Antminer S9 Price Guide: What Used Units Cost Now

The Antminer S9 ships rated 13.5 TH/s, 1,350 W at the wall, about 100 J/TH efficiency, SHA-256 algorithm, air cooling. Those four numbers — hashrate, wall draw, efficiency and cooling method — decide everything else: what circuit you need, how much heat the room absorbs, and how the machine behaves when network difficulty climbs. Everything a spec sheet lists beyond that is packaging.

Treat the rated figures as a ceiling rather than a promise. Air-cooled units typically settle two to five percent below datasheet output once intake air passes 30 degrees Celsius, while liquid-cooled siblings hold their numbers far more consistently. We bench every Antminer S9 before it ships and record the actual hashrate and wall draw on the test sheet that travels with the unit.

At 1,350 W continuous the Antminer S9 consumes roughly 32 kWh per day. At six cents per kilowatt hour that is about $2 of power per day; at twelve cents it doubles to roughly $4. Multiply by 30 and you have the only fixed cost that never negotiates.

Because power is the dominant cost, your electricity rate — not the purchase price — decides whether this machine belongs in your fleet. Run the numbers with hashprice cut by 30 percent before you buy. If the Antminer S9 still clears its own power bill in that stressed scenario, the purchase survives a bad quarter; if it does not, you are betting on price appreciation rather than on mining.

Payback in months equals delivered machine cost divided by monthly net profit, where net profit is mining revenue minus electricity. Revenue depends on hashprice, which moves daily with bitcoin price and network difficulty; electricity is fixed and known. Build the model in a spreadsheet with hashprice as the single variable and you will never be surprised by a market move again.

Model three scenarios for the Antminer S9: current hashprice, hashprice down 30 percent, and hashprice down 50 percent. If the middle case still shows positive cash flow, the purchase is defensible. If only the optimistic case works, either negotiate the price down or buy fewer units and keep the reserve.

Second-hand Antminer S9 hardware is where the best dollars-per-terahash usually live, but the risk profile shifts to you. Insist on a load test at full power for at least an hour, verify all hashboards report the expected chip count, check for corrosion on the board edges and PSU connectors, and look for replaced fans or reflowed chips — signs the unit has been repaired before.

Price used units at a level you would be comfortable writing off inside twelve to eighteen months. Every refurbished machine we sell is stripped, cleaned, load-tested and re-benched, and we publish the measured numbers rather than the datasheet numbers so you can price the risk honestly.

Ask five questions before money moves: what the measured hashrate and wall draw were on the test bench, which firmware version is installed, whether hashboards are original or replacements, what the warranty covers in writing, and who pays return shipping if the unit arrives dead. A supplier that answers all five in writing is a supplier worth buying from.

Refuse pressure tactics, avoid escrow-free peer-to-peer deals for machines you cannot inspect, and confirm delivered pricing including freight and duty. Our Antminer S9 stock is bench-tested and hashrate-verified from our Hong Kong facility, with worldwide shipping to the USA, Canada and Europe and written warranty terms.

If you want a cheap entry into SHA-256 hashing, compare against used S19-series machines at 21-30 J/TH — they cost more up front and earn far more per watt. Do not compare on headline hashrate alone; two machines with the same output can differ by 40 percent in electricity cost, which is the entire margin at most power rates.

Our sales desk will model two or three candidates against your actual electricity rate, available amperage and budget, then quote delivered pricing so you can compare like with like. That takes an afternoon of spreadsheet work off your plate and usually changes the shortlist.