviving hashpower may earn more per hash. The network targets its block interval; it does not target a dollar reward or reimburse a miner for its machines. Competition, fees, bitcoin price and efficiency all change the economics together.

A mining pool reduces payout variability, but does not manufacture revenue that the network never paid. Pool fees and payout contracts also mean an operator's receipts need not match this simple proportional model. The long-run fee market and the security budget remain uncertain.

For a company, the next capital cycle matters first.

A machine bought today will wear out long before the final subsidy. Read the company's useful-life assumptions, power contracts, fleet efficiency, debt maturities and share issuance. A firm can change its business or close while the Bitcoin network continues. A fee-only network is not a promise that today's listed miners will survive.

Try the subsidy selector in the power-to-bitcoin example. Then change efficiency or electricity price. A cheaper fleet can reduce a loss; it cannot make the missing subsidy reappear.