Previous
Used Antminer S19 units stacked in rows against a warehouse wall, fan grilles facing out, an old pedestal fan leaning on the stack.

Lot of 5 Refurbished Bitmain Antminer S19 95TH/s Miners – Braiins OS

Next

Lot of 10 Refurbished Bitmain Antminer S19 110TH/s Miners – Boosted, Braiins OS

Refurbished Bitmain Antminer S19 110TH/s miners stacked in dense rows against a warehouse wall, fan grilles facing out, with an old pedestal fan standing against the stack.

Lot of 100 Refurbished Bitmain Antminer S19 110TH/s Miners – Boosted, Braiins OS

Bulk pricing is by quote.

Tell us the quantity and where it is going and we will price this lot for your site — condition grade, freight, palletising and lead time included. No fixed sticker on a five-figure order that depends on all of the above.

💬 Get a bulk quote on WhatsApp  or email ns@ingmining.com · +1 781-363-1595

Lot of 100 refurbished Antminer S19 miners at 110 TH/s on Braiins OS — 11 PH/s. Boosted base S19s, so the real connected load is at least 376 kW rather than the 325 kW the stock sheet implies: a 51 kW gap that has to be caught before switchgear is specified.

Key specifications

Total lot power draw
at least 376.3 kW continuous (100 × 3,763 W)
Hashrate (per machine)
95 TH/s ±5%, SHA-256
Power draw (per machine)
3,250 W ±5%
Efficiency (per machine)
34.2 J/TH by arithmetic (3,250 W ÷ 95 TH/s); Bitmain publishes 34.5 J/TH ±5%
Input voltage (per machine)
200–277 V AC. This is not a 110 V machine and no adapter makes it one
Noise (per machine)
~75 dB from 4 fans
Full specification table

Buying from ING Mining

  • PriceQuoted, not listedWe price this on condition, quantity and freight to your address. Call or email for current price and availability.
  • FreightQuoted with the unitFreight to your address is confirmed with the price, before you pay — multi-unit and pallet orders included.
  • ConditionUsed — second-hand stockThe grade for this unit is on the listing above, and where a listing states no grade it says so.
  • GradingGraded before listingThis unit’s own grade is in the description above — where a listing states none, it says so. The six-step bench procedure behind a grade is published in full.
  • WarrantyStated per unit, not per storeAny term for this machine is in the description above. Ask for the return terms on this exact unit before you pay — full terms are in the Shipping & buying tab below.

Sold by ING Mining LLCING MINING® is a registered U.S. trademark.

Price:

Description

One hundred refurbished Antminer S19 units running 110 TH/s each on Braiins OS — 11 PH/s in total. At this scale the most expensive available mistake is planning the site from Bitmain’s stock 3,250 W. These machines are firmware-boosted: the real connected load is at least 376 kW, not 325 kW. Discovering a 51 kW shortfall after the switchgear and transformer are specified is not a cheap discovery, and it is the reason this page opens with power rather than hashrate.

The 51 kW gap, worked through

The metered draw of these units at 110 TH/s is not stated, and at this order size a made-up number would be actively dangerous. Use the floor: at the factory 34.2 J/TH, 110 TH/s requires 3,250 W × (110 ÷ 95) = about 3,763 W per machine, and since overclocking degrades efficiency the true figure sits above that. One hundred units is therefore at least 376.3 kW against the 325 kW a stock calculation returns. Specify to the metered figure with headroom on top, and ask us to meter a representative sample before you commit to the electrical design — we will send the readings in writing.

Boosted base S19s, not factory S19 Pros

Bitmain’s S19 Pro is also a 110 TH/s machine, but it achieves that with 114 chips per hashboard against 76 on the base S19, landing at 3,250 W and 29.5 J/TH. Firmware cannot add silicon; it can only run what is there harder. These are boosted base machines, which is why the load figure above is what it is. If your model assumes Pro-class 29.5 J/TH economics, it is the wrong model for this lot — and if you need genuine Pro hardware, tell us before ordering.

Antminer S19 (base model) — Bitmain's published specification

Stock hashrate 95 TH/s ±5%, SHA-256
Stock power at the wall 3,250 W ±5%
Stock efficiency 34.2 J/TH by arithmetic (3,250 W ÷ 95 TH/s); Bitmain publishes 34.5 J/TH ±5%
Input voltage 200–277 V AC. This is not a 110 V machine and no adapter makes it one
ASIC chip BM1398BB, 7 nm — 76 chips per hashboard, 228 per machine
Noise ~75 dB from 4 fans
Dimensions and weight 400 × 195.5 × 290 mm bare, 570 × 316 × 430 mm boxed. 14.2 kg net, 16 kg gross per unit — Bitmain’s published S19 specification (model 240-Ca). Freight is billed on the gross weight, not the net
Operating temperature 5–40 °C · Network: Ethernet RJ45
Generation launched May 2020, so a unit sold today is around six years old

One hundred boosted units, in site terms

Hashrate 11,000 TH/s = 11 PH/s at the tuned rate
Connected load at least 376.3 kW continuous (100 × 3,763 W)
Heat rejection at least 376 kW of heat to move. On a build this size the cooling design, not the miner supply, is the long-lead item
Freight about 1,420 kg of miners net and about 1,600 kg gross once boxed (100 × Bitmain’s published 14.2 kg net / 16 kg gross) — multiple pallets, LTL or better, and the quote is priced on the gross weight
Noise one hundred machines at ~75 dB. Site it where that is somebody’s plan rather than somebody’s complaint

The efficiency trade, in the only terms that matter at this quantity

Against the stock 95 TH/s lots you are buying about 15.8% more hashrate for more than 15.8% additional power. Whether that is a good trade is purely a function of your power price: on cheap or curtailed supply, hashrate per rack slot wins; on retail supply, the stock lots return more per dollar of electricity. That is arithmetic to run on your own numbers rather than accept from a product page — our buying guide sets out the break-even maths. It moves with difficulty and hashprice, both of which change daily, so no earnings figure appears on this page.

Spares, repair and realistic attrition

One hundred used, firmware-boosted, six-year-old machines will generate a continuous trickle of hashboard faults, and running them above their factory rating does not slow that down. Treat it as an operating cost rather than a defect: keep a spares float, run a repair bench, and stock hashboard testers, test fixtures and repair kits before you need them. Settle the DOA window and replacement policy in writing before the order is placed.

What’s in the box

Sold refurbished and tested. No condition grade, no warranty term and no statement about power supplies appears in this listing, so this page implies none of them. On a hundred-unit order the PSU question alone is a five-figure swing — get it, and the return terms, confirmed in writing before you order.

Who should buy this

Established operators with a power contract, a site and a repair capability already running. If any of the three is still hypothetical, buy five or ten first and prove the economics on your own meter before committing to a fleet.

Questions buyers ask about this lot

Will this run on a normal household outlet?

No. This lot needs a dedicated 200–277 V circuit. A North American 120 V household outlet will not run it under any adapter or configuration. Have an electrician confirm the supply before the pallet arrives.

How loud is it?

~75 dB from 4 fans — about as loud as a household vacuum cleaner running continuously, and it does not stop. This is a garage, workshop, outbuilding or hosted-facility machine, not something to put in a room where people live or sleep.

What condition is it in, and what does that grade mean?

This is second-hand stock. We describe it the way we found it: tested working means it ran on our bench at the stated hashrate, refurbished means it was also cleaned, re-pasted or had parts replaced, and as-is means it is sold without a working claim. The grade for this unit is stated in the description above; where a listing states no grade, it says so rather than implying one. Cosmetic wear — dust, scratches, label marks — is normal on hardware that has been mining and is not a fault.

What should I check the day it arrives?

Photograph the pallet or box before you open it, keep the packaging, and run the 48-hour acceptance test straight away. Freight claim windows are short, and a fault found on day two is a conversation while a fault found on day thirty is an argument. Our own six-step bench procedure is published in full, so you can repeat exactly what we did.

Shipping cost for this exact item, and how warranty and returns work here, are answered in the Shipping & buying tab on this page. Anything else, ask a person on +1 (781) 577-0377 or ns@ingmining.com — we would rather talk before you buy.

A hundred boosted machines is a power contract, not a purchase

A hundred S19s tuned to 110 TH/s is roughly 11 PH/s, and the load is what you are really buying. Filled with boosted machines a megawatt takes about 265 of them and produces 29.15 PH/s; filled with stock 95 TH/s machines it takes about 307 and produces 29.16. The two land within one per cent of each other, which is the entire point — at fixed power the boost changes how many machines you have to maintain, not what comes out of the room.

So size this lot against the service you actually have, not against the hashrate in the title. On the hashprice derived at block 961,911 on 10 August 2026 on used ASIC miner fleet economics, a used S19 fleet of either flavour breaks even at about 3.89 cents per kilowatt-hour. Below roughly four cents it wins and it is not close. Above your own break-even the payback column stops existing: there is no slow payback on a machine that loses money every hour, and no purchase price fixes it. That is why a lot like this one is behind-the-meter, curtailed-power or flare-gas hardware and not hosting hardware.

Ask us for the measured wall draw at the shipping firmware before you size a single breaker, because a 200 W per-machine error is 20 kW across a hundred. The payback table against new S21 Pro and SealMiner fleets at three, five and seven cents, and every assumption underneath it, is on used ASIC miner fleet economics.

Related hardware and where this sits in the store

Other fleet lots and single units from the same family:

Categories: Bitmain Antminer, Bitcoin miner, Used ASIC miners and Air Cooling.


Before comparing this against the 95 TH/s lots, note that a Braiins-boosted 110 TH/s S19 is less efficient than its stock rating, not more. The used-miner buying guide shows why, and what that does to your break-even power rate.

Delivery day on a lot this size is its own procedure — what to write on the freight receipt before you sign, reconciling the count and the serial numbers, why the machines come up in groups rather than together, and how to sort the ones that do not: receiving a used miner lot.




Where this listing sits

This one is on the used and refurbished shelf. It is second-hand hardware. The grade for this specific unit is stated in the description on this page — and where a listing states no grade, it says so rather than implying one.

What our grades mean

These are the only words we use, and each one means a specific thing. They are not interchangeable and we do not soften them to move stock.

NewUnused stock. Factory sealed unless the listing says otherwise — “new, open box” means exactly that, and the listing will say it.
Tested workingSecond-hand, and it ran on our bench at the hashrate the listing states. We measured it; we are not repeating what the last owner told us. Expect cosmetic wear.
RefurbishedTested working, plus work done: cleaned, thermal paste and fans dealt with, failed components replaced. The listing says what was actually done to that unit rather than using the word as decoration.
As-is / for partsSold with no working claim at all, at a price that reflects it. Buy these to repair or to harvest. If a listing says as-is, assume it does not hash until you have proved otherwise.
UntestedWe have it, we have not powered it, and we are not going to pretend we did. Rarer than as-is and always stated explicitly.

What a grade here does not mean

  • Cosmetic wear is not a fault. Dust, scratches, scuffed labels and discoloured plastics are normal on hardware that has been earning. We photograph what we have; if the pictures matter to you, ask for more and we will take them.
  • A grade is not a warranty. Warranty terms are stated per listing, never per store — see the Shipping & buying tab. A tested-working grade describes what the machine did on our bench, not what we owe you in six months.
  • A boosted hashrate is not a spec. Where a unit ships on tuned firmware we say so and we give the factory figures too, because overclocking raises power faster than it raises hashrate — a tuned machine is less efficient per terahash than stock, not more.
  • We will not invent a grade to fill a field. If a unit came in without history and has not been through the bench, its listing says that.

How the grade gets decided

A used-hardware seller who will not tell you how they test is asking you to take the grade on trust. Ours is published: the six-step bench procedure every machine goes through, and the five measurements that get recorded — per-board hashrate, chip count, measured wall power rather than the nameplate figure, hardware error rate, and temperature under sustained load. When it reaches you, run the 48-hour acceptance test straight away, while the freight claim window is still open.

Buying second-hand for the first time? How to buy a used ASIC miner without getting burned is the long version, written from the repair bench rather than from a spec sheet.

Shipping

This listing is quoted rather than priced, and freight is quoted with it. Tell us where it is going and we will confirm the shipping cost to that address with the price, before you pay — on a multi-unit or pallet order that is the only honest way to do it.

Outside the United States the checkout quotes a flat international rate. On multi-unit or pallet orders, contact us before ordering so freight to your destination can be confirmed properly. What to check when the pallet arrives, and how a freight claim works, is covered in the mining hardware shipping guide.

What to do the day it arrives

Photograph the pallet or the box before you open it, and keep the packaging until you are satisfied. Freight claim windows are short, and damage found on day two is a conversation while damage found on day thirty is an argument.

Then run the 48-hour acceptance test straight away, while that window is still open. It is the same ground our own bench covers, so you can compare like for like: how the grade on this listing was arrived at, and the five measurements we record, are set out in the Condition & grading tab on this page.

Condition, warranty and returns

Condition: Used — second-hand stock. The grade for this unit is on the listing above, and where a listing states no grade it says so.

Warranty is stated per listing, not across the catalogue. Some units carry a manufacturer term and say so above, naming the term and the date it runs from. Many units, particularly used and AS-IS stock, carry none, and those listings do not claim one. There is no blanket warranty covering a catalogue that runs from sealed factory stock to a hashboard sold for parts. For what a mining warranty normally covers, see ASIC miner warranty explained.

Returns: ask for the terms on this unit before you pay. No single returns window is published for the whole store, because the sensible terms on a tested working machine and on a board sold AS-IS for parts are not the same terms. Email or call and ask what applies to this exact item — it is a two-minute conversation and a great deal cheaper than a dispute.

Where you are going to run it

Two questions decide whether a machine like this works where you live, and neither of them is about the miner: can the circuit carry it, and what does your electricity cost? Most full-size machines here want a dedicated 240 V circuit and run at 70–75 dB continuously. A few do not — the Avalon Q publishes 45–65 dB and runs on 110–240 V — so read the voltage and noise figures on this listing rather than assuming. ASIC miner power requirements covers breaker sizing and what the install actually involves; cooling covers where the heat and the noise end up.

If your power is not genuinely cheap, hosting is the question to be asking instead of which miner to buy. Break-even electricity rate is hashprice divided by (24 × J/TH), and hashrate cancels out of it entirely — which is why a cheap old machine on expensive power loses money faster than an expensive new one does. The buying guide has the arithmetic; work out your own number before you buy anything.

If hosting is the answer, get written answers on three things before you sign: who pays for repairs, what the uptime commitment actually pays out when it is missed, and how curtailment is handled. A flat per-machine monthly rate is where most of the unpleasant surprises live. Types of mining farm compares the models and hosting versus cloud mining covers why owning the hardware is the better end of that trade. Ask us before you commit — it is a five-minute call and a great deal cheaper than a contract you cannot get out of.

What we do once it is racked, wherever that is. Deployment, firmware and monitoring are a service we run rather than a line on a page: pool and network setup, Braiins OS or VNish tuning, offline alerts, and board-level fault-finding when a machine stops hashing — ASIC mining consulting sets out what that covers. The hosting agreement itself is between you and the facility. What we can give you is the number most hosting disputes turn on: what this machine measured at the wall on our bench, so a shortfall on site is a figure to compare against rather than an argument about what it should have been doing.

Who you are buying from

ING Mining LLC, Hanson, Massachusetts. ING MINING® is a registered U.S. trademark; our marketplace feedback is on the reviews page and the longer version of who we are is on about. Reach a person on +1 (781) 577-0377 or ns@ingmining.com.