Bitcoin’s computing power has grown a hundred-trillion-fold since 2010. Fitted the way the price is, it follows a power law over any stretch you pick, but a different one each time: the growth keeps slowing, and the price moves first.
Dr. Giovanni Santostasi, who first framed bitcoin’s price as a power law of time, has fitted the hash rate and active addresses the same way. Stephen Perrenod fitted the mining difficulty, which tracks the hash rate, and found an index of about 11.4 (“Bitcoin’s power laws”). The fits below, the rolling exponent and the price comparisons are this site’s own, computed from Blockchain.com’s hash-rate estimates and the site’s price record.
Miners compete to add the next block by brute force. There is no calculation to solve and no shortcut: each miner takes the block it wants to add, adds a number, and runs it through a hash function, which returns what is in effect a random number. If that number is below a target, the block is valid; if not, change the number and try again. It is closer to rolling dice trillions of times a second than to working anything out. The hash rate is how many of those guesses the whole network makes each second, measured in hashes per second. Today that runs to zettahashes: a zettahash is 1021, a thousand billion billion. Nobody reports the figure: it is estimated from how hard the target is (the difficulty) and how fast blocks are actually found, so a single day’s figure is noisy. This page uses weekly averages.
It matters because it is the cost of attacking the network. To rewrite recent blocks, an attacker would need more computing power than everyone else combined, and every rise in the hash rate raises that bar. That work is paid for in two ways. The first is the block reward: a fixed number of new bitcoin per block (3.125 since April 2024), halved about every four years, whose value to the miner is that number times bitcoin’s price in dollars or other currency, the way most participants judge it. The second is the fees on the transactions the block picks up from the mempool, the queue of transactions waiting to be confirmed. A higher price can mean higher fees, but not necessarily: fees depend mostly on how crowded the mempool is and how urgently people want their transactions confirmed. Both streams, and so the hash rate miners can afford, rise and fall with the price.
A power law grows quickly at first and more slowly later, but at a steady rate in the right units: the exponent should be the same whatever stretch you fit. Fit the hash rate from three starting years and the exponent falls each time.
| Fit | Exponent | R² | Implied doubling time | Today against it |
|---|
Most of the early speed came from the machines. Mining moved from ordinary processors to graphics cards to chips built for nothing else, and each step multiplied the hashes a dollar of electricity could buy. Those leaps have run their course: chips still improve, but by smaller steps, and the hash rate now depends more on how much miners spend than on leaps in efficiency. That is why the exponent fell, and why the hash rate today sits well below a line drawn through its fast early years.
Not one power law. Bitcoin’s price, fitted the same way, keeps roughly the same exponent as the starting year moves; the hash rate’s falls steadily. Here are both, from the same starting years:
| Fit from | Price exponent | Price R² | Hash-rate exponent | Hash-rate R² |
|---|
This page’s reading: the hash rate is not a law of its own. It is downstream of the price. A miner’s revenue is the reward and fees, valued at the price, and over time miners spend roughly what mining pays on machines and electricity. When chips were getting dramatically better each generation, the same spending bought ever more hashes, and the hash rate raced ahead of the price. As those gains shrink, the hash rate grows more nearly in step with what is spent on it, which follows the price. If that is right, the hash rate’s exponent should settle near the price’s, plus whatever chips still improve, rather than keep falling, and the next few years of data will test it. It is a reading of the record, not a law; the halvings, which cut the reward in half every four years, pull the other way.
Each dot below is a month: bitcoin’s price against the hash rate that month, both on log scales. A straight band of dots means one grows as a power of the other.
Take each month’s change over the year before, in price and in hash rate, and line the two up with one shifted against the other. The taller a bar, the more closely the two changes track at that shift.
That order is what the economics would predict. A higher price raises what each block pays, miners buy and plug in more machines, and the new machines take months to order, ship and power up. The hash rate is the security the price pays for, and it follows the price; it does not forecast it. A rising hash rate says the network is getting harder to attack, and that miners expect mining to pay. It says nothing about where the price goes next.
A high hash rate is not the whole of security. How the mining is spread matters too: across pools, across countries, across power sources. Risks to Bitcoin covers pool concentration and where the miners are. For bitcoin’s price as a power law of time, the trend this page measures the hash rate against, start with The Power Law.
Hash rate: Blockchain.com’s daily estimate (chart), averaged by week from 2010; today’s figure is the mean of the last seven daily estimates. Other trackers estimate it differently and can differ by a few percent.
The fits: least squares of the logarithm of hash rate on the logarithm of days since the genesis block (3 January 2009), the same form as the price power law, from 2011, 2016 and 2020, and over rolling four-year windows, refitted each time the page loads. Doubling times come from the slope of the hash rate’s logarithm against time over each window.
Price against hash rate: the site’s price record (the samples behind The Power Law) against the hash rate of the same week. Which moves first: year-over-year changes in the logarithms of price and hash rate, monthly from 2014, correlated with one shifted by up to twelve months. A correlation shows the two move together at that shift; it does not by itself prove which causes which.
Sources: Blockchain.com, hash rate; Stephen Perrenod, “Bitcoin’s power laws”; Giovanni Santostasi’s power-law work on hash rate and addresses.
Get the essays and tool updates
A few emails a month when something new ships — an essay, a new tool, a meaningful update. No funnel, nothing for sale, unsubscribe anytime.
Every page on this site has been improved by someone pushing on it. Ask a question, flag an error, or suggest what’s missing — it goes straight to the author, never published.
I’m also available for one-on-one conversations and open to collaborations — Work with me →