In 2018, entrepreneur and physicist Dhruv Bansal gave a low-key talk that introduced a big idea. The idea emphasized the fraction of world electricity used to secure the Bitcoin network.1 Aptly named the “Nakamoto Ratio”, Bansal argued that this ratio would continue to increase until it reached its saturation point; which, to his conjecture, was the point at which marginal revenues from securing Bitcoin equaled the marginal revenues from selling electricity.
Bansal deliberately chose this equilibrium because he understood incentives. If energy companies can make more money mining bitcoin rather than selling it to end-users, they will arbitrage that difference until profitability flattens or flips in the other direction. Today’s energy companies will become tomorrow’s miners, and today’s miners will become tomorrow’s energy companies.
Four years ago, bitcoin miners were considered large if they consumed two-digit megawatt-hours’ worth of electricity.2 But it’s no longer uncommon to hear about developments in the hundreds of megawatts. Bitcoin miners are increasingly partnering with large energy companies to tap into their vast portfolio of electricity-generating assets. The pace at which they’re growing has even spooked the most business-friendly state in the U.S.—yes, I’m talking about Texas—after the local grid operator hit the pause button on new buildouts, citing the need to reassess mining’s impact on the grid.
When Bansal gave his talk in 2018, the Nakamoto Ratio was hovering around 0.17%. Nearly all of Bitcoin’s power demand could be satisfied with one large nuclear power plant. Fast forward to 2021, and that ratio has grown to 0.37%, a power demand that would need the equivalent of three large nuclear power plants.3 Be that as it may, the takeaway is this: though Bitcoin’s energy consumption continues to rise, it’s a popular misconception in the media that Bitcoin demands a large share of the world’s energy, but the data proves the opposite — Bitcoin is still a minuscule amount of demand compared to total production.
While Bansal accurately predicted the merge between energy companies and bitcoin miners, the growth in the Nakamoto Ratio proves that the bitcoin-energy arbitrage still exists; which begs the question:
Will the Nakamoto Ratio ever reach its saturation point?
Or, are we doomed to an ever-increasing fraction of electricity whereby every electron in the world goes to mining bitcoin? “Sorry, you can’t charge your iPhone—those electrons are mining bitcoin.”
Maybe every energy company will start mining bitcoin. Stove isn’t lighting? All the world’s natural gas is being used to generate electricity—to mine bitcoin. Perhaps bitcoin will consume all the Earth’s resources, then the Sun’s, and eventually trigger the Heat Death of the Universe.
Obviously, the idea that Bitcoin will consume all of our energy sounds ridiculous, and I am skeptical that public policy will need to intervene to prevent that from even becoming a remote possibility. Gold mining companies still exist today, mining ever more gold as extraction technology advances, yet there are physical and financial constraints that limit gold mining’s impact on global energy use. The same is true for bitcoin mining.
Unlike gold, however, bitcoin mining will never benefit from innovations in exploration or extraction. Its future supply is certain. Bitcoin’s supply gets cut in half roughly every 4 years by design. Which leads me to my own conjecture — the Nakamoto Ratio will reach its saturation point when bitcoin’s halving decreases mining revenue to the point of equilibrium.
I know that doesn’t answer the “when”, but I do think it answers the “will it ever”. While we can make dozens of assumptions about bitcoin’s price, future transaction fees, and so on and so forth, the biggest certainty we have is bitcoin’s halving, which—all else being equal—cuts mining revenues in half. Nonetheless, bitcoin’s price has continued to increase with every halving, and the last bitcoin is scheduled to be mined in 2140. That’s a long runway for the Nakamoto Ratio to continue growing.
Without getting into too much math, mining revenues for the past 12 months was 12.9 cents per kWh. The average retail price of electricity in 2020 was about 10.6 cents per kWh. Since wholesale electricity markets tend to trade much lower than that, let’s assume a vertically integrated miner can sell electrons for 4 cents per kWh. Depending on your assumptions around bitcoin’s price and power demand, mining revenues could fall below this threshold as early as the next halving (in 2024), or as late as the next decade, possibly even into the 2040s.4
If you’re interested in Dhruv’s idea around the Nakamoto Ratio, and want to toy around with some of these assumptions, I encourage you to plug and play with this (very basic) model that projects future mining revenues. Maybe the results will excite you, or maybe they’ll scare you. In the end, no one can predict bitcoin’s price. But we can predict bitcoin’s supply. Whether we can predict a happy marriage between the Bitcoin and Energy sectors remains to be seen.
- Dhruv’s introduction of the Nakamoto Ratio actually talked about the fraction of world electricity used to secure “proof of work money”, but since every other digital asset besides bitcoin uses a negligible amount of electricity, I’ve chosen to focus solely on bitcoin for simplicity. ↩
- To put this in perspective, 1 megawatt can power approximately 200 family homes. ↩
- Sources: Cambridge Center for Alternative Finance; BP. Bitcoin’s annual electricity consumption in 2018 was approximately 45 TWh, and in 2021 it was approximately 105 TWh. Total world electricity generation in 2018 was approximately 26,677 TWh, and in 2021 it was approximately 28,466 TWh. A “large nuclear power plant” refers to a 5,000 MWh power plant with a 75% capacity factor. ↩
- Sources: Clark Moody Bitcoin; EIA. Mining revenues per kWh based on $14.3 million in mining revenues divided by 111,210 million kilowatt-hours of electricity consumption. Average retail price of electricity taken from the EIA’s Electric Power Annual 2020 report. Since the average price of electricity being sold from energy-only providers (i.e. generation plants) to industrial customers (e.g., bitcoin miners) was around 4.89 cents per kWh in 2020, I decided to use 4 cents as a conservative estimate for vertically integrated miners. ↩
Cover Photo from Talen Energy