The 50 Percent Rule: Should You Repair or Replace Your ASIC Miner?

Your ASIC miner just threw an error code. One hashboard is dead. The fans are screaming like a jet engine with bearing failure. You are staring at a machine that cost you serious sats, and the question hits: do I fix this thing, or do I rip it out and replace it?

This is not a hypothetical. Every miner who runs hardware long enough faces this exact crossroads. And getting the answer wrong can bleed your operation dry — either by pouring money into a machine that should have been retired, or by scrapping hardware that had years of profitable hashing left in it.

The 50 Percent Rule gives you a framework. But like every rule in Bitcoin mining, the real answer requires understanding the variables underneath. Here is how to think about it like a Mining Hacker — not a spreadsheet jockey.

What Is the 50 Percent Rule for ASIC Repairs?

The 50 Percent Rule is straightforward: if the total cost to repair your ASIC miner exceeds 50% of the cost to buy an equivalent replacement, replace it. Below that threshold, repair is generally the smarter move.

“Total cost” means everything — parts, labor, shipping both ways if you are sending it to a repair shop, and any lost revenue during downtime. “Equivalent replacement” means a machine with comparable hashrate and efficiency, not necessarily the same model. If your S19j Pro dies and an S19k Pro costs the same but hashes faster, that is your replacement benchmark.

The math looks simple on paper. In practice, the 50% line is where the decision gets interesting — and where the factors below start to matter.

Why the 50 Percent Rule Is Not Enough on Its Own

The rule gives you a starting point, not a final answer. A pure cost comparison ignores critical variables that can flip the verdict entirely. Here is what the spreadsheet misses.

Efficiency Gains from Replacement

Mining hardware improves on a steep curve. An Antminer S19 (95 TH/s at ~34 J/TH) and an Antminer S21 (200 TH/s at ~17.5 J/TH) are separated by only a few years, but the S21 delivers more than double the hashrate at roughly half the energy per terahash. If your repair candidate is two or more generations behind the current flagship, the 50 Percent Rule understates the case for replacement because the new machine does not just replace your old hashrate — it dramatically improves your joules-per-terahash ratio.

Opportunity Cost of Downtime

Every day your miner sits on a bench — yours or a repair technician’s — is a day of zero hashrate. In a bull market, downtime is expensive. If repair turnaround is two to four weeks versus next-day delivery on a replacement, the lost mining revenue can push total repair cost well above 50%. Factor in your daily revenue when calculating the real repair cost.

Remaining Useful Life After Repair

Replacing a single hashboard on a three-year-old S19 might bring it back to full spec today, but what about the other two hashboards that have been running at the same temperature and duty cycle? A repaired machine with aging components can develop cascading failures. Ask yourself: how many more months of reliable operation do I realistically expect after this repair?

The Variables That Actually Drive the Decision

When the 50 Percent Rule lands you in the gray zone — repair costs between 30% and 70% of replacement — these factors break the tie.

1. Machine Efficiency (J/TH)

This is the single most important metric. A miner’s joules-per-terahash ratio determines whether it remains profitable as network difficulty climbs. In 2026, with post-halving block rewards at 3.125 BTC and difficulty at all-time highs, efficiency separates survivors from e-waste.

Legacy machines like the S9 are almost never worth repairing for pure mining profitability in 2026 — but they are absolutely worth repairing if you are converting them into Bitcoin space heaters. When the heat output is the primary value and the sats are a bonus, the economics change completely.

2. Parts Availability

This is where many miners get stuck. Current-generation hashboards, control boards, and fans are readily available. But try sourcing a hashboard for an Innosilicon T2T or a Whatsminer M21 in 2026 — good luck. When parts are scarce or discontinued, repair costs spike from supply constraints alone, often pushing you past the 50% threshold regardless of labor costs.

D-Central maintains one of the largest ASIC parts inventories in Canada, including hard-to-find components for legacy and mid-generation machines. But even with deep inventory, some parts simply stop being manufactured.

3. Electricity Cost

Your power rate determines the profitability floor for any machine. At $0.05/kWh, an S19j Pro (104 TH/s, ~29.5 J/TH) can still turn a profit in 2026. At $0.12/kWh, that same machine is underwater, and repairing it is throwing good money after bad. Use a mining profitability calculator with your actual electricity rate before committing to any repair.

4. Network Difficulty Trajectory

Bitcoin’s mining difficulty adjusts every 2,016 blocks (~two weeks). In 2026, difficulty has been trending upward as institutional miners deploy next-gen hardware at scale. This relentless difficulty increase compresses margins for older machines. A repair that makes sense at today’s difficulty might not pencil out six months from now when difficulty is 15–20% higher.

5. Bitcoin Price and Market Cycle

Market conditions affect the repair decision in two ways. First, high BTC prices increase the opportunity cost of downtime (every offline hour costs more sats). Second, high prices inflate the used hardware market, which raises both replacement costs and the residual value of your current machine. During bear markets, replacement hardware is cheap and plentiful — sometimes buying a used machine is cheaper than repairing your broken one.

6. Your Repair Capabilities

If you can diagnose and replace a hashboard yourself, your repair costs are parts-only — dramatically lower than sending the machine to a shop. Home repair capability shifts the 50% line heavily in favor of repair. If you are paying for professional diagnostics, labor, and round-trip shipping, those costs stack up fast.

When to Always Repair (Regardless of the 50% Rule)

Some situations override the math entirely:

• Current-gen machines with minor faults. A fan failure or firmware glitch on an S21 is a no-brainer repair. The machine has years of competitive life ahead.
• Space heater conversions. If the miner’s primary job is heating your home and the sats are a bonus, repair economics are completely different. A $150 hashboard repair on an S9 space heater that saves you $200/month in heating costs pays for itself in weeks.
• Machines with custom firmware. If you are running BraiinsOS, LuxOS, or other aftermarket firmware that you have tuned for your specific power and thermal environment, there is real value in keeping that specific machine running rather than starting over with a new unit.
• Warranty-covered repairs. If the manufacturer or reseller covers the repair, the cost calculation is irrelevant — get it fixed.

When to Always Replace (Regardless of the 50% Rule)

Some machines should be retired, even if repair costs are technically under 50%:

• Machines older than five years with no heat recovery use. An S9 in 2026 operating purely as a miner — not a heater — is almost certainly unprofitable at any reasonable electricity rate. Repairing it to keep mining is sending good money after bad.
• Machines with multiple simultaneous failures. Two dead hashboards plus a failing PSU on a mid-gen machine signals systemic end-of-life, not a one-off component failure. Even if repair cost is under 50%, the probability of another failure within months is high.
• When replacement delivers a 2x+ efficiency gain. If replacing an S19 (34 J/TH) with an S21 (17.5 J/TH), you are not just restoring hashrate — you are cutting your power consumption in half for the same output. The efficiency delta makes the repair cost comparison misleading because the replacement is a fundamentally better machine.

The D-Central Approach: Repair with Intelligence

At D-Central Technologies, we have repaired thousands of ASIC miners since 2016. Our approach is not “fix everything” or “replace everything” — it is diagnose first, advise honestly, and repair only when it makes economic sense for the miner.

Here is what that looks like in practice:

1. Full diagnostic before any quote. We identify every issue — not just the obvious one. A dead hashboard might be the symptom; the root cause might be a failing PSU that took the hashboard with it. Fixing only the hashboard means the PSU kills the replacement board next.
2. Transparent cost breakdown. Parts, labor, and estimated turnaround — no surprises. We will tell you upfront if a repair does not make economic sense.
3. Second-life recommendations. If your miner is not worth repairing for pure mining, we will tell you whether it is a candidate for a space heater conversion, whether the parts have salvage value, or whether it should be responsibly recycled.

We are Bitcoin Mining Hackers. We do not waste your sats on repairs that do not pencil out, and we do not scrap machines that have life left in them.

A Step-by-Step Decision Framework

Use this checklist before committing to repair or replacement:

1. Get a repair quote — either self-diagnose or send it to a professional.
2. Price the replacement — check current market prices for an equivalent or better machine.
3. Calculate the ratio — repair cost divided by replacement cost. Under 50%? Lean toward repair.
4. Check the efficiency gap — is the replacement significantly more efficient (J/TH)? If yes, that shifts the needle toward replacement even if repair is cheaper.
5. Factor in downtime — add lost revenue during repair turnaround to the repair cost.
6. Assess remaining life — will the repaired machine run reliably for at least 12 more months?
7. Run the profitability math — plug your electricity rate and the machine’s specs into a mining profitability calculator. If the machine is unprofitable even after repair, neither option makes sense for pure mining.
8. Consider heat recovery — if the machine can serve as a space heater, the economics change dramatically in favor of repair.

Frequently Asked Questions