Skip to content

Bitcoin accepted at checkout  |  Ships from Montreal, QC, Canada  |  Expert support since 2016

Cost to 51%-Attack Bitcoin: A Live Security-Budget Calculator

Quick answer

To 51%-attack Bitcoin you must out-hash the honest network, so you need roughly as much hashrate as the whole network already runs — about 882 EH/s right now. That is ~791,783 of the most efficient current ASICs drawing ~8.7 GW. Two honest framings: renting that hashrate has a theoretical floor of about 1,757,336 CAD/hour (you must at least out-bid what it earns mining honestly) — but nowhere near half the network is actually available to rent. Acquiring the fleet is roughly 10.1B CAD in hardware plus power — and you cannot buy that many ASICs without exceeding global production. Either way it is a transparent model, not a single authoritative number — adjust the assumptions below.

A 51% attacker can double-spend their own recent transactions and censor or reorg blocks — they CANNOT steal other people's coins, mint new coins, or change the rules. The cost below is the price of temporary control, not ownership of the network.

Method & assumptions. To control a share S% of hashpower while honest miners keep mining, an attacker must add hashrate A = H × S / (100 − S), where H is the live network hashrate (from our hashprice feed, mempool.space-sourced) — at 51% that is ~1.04× the whole network. Miner count N = A / (miner TH/s); power P = A × (miner J/TH). Rent (theoretical floor): the least a rational hashrate owner would rent for is what it earns mining, so cost ≥ A × hashprice (CAD/TH/day) × hours/24 — a floor, because in reality only a tiny single-digit percentage of Bitcoin's hashrate is listed on rental markets, far below 50%. Acquire: CAPEX = A × CAD/TH (default is the median used-market data-centre ASIC price from our Used ASIC Price Index) plus OPEX = power × CAD/kWh × hours — and you cannot buy millions of current-gen ASICs without exceeding global production and moving the market, nor site multiple gigawatts of power overnight. This is a transparent model for scale, not a single authoritative figure; every input above is adjustable. Educational only, not security or investment advice. Credit: the "cost to attack" framing follows crypto51.app and the Bitcoin security-budget literature. Related: 51% attack (glossary), the Bitcoin security-budget analysis, and Canadian electricity rates.