What Is BTCB2? The Bitcoin BLAKE2b Fork, Explained for People Who Own Machines

On 8 August 2026 Bitcoin split, and for the first time in its history the minority chain did not keep SHA-256. It kept the ledger and changed the algorithm. That single decision is why a $12,000 Antminer cannot mine BTCB2 at all, and why a Goldshell that had been sitting unplugged in a garage since the Siacoin days suddenly can.

Most coverage of BTCB2 (Bitcoin BLAKE2b) has been written for traders, which means it is about price. This one is written for the person who owns machines, because for a hardware owner the interesting question is not what the coin is worth today — it is what the fork actually changed at the protocol level, which hardware it made relevant, and what can still go wrong.

Here is the whole thing, in order, with the uncomfortable parts included.

First, the stakes: what actually happened on 8 August

Bitcoin forks are usually governance arguments that end in a chain split. This one was no different in shape, only in what the losing side did next.

The dispute was over BIP-110, a proposal associated with Luke Dashjr's group to restrict the storage of non-monetary data in Bitcoin transactions — the long-running argument about whether Bitcoin's block space should be for payments or for arbitrary data. When it came to a head, roughly 97–98% of SHA-256 hashrate declined to enforce the change. The minority that wanted it did what minorities in Bitcoin have always done: it split away, carrying a full copy of the ledger up to block 961,632.

What happened next is the unusual part. Instead of continuing on SHA-256 as a tiny, defenceless fraction of Bitcoin's hashrate, the new chain changed its proof-of-work. At block 961,640 on 30 August 2026, 06:14 UTC, BTCB2 activated BLAKE2b and reduced its block size to suit the hardware that runs it.

There is a hard-headed reason for that choice. A minority chain that keeps the majority's algorithm is permanently at the mercy of it: any large SHA-256 operator can point spare hashrate at the small chain and rewrite its history for the cost of an afternoon's electricity. Change the algorithm and that attack requires hardware the attacker does not own and cannot quickly buy. Switching to BLAKE2b did not just pick a hash function — it moved the chain behind a completely different wall.

Bitcoin and BTCB2, side by side

What was inherited, and what was replaced.

Why a six-year-old Siacoin miner is the right key

BLAKE2b was not chosen at random. It is the algorithm Sia has used since 2015, which means a mature installed base of ASICs already existed — idle, depreciated, and largely unemployed, because mining Siacoin directly had been marginal to negative for years.

Those machines are the ones in our BTCB2 collection. Six models, from the 950 W Goldshell SC Lite at 4.4 TH/s to the 19 TH/s iBeLink BM-S3. Put together, those six units represent roughly 77 TH/s — on a network reported near 5.1 PH/s in early September, that is about one and a half percent of the entire chain from a single shelf.

Supply of new Blake2b hashrate cannot respond quickly either. Nobody has been building Blake2b ASICs at volume, because until August there was no reason to. The hashrate that has arrived so far is overwhelmingly repurposed Sia hardware and rented capacity — and it is arriving fast.

How you actually mine it

The practical setup is closer to Siacoin mining than to Bitcoin mining, because mechanically it is Siacoin mining pointed somewhere else.

1. Use compatible hardware. Blake2b ASICs only. Stock Sia firmware works on compatible machines such as the iBeLink BM-S3 — there is generally no exotic flashing step.
2. Point it at a pool that supports BTCB2. B2Pool serves BLAKE2b hardware over Stratum and adjusts worker difficulty once shares start arriving. The alternative is running your own Bitcoin Knots node paired with a BLAKE2b DATUM gateway.
3. Set the payout address carefully. A BTCB2 address is not a Bitcoin address in practice even where the format looks familiar. Confirm the wallet you are paying into actually supports the chain.
4. Verify the worker is real. Stable connection without constant reconnects, accepted rather than rejected shares, hashrate developing toward the machine's rated range, and steady temperatures and fan speeds.

The risks, stated plainly

Replay exposure on pre-fork coins

Coins that existed before block 961,632 exist on both chains. Until the relevant UTXOs are separated, a transaction broadcast on one chain can in some circumstances be replayed on the other. Separate before you move anything of size, and do not assume your wallet handles it for you.

Liquidity and price discovery

BTCB2 printed an all-time high of $1,799 on 5 September and traded between $750 and $1,000 the following day, on roughly $1 million of daily volume concentrated in one pair. Some trackers quote materially lower figures. It is not listed on CoinMarketCap or CoinGecko. A fully diluted valuation computed from those trades has been described — fairly — as shaky.

Security budget

A network of a few petahash is a small network. It is well defended against Bitcoin's SHA-256 fleet, which cannot touch it, but it is not well defended against someone assembling a large share of the world's Blake2b capacity. That capacity is finite and rentable, which cuts both ways.

Rules that are explicitly temporary

Some of the chain's data-handling rules are scheduled to expire on 1 September 2027. A protocol with a dated clause is a protocol with a governance event already on the calendar.

Custody and infrastructure

Wallet support, explorers, exchange integrations and accounting tools are all immature. Plan to hold the coin somewhere you have verified works, and expect the tooling around it to change.

Final thoughts

Forks are usually arguments about what Bitcoin is for. This one produced something unusual: a chain that inherited Bitcoin's monetary parameters and then deliberately made itself unmineable by Bitcoin's hardware. That decision is what turned a category of obsolete ASICs into working machines overnight.

Whether BTCB2 lasts is a question nobody can answer five weeks in, and anyone who tells you otherwise is guessing. What is knowable is the hardware side: the machines are real, you own them outright, they can mine any Blake2b chain, and they can be resold. That is a very different risk profile from buying an abstraction.

Understand the chain before you buy the machine. Then price the machine on what it earns, not on what the fork story sounds like. Our comparison of all six Blake2b machines is where to do that.