The Canaan Avalon Q is the more efficient machine and the wrong buy for almost everybody: $22.21 a terahash against the refurbished S19 XP’s $4.96. What the premium really buys is a 120 V circuit and a lower noise floor.
One new Antminer S21 Pro 245T costs the same as nine used S19k Pros once freight is counted. We ran both baskets on our own shelf prices and a chain snapshot taken this morning, and the window where the new machine is the better buy turns out to be 0.55 cents per kilowatt-hour wide.
Two used Antminers on the same shelf, $320 apart. The XP is 7% more efficient and 23% denser, and for most buyers it is still the wrong one. The arithmetic, the crossover, and the two sites where paying double is right.
Used wins only where power is cheap enough that a 34 J/TH machine still clears its bill. The two gates that decide it, a crossover calculation that needs no Bitcoin price, what actually fails on a used Antminer at our own shelf prices, and the one-line rule for when not to repair.
A boosted Antminer S19 is $130 cheaper than an S19k Pro and burns about 816 kWh a month more to mine slightly less. Break-even, payback and the two-cent crossover, worked out on both machines.
The desk miner is the efficient one, at 16.2 J/TH against 34.2. Break-even rates, the 3.75-cent crossover, the real solo odds, and which buyer each machine is actually for.
Kaspa miner vs Litecoin miner: the Antminer L9 17G breaks even at 9.15 c/kWh, the KS5 Pro at 5.32 c. And 87% of the L9 is Dogecoin, not Litecoin.
The used S19 XP returns capital faster below 5.5 c/kWh; above it the new S21 Pro wins. Break-even, payback and the 208 V trap, with the assumptions shown.
Both L9 bins run at 210 J/GH, so efficiency cannot decide between them. What can: your breaker, cost per gigahash, and the fact that 90% of the revenue is Dogecoin.
Read the BHB code on the board, not the badge on the lid. What actually fits an S19, S19 XP, S19k Pro and S21 — hashboards, APW12 vs APW17 supplies, control boards, fans and chips — and the arithmetic that decides whether the repair is worth buying.
