I used to think Bitcoin mining involved tiny pickaxes and a very patient canary. The joke is bad, but the correction is useful: miners are not digging coins out of the internet. Specialized computers repeatedly hash block headers, trying to produce a result below the network’s target. The first valid result helps add a new block to Bitcoin’s transaction history. Bitcoin’s developer guide explains the mechanics without requiring a decoder ring.

So where does the money come from? A miner’s reward is two things: a block subsidy—newly created bitcoin—and the transaction fees attached to the transactions in that block. The protocol’s block-chain documentation calls these together the block reward. A solo miner keeps the whole payout, but may wait a long time between wins. A pool spreads the work among many miners and pays smaller, steadier amounts based on each miner’s share of the work. Think lottery ticket versus office pool, except the office has a warehouse full of humming machines.

The bill arrives every hour. Electricity is the obvious cost, but not the only one. There is specialized ASIC hardware, cooling, real estate, internet service, maintenance, financing, and the occasional experience of discovering that a hot computer room is not a spa. A miner is profitable only when the expected value of rewards covers those costs.

This is why energy price and machine efficiency matter so much. The Cambridge Centre for Alternative Finance’s methodology models profitability by comparing mining revenue with the electricity needed to run different generations of hardware. Its efficiency measure is joules per terahash: lower is better. Cambridge also notes that its electricity-price input is an assumption, not a universal fact, and that its model does not include every cost, such as maintenance and cooling. That is a helpful humility flag. A spreadsheet can be precise and still not know everything about the warehouse.

Competition adds another wrinkle. More miners do not make Bitcoin permanently easier to mine. The network adjusts difficulty every 2,016 blocks—aiming for about two weeks—so the target becomes harder or easier depending on how quickly blocks were found. The protocol documentation lays out that adjustment. More machines can mean more total computing power, but also a larger crowd reaching for the same scheduled reward.

And the reward itself is not fixed forever. The subsidy is cut at programmed intervals, roughly every four years, while transaction fees remain tied to demand for block space. That makes the business a moving target: price, fees, difficulty, electricity, machine prices, and the calendar all matter. The Congressional Research Service puts the broader point plainly: mining energy intensity tends to rise and fall with profitability, and when electricity and maintenance costs outrun revenue, operators may shut machines down or delay new purchases.

What happened? A digital currency turned security work into a competitive energy business. What did it mean? Mining is less like finding buried treasure and more like running a factory whose product, costs, and rules can all change. What can we carry forward? Whenever a crypto pitch says “the reward,” ask the quieter questions too: What are the costs? How often do they change? Who gets paid, and what happens when the easy money gets harder?