Used Antminer S19k Pro 115TH/s – Tested, Braiins OS Boost to ~125TH/s

Description

The Antminer S19k Pro 115TH/s is a high-performance ASIC designed for SHA-256 mining. It delivers a stable 115TH/s hashrate and carries a firmware tune advertised at up to 125TH/s. Worth being precise about what that buys you: overclocking raises power faster than it raises hashrate, so a boosted unit is less efficient per terahash than the stock 23 J/TH, not more. You are buying hashrate, not efficiency. This S19k Pro miner 115TH/s is fully tested and ready for deployment in professional mining environments. It includes a PSU and is suitable for long-term Bitcoin mining operations.

What’s in the box

• 1x S19k Pro Bitcoin miner 115TH/s
• Braiins OS installed for optimization
• Boosted performance up to ~125TH/s (at a worse J/TH than stock)
• Power Supply Unit (PSU) included

Condition of This Bitcoin Miner

• Used but fully tested S19k Pro unit
• May show cosmetic wear from mining use
• Performance remains stable and reliable
• Ready for immediate mining setup

Antminer S19k Pro Performance Details

Model	S19k Pro Bitcoin Miner
Algorithm	SHA-256
Coins mined	Bitcoin (BTC)
Hashrate	115TH/s (boosted ~125TH/s)
Power consumption	2,645 W ±5% at the stock 115 TH/s; more on the boosted profile
Efficiency	23 J/TH ±5% at stock, worse on the ~125 TH/s tune
Input voltage	200–240 V AC only, 20 A — not a 110 V machine
Noise	75 dB
Condition	Used, fully tested on our bench. Cosmetic wear consistent with mining service
Power supply	PSU included
Size and weight	400 × 195 × 290 mm bare, 570 × 316 × 430 mm boxed. 12.85 kg net, 14.5 kg gross — Bitmain’s published S19K Pro specification. Bitmain publishes that table against the 120 TH/s bin; the chassis is common to the S19K Pro bins. Freight is quoted on the gross weight, not the net

Important Mining Information

The S19k Pro is designed for industrial use, is rated at 75 dB, and takes 200–240 V only — it will not run from a North American 110 V household outlet under any configuration. It needs a properly ventilated location and a 240 V circuit. If the machine has to live where people do, read our Avalon Q versus Antminer S19k Pro comparison first; the Avalon Q is the one that runs on household voltage. Performance may vary depending on power supply, cooling system, and firmware configuration of the S19k Pro ASIC miner.

Shipping & Delivery

Each S19k Pro unit is securely packed and shipped with tracking. Please ensure correct delivery details before ordering.

Useful Resources

Learn about Bitcoin mining: Bitcoin Official Website Explore more mining hardware: ASIC Miner Collection Firmware optimization tool: Braiins OS Official

Final Note

The S19k Pro Bitcoin miner is a reliable choice for miners looking for strong performance, tested hardware, and optimized efficiency for long-term mining operations.

The other S19k Pro on this shelf costs the same $379.99 — the honest comparison

We have two used Antminer S19k Pros on the floor at the same price. This one, at 115 TH/s with Braiins OS installed, and a 120 TH/s unit running VNish — same model, same 23 J/TH, same $379.99. Neither the price tag nor the spec sheet separates them, so here is what actually does.

	115 TH/s, Braiins OS	120 TH/s, VNish
Price	$379.99	$379.99
Factory rating	115 TH/s at 2,645 W	120 TH/s at 2,760 W
Efficiency, stock	23 J/TH	23 J/TH
Firmware as it ships	Braiins OS+	VNish
Tune advertised on the listing	~125 TH/s	130 TH/s
Condition grade stated	Used, tested working on our bench	None stated — ask us in writing before you pay
Gross at 2026-08-09	$3.68/day	$3.84/day
Break-even power price	5.80¢/kWh	5.80¢/kWh
Voltage / noise	200–240 V only, 75 dB — identical machines in this respect, and neither belongs in a house

Both machines stop earning at exactly the same electricity price

At BTC $64,927 and network difficulty 127,479,855,693,691 (2026-08-09, block subsidy only —
transaction fees were adding about 0.3% on the day and are left out to keep this
conservative), a terahash of SHA-256 grosses $0.0320 a day. Run that
against each machine’s own wall draw and both land on the same number:
5.80¢/kWh. They land there because they are the same silicon at the
same 23 J/TH — the break-even rate is efficiency divided into hashprice, and the
hashrate cancels out. These figures move every day with price and difficulty.
Re-run them against your own rate before you buy anything.

The 120 TH/s unit earns about $58 a year more, for the same money

Five extra terahash at today’s hashprice is $0.16 a day, and
because the efficiency is identical those terahash cost the same to run. On the
arithmetic alone, at an identical $379.99, the VNish 120 TH unit is the better buy
and we are not going to pretend otherwise on our own shelf. What the 115 TH/s
listing has instead is a stated condition grade: it is described as
used and tested working on our bench, and the 120 TH listing says plainly that it
states no grade. At the same price that is the whole trade — $58 a year
against a written test result.

Braiins OS or VNish: the part that is actually about firmware

Both are real, both arrive on these units already installed and configured,
and both take a cut. Braiins publishes a 2–2.5% dev fee depending on
model, collected by pointing that share of your hashing power at their pool; Braiins
also states that mining on Braiins Pool with Braiins OS gives a 0% effective pool
fee instead of their standard 2.5%, so on that pool the fee is offset rather than
paid twice. It is the only firmware here with native Stratum V2.
VNish does not publish a single fee figure on its own site; the
independent firmware guides put the S19-family dev fee at 2–2.8%. Ask us to confirm
the installed version and its fee setting on the actual unit before you pay — on a
used machine that arrives pre-flashed, that setting is a running cost you inherit,
and roughly 2.5% off the top is about 0.15¢/kWh of your break-even headroom gone.
No listing on this site said that before today, ours included.

The difference that decides the purchase is not the fee, it is the direction each
one tunes. Braiins OS+ autotunes to a power target you set — you tell it
a wattage ceiling and it finds the best hashrate inside it. VNish tunes to a
hashrate target, with per-domain voltage and per-chip frequency control and
more aggressive overclock profiles. One is built to push a machine down, the other
to push it up.

When the boost is worth running, in one line

Overclocking never improves efficiency — power climbs faster than hashrate on the
voltage-frequency curve of every ASIC ever made, so a boosted unit is worse
than 23 J/TH, not better. Both of our listings already say so. What neither said is
where the line falls, and it is one division: any terahash an overclock adds
has to clear the same J/TH bar the whole machine has to clear — hashprice
divided by 24 hours times your rate.

Your all-in power rate	Bar the added terahash must beat
3.0¢	44.5 J/TH
4.0¢	33.4 J/TH
5.0¢	26.7 J/TH
5.8¢	23.0 J/TH — your break-even; the bar is now the machine itself
7.0¢	19.1 J/TH
12.0¢	11.1 J/TH

Read the bottom of that table carefully. At your break-even rate the bar
is the machine’s own stock 23 J/TH, so above it an overclock would have to
be more efficient than the machine is at stock — which is arithmetically impossible.
Below about 4¢ there is real headroom and the boost pays; above about 5¢ it
cannot. That is the whole case for each firmware: if you are paying under
roughly 3.5¢ — comfortably inside that headroom — you want the machine tuned up, and
if you are paying more than that you want it tuned down.

The gate that applies to both: 612 days

The next Bitcoin halving is block 1,050,000. The chain was at block 961,818 on
2026-08-09, which is 88,182 blocks — about 612 days, roughly 20
months at ten minutes a block. Block revenue halves that day; the power bill
does not. Holding price and difficulty still, the break-even on both of these
machines goes 5.80¢ → 2.90¢/kWh. So the real question is not what
either machine earns, it is whether it returns $429.98 — $379.99 plus $49.99
US freight — before that date.

Payback on this unit, after its dev fee, at today’s numbers:

Your power rate	Time to return $429.98
3¢/kWh	255 days (8 months)
4¢/kWh	409 days (13 months)
5¢/kWh	1033 days (34 months) — past the halving
6¢/kWh	never — it runs at a loss

So which one should you buy?

• Under about 3.5¢/kWh, and you want maximum hashrate: take the
  120 TH/s VNish unit. Five more terahash for the same money at the
  same efficiency, and its firmware is the one built to push. Ask us in writing for its
  condition grade first — that listing does not state one, deliberately.
• Between about 3.5¢ and 5¢/kWh: take the
  115 TH/s Braiins unit. At that rate you will be tuning the machine
  down, not up, and a power-target autotuner is the right tool for it. The stated
  test result is worth more than $58 a year of gross when margins are that thin.
• Above about 5¢/kWh: buy neither. Payback runs to 34 months, which
  is well past the halving that cuts the revenue in half, and after it a 23 J/TH machine
  needs 2.90¢ power to break even at all. That is not a machine purchase, it is a
  donation. Spend the money on a current-generation miner instead, or don’t spend it.
• If it has to live where people live: neither. Both are 75 dB and
  200–240 V only, and no adapter changes that. The
  Avalon Q versus S19k Pro
  comparison is the one to read — the Avalon Q runs on household voltage at about
  45 dB.

Everything above is arithmetic on public numbers, and you can redo it: the
used-miner buying guide
carries the same break-even method for every used S19 variant we stock, and the
used ASIC miner shelf
lists what else is on the floor. If you want the measured draw and the measured hashrate
on the specific unit rather than the number in the title, ask — we bench them, so we have
it. +1 (781) 577-0377 or
ns@ingmining.com.

There is a $249.99 S19 on the same shelf — here is when it beats this one

We stock a refurbished Antminer S19 boosted to 110 TH/s for $249.99, and it is $130 cheaper than this machine for almost the same headline hashrate. On our own numbers it is the better buy in exactly one situation, and we would rather set that out than let the price tag decide it for you.

This S19k Pro makes 115 TH/s on 2,645 W, which is 23 J/TH and a break-even of 5.86¢/kWh. The boosted S19 makes 110 TH/s on at least 3,763 W — 34.2 J/TH at absolute best, and a break-even of 3.94¢/kWh. Nearly two cents of headroom is the difference between a machine that survives a hosting contract and one that only works behind your own meter.

The crossover is 2.04¢/kWh. Below it the cheaper machine returns its purchase price first. Above it, this one pays back sooner at every single rate — 250 days against 396 at 3¢/kWh — and clears more money in total at every rate, including below the crossover. This one also ships with its PSU and sits comfortably on a 20 A circuit at 240 V, where the S19 wants 30 A and its listing does not state whether a PSU is included at all.

At $32.33/PH/day, BTC $65,231, difficulty 127,479,855,693,691, 10 August 2026, after a 2.5% firmware dev fee, landed cost including $49.99 US freight. These move daily.

Both machines and the full payback tables are in Antminer S19 vs S19k Pro.

And the $699.99 S19 XP above it — when paying double is right

The machine one step up our used shelf is a refurbished Antminer S19 XP at $699.99, 141 TH/s at 21.35 J/TH. It is 7% more efficient than this one and costs 84% more, and that trade is worse than it sounds: per terahash you would be paying $4.96 against $3.30 here. Buy the same total hashrate either way and the two fleets earn exactly the same revenue, so the price of Bitcoin drops out of the comparison entirely and what is left is $1.66 more capital per TH/s against 1.21 kWh per TH/s per month saved — roughly 46 months to break even at 3¢/kWh.

The XP earns its premium in exactly two situations. If your machine positions are fixed rather than your budget, it puts 22.6% more hashrate on the same 20 A, 240 V circuit. And if your power sits between about 2.8 and 3.0¢/kWh, the halving in roughly 608 days takes this machine below its break-even and leaves the XP above its own. Outside those two cases, this is the machine to buy. The full comparison is in Antminer S19k Pro vs S19 XP.

The $4,200 question: nine of these, or one new S21 Pro?

The most common cross-shop on this machine is not another used unit — it is a new Antminer S21 Pro 245T at $4,200. Worked as a fixed budget rather than as a spec sheet, that is what the two really cost:

	Nine of these	One new S21 Pro 245T
Goods	$3,419.91	$4,200.00
Freight at $49.99 a unit	$449.91	$49.99
Landed	$3,869.82	$4,249.99
Hashrate	1,035 TH/s	245 TH/s
Landed cost per TH/s	$3.74	$17.35
Wall power	23.8 kW	3.7 kW
20 A / 240 V circuits	9	1
Hashboards in service	27, used	3, new

Two things that table does not say out loud. Freight eats two machines: eleven of these at the sticker price becomes nine once the pallet is paid for, which is why the honest capital cost of this miner is $3.74 per terahash and not the $3.30 the price tag implies. And the verdict flips on electricity rather than on hardware — under roughly 5.1 cents a kilowatt-hour the nine win on net dollars as well as on hashrate; above this machine’s own break-even the nine are not the cheaper option, they are not an option at all. We work the whole thing through, with the crossover, the freight and the spares shelf costed.

Buying more than one or two? The ten-machine lot of this exact unit is quoted rather than listed — on condition, quantity and freight to your address.

Related hardware and where this sits in the store

Other used and refurbished units currently on the floor:

• Used Antminer S19k Pro 120TH Bitcoin Miner – VNish 130TH Tune
• Lot of 10 Used Bitmain Antminer S19k Pro 115TH/s Miners – Tested, Braiins OS, PSUs Included
• Bitmain Antminer S19K Pro 120T – Efficient Bitcoin Miner

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

At 23 J/TH this is one of the few used machines that still clears realistic hosting rates, which is the test that matters more than the terahash figure. The used-miner buying guide has the break-even table for every used S19 variant.