Selling the heat: what mining waste heat is actually worth

A mining machine is, thermodynamically, a resistive heater that happens to compute. Essentially all the electrical energy you pay for leaves as heat. Most operators spend money getting rid of it. A growing number sell it, and in places with expensive power that second line is the difference between a marginal site and a good one.

The number that decides everything

Not how much heat, but how hot. Heat is only useful to a buyer at a temperature above what they need, and the routes divide sharply on that.

• Air cooled exhaust leaves the machine warm rather than hot, typically somewhere in the thirties to forties in Celsius depending on inlet temperature and load. Useful for drying and for space heating in a building you already control. Difficult to sell to anyone with a pipe network.
• Hydro loops deliver water at a genuinely useful temperature, commonly in the fifties to sixties depending on the machine and how the loop is configured. That is the range where district heating, greenhouses, pools and industrial process heat become real conversations.
• Immersion sits in between and can be pushed higher with the right fluid and heat exchanger, at the cost of running the chips warmer.

If heat reuse is part of the business case, it points hard at hydro. That is a hardware decision made for a revenue reason, which is unusual and worth being explicit about when you shortlist machines.

Who actually buys heat

District heating networks are the largest buyers and the hardest to join: they want guaranteed supply, metered quality and a contract measured in years, and they are concentrated in northern Europe. Greenhouses are the most approachable, because they need low grade heat at night and in winter and are often already sited near cheap industrial land. Beyond those: swimming pools and leisure centres, timber and grain drying, aquaculture, fruit and vegetable dehydration, and industrial washing.

What all of them have in common is that they were paying for heat already. You are not creating demand, you are undercutting a gas bill. That is why the conversation works, and also why it fails the moment your supply is unreliable: they cannot let a crop or a building go cold because hashprice dropped.

The obligation you are taking on

This is the part operators underestimate. Selling heat converts a fleet you can switch off at will into a fleet with a delivery obligation. A curtailment event, a firmware rollout or a bad month for hashprice now has a counterparty attached. Every serious heat contract therefore needs a backup source, usually the buyer's existing boiler kept warm, and a clear statement of who pays when you cannot deliver.

Negotiate the right to switch off explicitly. If you sign away the ability to curtail, you have also signed away the strategy in curtailment and demand response, which is often worth more than the heat.

What it is worth

Price the heat against what the buyer currently pays per unit of energy, not against what you pay for electricity. In practice a deal lands somewhere below their existing cost, because you need to be clearly cheaper to be worth the switching effort, and above zero, because you have capital in the heat exchanger. The realistic effect is a reduction in your effective power price rather than a business in itself. A cent or two per kilowatt hour does not sound like much until you run it through your breakeven.

Getting the engineering right

• Put a heat exchanger between your loop and theirs. Never share fluid, and never let their maintenance failure become your outage.
• Meter at the boundary, on both sides, and agree whose meter settles the invoice.
• Size the loop for their peak, not your average, or the deal fails in the one week it mattered.
• Keep the dry cooler. It is the dump you need whenever the buyer is not taking heat, and the day you rely on the buyer as your only heat sink is the day you overheat.