# What is the average amount of electrical power required to complete one Bitcoin transaction? [duplicate]

Once a bitcoin has been mined, if I want to use it to pay for something, what is the average amount of electrical power required to process that payment or bitcoin transfer? Does this vary with the value of the transaction, and if so in what manner?

Please note that I am not asking about the electricity cost, as that would clearly depend on the location as well as what deal you've been able to strike with the power company. Hopefully the number of kWh, J, or whichever other unit happens to be convenient, will be easier to answer reasonably objectively.

Answers for any point in time during which Bitcoin has existed are acceptable, but more recent data is better. It is acceptable to exclude costs that are not actually related to the data being transferred being about bitcoins per se (for example, I see no real need to calculate how large a percentage of Internet core router traffic is for bitcoin transfers and account for that).

(I did find e.g. Energy to confirm one block? and Can it really be true that the Bitcoin system consumes almost \$1.000.000 in electricity per day? as well as (on Skeptics) Does one transaction on Bitcoin require 215kWh? but all of these appear to conflate the total power usage of the Bitcoin network with the power required for a transaction using already mined Bitcoin by lumping the two together.)

• They are lumped together because they are pretty indivisible. You can make a transaction to send me 1 BTC and it can show in my wallet, but it will show unconfirmed until the next block is mined. Until its confirmed you could always re-spend (double-spend) the same 1 BTC somewhere else, and so the transaction isnt securely completed until at least one confirmation has occurred. Its possible for the blockchain to re-organise, and so for very large (high risk) transfers the recipient may wish to wait for many more confirmations. Nov 3, 2017 at 14:34

One new block gets mined every 10 minutes. This makes 144 blocks per day. Today there have been 274000 transactions (source).

This makes around 2000 transactions per block. The current difficulty is 1,452,839,779,146, so that mining one block requires on average 2^32 * 1,452,839,779,146 hashes.

One AntMiner S9 can do 14 Thashes/s, so it would require 123,800 hours to find a block. It consumes 1.4 kW of power, so the block has required 173,320 kWh.

This makes 87 kWh per confirmed transaction, today.

• Dividing the total energy expenditure of Bitcoin by the number of transactions included in the blockchain does not provide a meaningful estimate of the “energy required to complete one Bitcoin transaction” since blocks are mined regardless of whether transactions are included or not.
– Murch
Dec 6, 2022 at 21:03
• How would you define the "average amount of electrical power required to complete one Bitcoin transaction" then? Dec 8, 2022 at 8:47
– Murch
Dec 8, 2022 at 14:27

Once a bitcoin has been mined, if I want to use it to pay for something, what is the average amount of electrical power required to process that payment or bitcoin transfer?

That a transaction is processed could mean it has gone all the way from its generation up to its addition on the blockchain within a mined block confirmed by all nodes.

The only significant consumption in the process is the guessing of the correct hash of the block that includes this transaction along with `cca 2000` (meaning around `500 byte` transaction size on average) other transactions included in it. This guessing (lottery) difficulty varies with time thus the energy consumption varies. Current guessing `difficulty` is `1,452,839,779,146` (https://bitcoinwisdom.com/bitcoin/difficulty), and the average number of guesses to make is `2^32 * difficulty` (`2^32 * 1,452,839,779,146 = awfully high`).

Now you can find the energy consumption for example by examining some high end mining device like `AntMiner S9` and checking its hashing rate (for this example `14,000,000 Mhash/s`, https://en.bitcoin.it/wiki/Mining_hardware_comparison) and its power consumption (`1,375 W`).

``````Average_number_of_guesses / hashing_rate = Guessing_time
Guessing_time * power_consumption = Energy_spent_for_mining_a_block
Energy_spent_for_mining_a_block / number_of_transactions_in_block = Energy_spent_per_transaction
``````

If I've done calculation correctly this should be around `85 kWh` per average transaction. (Note that not every mining device on the network is AntMiner S9 so this figure should be used with reserve).

Does this vary with the value of the transaction, and if so in what manner?

It does not vary with transaction value but the size of transaction in bytes and this varies by the number of inputs and outputs of transaction. If has more inputs and outputs it will be bigger in size thus allowing less transactions in a block and the last equation will change accordingly. For example if transaction is `1000 bytes` big instead of average `500` used in last equation it will use around `170 kWh`

• How do you come up with the 0.02 kWh per transaction? See my own answer. One of us is making a mistake, but who? Nov 4, 2017 at 8:15
• I cannot redo my calculations. It looks I accidentally divided twice with 3600 or something. But if so is true than 85kWh is awfully high and bitcoin has no future. I will try to correct the procedure. Nov 4, 2017 at 10:52
• Dividing the total energy expenditure of Bitcoin by the number of transactions included in the blockchain does not provide a meaningful estimate of the “energy required to complete one Bitcoin transaction” since blocks are mined regardless of whether transactions are included or not.
– Murch
Dec 6, 2022 at 21:03
• @Murch It does provide a meaningful estimate. The average number of transactions included in a block provides the average consumption per transaction. If less transactions are included than average that would give bigger consumption per transaction for that block. If no transactions are included in some block, that would give infinite energy spent per transaction, since energy was spent, but not a single transaction was produced. And that is perfectly meaningful explaination. Dec 11, 2022 at 9:35
• @croraf: You just pointed yourself how the count of transactions is inversely correlated to the “energy spent per transaction”. So, adding more transactions clearly lowers the “energy spent per transaction” which indicates that the energy is not spent on the transaction but something else.
– Murch
Dec 11, 2022 at 15:51