Sound ASIC mining economics includes a replacement decision, not just purchase ROI. A machine can remain functional yet no longer fit the operating site because efficiency, repair cost, downtime or changing network conditions reduce its contribution.
Three decisions, one evidence set
| Decision | When it may fit | Evidence required |
|---|---|---|
| Repair | The fault is known and cost is modest relative to remaining use. | Diagnosis, quote, expected downtime and prior repair history. |
| Replace | Downtime, efficiency or maintenance no longer support the workload. | Delivered replacement cost and comparable operating assumptions. |
| Retire or repurpose | The machine cannot operate safely or economically in its current role. | Site limits, lifecycle cost and responsible disposition option. |
Keep a lifecycle record
Track delivered cost, uptime, repairs, energy use and output. This makes the next hardware decision evidence-based instead of driven by the latest headline estimate. Separate observed operating facts from forecasts: a week of good output does not erase recurring thermal, electrical or maintenance problems.
Compare the next step on equal assumptions
- Use the same power rate, availability assumption and cost boundary for old and new equipment.
- Include shipping, installation, electrical changes and any host charges for the replacement.
- Record repair lead time and the expected cost of an idle machine.
- Check site compatibility, cooling requirements and service availability.
- Define the evidence that would make a replacement unattractive before paying for it.
Separate a repair quote from a repair decision
A repair price is only one variable. Ask whether the repaired unit can reasonably operate for the period assumed, whether the fault may recur, how long the work will take and what the power cost will be after it returns. A replacement quote deserves the same treatment: compare delivered cost, commissioning delay, expected energy use and service support instead of comparing list prices alone.
Set retirement triggers early
Triggers can include repeated faults, unavailable parts, unsafe operating conditions, a sustained inability to cover direct operating cost or a hosting environment that no longer supports the machine. A trigger is a request to recalculate, not an automatic order to sell or shut down. It helps the operator avoid defending a past purchase with outdated assumptions.
| Trigger | Review action |
|---|---|
| Repeated downtime | Update availability and maintenance reserve. |
| Power-rate change | Recalculate direct operating contribution. |
| New repair quote | Compare repair against replacement on the same timeline. |
| Site change | Confirm electrical, cooling and contract compatibility. |
Conclusion
Replacement planning protects capital by treating hardware condition and operating cost as continuing inputs, not one-time assumptions.
Create a replacement file
Keep one record for every machine with its model, serial or internal identifier, installation date, firmware and site assignment. Add dated entries for repairs, parts, downtime, energy changes and any configuration adjustment. Use the same units for energy and cost throughout the file. This makes it easier to compare equipment without relying on memory or mixing estimates from different operating periods.
For a proposed replacement, record what the new unit changes and what remains unchanged. A higher rated efficiency does not by itself establish the result at a particular site: actual workload, cooling, pool settings and availability still matter. Note which values come from measured operations and which are supplier specifications. Preserve the quote and the date it was received so that later comparisons use the right assumptions.
Compare downtime over the same period
Estimate the practical effect of a repair or replacement across the same planning horizon. Include the work window, shipping or installation delay, testing time and the possibility that a repair does not resolve the fault. Fixed hosting costs may continue while a unit is idle, while output can be lower than forecast during commissioning. These factors do not guarantee one choice is preferable; they show which information must be collected before committing funds.
After the work, compare the actual return-to-service date and observed operating data with the estimate. Record the difference and update the next decision. A consistent process helps an operator learn from equipment history without assuming past performance will repeat.
