Solo Mine Bitcoin With an Old Antminer S9 or S17

On April 9, 2026, a single mining machine — roughly the power of one 2019-era Antminer S17+ that big farms threw away years ago — solved Bitcoin block 944,306 and walked off with 3.128 BTC, about $222,000. No farm. No data center. One old box, running solo, beating odds of roughly 1-in-100,000 on the day it won. This is not a fairy tale. It is the single best argument for doing something the entire mining industry told you was pointless: pointing a cheap, “dead” Antminer at the Bitcoin network and pulling the lever.

Here is the 30-second answer. Yes, you can absolutely solo-mine Bitcoin with an old Antminer S9 or S17 in 2026. No, it will not “pay for itself” the way mining did in 2017 — at normal electricity prices these machines lose money on raw economics. What they buy you instead is three things at once: a real, life-changing lottery ticket, a space heater that happens to mine, and a vote for Bitcoin’s decentralization. For a $30–$300 machine, that is a remarkable bundle. Let’s break down exactly how it works.

The wins that keep the dream alive

Solo block finds used to be once-a-year curiosities. In 2026 they are routine. Through a rolling 12-month window, solo miners on CKPool alone found about 22 verified blocks — roughly one every 18 days, distributing well over 69 BTC. These were not warehouses. They were individuals, many running a single ASIC at home.

The standouts read like tall tales:

• The S17+ jackpot (April 2026): ~70 TH/s solved block 944,306 for 3.128 BTC (~$222K). Mathematical expectation to find a block at that hashrate: around 300 years. It hit in months.
• The S9-class miracle (November 2025): a solo miner contributing just 6 TH/s — roughly an aging S9 — found a block. CKPool’s creator pegged that miner’s odds at about 1 in 170 million on any given day. They won anyway.
• The $281K block (December 2025): block 928,985, 3.128 BTC, found solo.

The point is not that any single machine is likely to win. It is that every block is a fresh, independent draw, and ordinary people keep drawing the winning ticket. That is the nature of a lottery — and old Antminers are the cheapest tickets in the building.

Why “obsolete” hardware is the perfect lottery ticket

Here is the counterintuitive truth at the heart of this: when you solo-mine, the block reward is the same whether you find it with an old S9 or a brand-new S23. A block is a block. The network does not pay you more for using expensive hardware — it pays whoever solves it, full stop, 3.125 BTC plus fees.

So the real question is not “how powerful is my machine?” It is “how cheap is my ticket, and how many tickets can I afford to leave running?” An old Antminer answers that beautifully: it costs almost nothing to acquire, and it produces real hashes against the real network. You are buying lottery tickets at a steep discount — and unlike a scratch card, the ticket keeps playing 24/7 and warms your room while it does.

S9 vs S17: the spec sheet

These two machines define the budget end of the used market. Here is what you are actually getting:

To put efficiency in context: the S17 is about 2.1× more efficient than the S9, but still 1.3× less efficient than an S19 and 2.6× less than an S21. By 2026 standards these are thirsty machines — but for a few hundred dollars of hashpower that you intend to run as a lottery ticket and a heater, “thirsty” is a feature, not a bug. One caution on the S17: it earned a reputation for reliability quirks — hashboard solder joints and thermistor failures are the usual suspects — so buy from a seller who tests and, ideally, reinforces the boards.

The honest economics: this loses money (and that’s the point)

Let’s not insult your intelligence. At normal electricity prices, an old Antminer mining Bitcoin runs at a loss on paper. An S17+ pulls about 2,920 W — at $0.10/kWh that is roughly $7 a day in power, against pennies of expected daily reward. The raw spreadsheet is red.

So why do tens of thousands of people do it anyway? Because the spreadsheet is missing two columns:

• The lottery column. Every day you run, you hold a real ticket to ~$200K+. The expected value is tiny, but the payout is life-changing and the downside is a known, small electricity bill. People pay far more for far worse odds at an actual lottery counter.
• The heat column. An S17 throws off about 8,600 BTU/h — equivalent to a mid-sized space heater. If you would otherwise be running electric heat in winter, that electricity is not wasted. It is doing double duty: heating your room and securing Bitcoin. D-Central even sells purpose-built “Space Heater Edition” cases for exactly this. Suddenly the machine runs “for free,” and any block it finds is pure upside.

Run it as a heater that happens to mine, in a cold room, on the cheapest power you can find — solar, off-peak, a basement you were heating anyway — and the economics flip from “loss” to “free lottery.” That reframe is the entire game. (If heat management is new to you, our cooling guide covers airflow and noise.)

Firmware: stock, or unlock it on GitHub

An old Antminer will mine on its stock Bitmain firmware, but the open-source firmware scene can make it quieter, cooler, more efficient, and frankly more fun to own. This is also where the “it’s open hardware” magic lives — you are not locked into anything. The main options in 2026:

• Braiins OS+ — the veteran. Open-source code with per-chip auto-tuning, a low-power efficiency mode, and native Stratum V2. It supports the S9 (S9/S9i/S9j) and the entire S17/T17 family, and you can build the open-source edition with the —open-source flag. Docs and source live on GitHub at tgr-braiins/braiins-os-docs. There is a 2% dev fee on the plus edition, which you can offset by mining on Braiins’ own pool.
• VNish — the aftermarket workhorse, with fast boot, self-tuning, and up to ~30% power savings. Crucially, it supports Antminers with Amlogic control boards, which Braiins OS+ does not.
• DCENT_OS — the newcomer worth watching: a free, fully open-source firmware written in Rust, with 0% dev fee by default, per-chip auto-tuning, Stratum V2, Home Assistant integration, and a genuinely clever PID space-heater mode that throttles hashrate to hold your room at a target temperature. It is in beta on the S9 (BM1387) with S17/S19/S21 support planned.

One critical compatibility check before you flash anything: the S17/T17 series shipped with two different control boards — Xilinx (Zynq) and Amlogic. Braiins OS+ supports the Xilinx board only; for an Amlogic board you must use VNish (or swap the board). Flash the wrong firmware and you can brick the miner. Always identify your control board first. The S9’s standard control boards are broadly supported.

What are your actual odds?

Honesty is the whole brand here, so let’s be precise. At today’s Bitcoin network size — about 895 EH/s, with difficulty having just dropped ~10% to 124.93 trillion — here is roughly how often each old machine would find a BTC block solo, on average:

Read the “daily odds” column and then remember: the S17+ that won in April was facing exactly those ~1-in-90,000-ish daily odds, and it hit in months. The “average time” is a long-run statistic, not a countdown — variance is the whole story, and variance cuts both ways. We unpack the real math behind this in our Poisson variance guide, and put the odds in plain language in can a tiny miner find a block?

The move that actually changes your odds

Here is the part most “old Antminer” articles never tell you. You do not have to point your S9 or S17 at Bitcoin. The same SHA-256 hashpower works on smaller chains where the difficulty is a tiny fraction of Bitcoin’s — Bitcoin Cash, eCash (XEC), and others. On those networks, your old machine’s odds go from “centuries” to genuinely realistic: days to weeks, not lifetimes.

That is the entire reason SoloFury is multi-coin. Our non-custodial pool lets you point one old Antminer at BTC, BCH, BC2, BCH2 or XEC, with the full block reward paid straight to your own wallet at a 1% fee. Want to see which chain is easiest to hit right now? The live Network Radar shows real-time difficulty and retarget countdowns for every supported coin. For the strategy of which coin to chase with limited hashpower, read the best coins to solo mine in 2026 and our coin selection guide. And if BCH is your target, we have a dedicated how to solo-mine BCH walkthrough.

Setup: from dusty box to hashing in minutes

The flow is the same as any Antminer:

1. Plug in the PSU and Ethernet, power on, and find the miner’s IP on your network.
2. (Optional) Flash Braiins OS+, VNish, or DCENT_OS for tuning, quieter operation, and Stratum V2.
3. Open the dashboard, enter your pool’s stratum URL, and set your own wallet address as the worker — that is where a winning block pays out.
4. Save, confirm it is accepting shares, and watch your best-ever share difficulty climb.

Our solo start guide covers the pool side end to end, and the Antminer setup guide walks through the dashboard and worker configuration that applies directly to an S9 or S17.

Verdict: should you do it?

Do it if you find the idea of a $50 machine that could pay out $200,000 irresistible, you have cheap or otherwise-spent electricity (especially heat you would run anyway), and you care about being one more independent node keeping Bitcoin decentralized. It is the most fun you can have with “obsolete” hardware, and the SHA-256 altcoin route gives you genuinely catchable odds.

Skip it if you expect reliable profit at retail power prices — that is not what this is. And if you want quiet, efficient, always-on hashing for a desk, a modern Bitaxe or the Nexus S1 is the better tool. The old S9/S17 is a heater, a lottery ticket, and a statement — buy it knowing that, and it is one of the best deals in Bitcoin.

Plan your old-rig setup with SoloFury’s free tools

• Solo mining calculator — real odds and break-even for your exact S9/S17 hashrate and power price.
• Profitability tool — compare expected return across BTC, BCH, BC2, BCH2 and XEC.
• Network Radar — live difficulty and retarget countdowns, so you aim your old rig at the easiest chain.
• Start mining — point your Antminer at SoloFury in about two minutes, payouts straight to your wallet.

Frequently asked questions

Can you still solo mine Bitcoin with an Antminer S9 in 2026?
Yes. An S9 produces about 13.5 TH/s of real SHA-256 hashpower and can solo-mine through a pool like SoloFury. Its odds of finding a BTC block are extremely long (averaging over a thousand years), but it remains a valid lottery ticket — and the odds are far better on smaller SHA-256 chains like BCH or XEC.

How much does an old Antminer S9 or S17 cost?
On the 2026 secondary market, an S9 runs about $30–$50 and an S17/S17+ about $100–$300, depending on condition and whether the hashboards have been tested or reinforced.

Is solo mining with an old Antminer profitable?
Not on raw economics at normal power prices — they typically run at a loss mining BTC. The value comes from the lottery upside, using the heat for home heating, and supporting decentralization. With cheap or already-spent electricity, the effective cost drops toward zero.

What firmware should I use on an old Antminer?
Braiins OS+ (open-source, auto-tuning, Stratum V2; supports S9 and S17/T17 with Xilinx boards), VNish (supports Amlogic boards, up to ~30% power savings), or the newer open-source DCENT_OS (0% dev fee, PID space-heater mode). Always check your control board type before flashing.

What are the real odds of an old Antminer finding a block?
On Bitcoin: an S9 averages ~1,260 years, an S17 ~305 years, an S17+ ~233 years per block — but variance means people win far sooner (an S17+ won a $222K block in April 2026). Odds improve by orders of magnitude on smaller SHA-256 chains.

Can an old Antminer heat my home?
Yes. An S17 outputs about 8,600 BTU/h — comparable to a mid-sized electric space heater. In cold climates that heat offsets your existing heating bill, effectively making the mining “free.”