Bitcoin Is Worth the Cost in Fees and Electricity

Saturday set a new record for cost per transaction, according to this chart at

For the first time ever, that value broke $100, clocking in at $103.

Fees are also rising. The relevant chart at shows a median fee of $16.75 for Friday (which also saw an all time high for the bitcoin price breaking $17,000).

Looking at the chart for cost per transaction, anybody familiar with the bitcoin price history will immediately recognize a high correlation between cost per transaction and bitcoin prices. Previous spikes in cost per transaction were in summer 2011 and in late 2013/early 2014.

This is of course easily explained by the fact that the largest part of the cost per transaction is the mining reward. Users pay only a small percentage of the cost as fees (right now this value is sitting at 18.46%). Therefore it is only to be expected that the cost per transaction goes up when miners receive more dollars for their mining rewards.

If so, higher transaction costs and higher fees only reflect the fact that bitcoin prices are rising. If you like higher bitcoin prices (I certainly do), the associated higher fees are just an inevitable sad fact of life.

Low value transactions will be priced out of the Bitcoin network until level two scaling solutions are deployed. But the percentage of cost compared to funds moved actually is stabilizing at low levels. It has not exceeded 1.25% in the whole of this year (7 days average). In contrast, in 2011 it was routinely breaking 10%, and on this day (December 11th) in 2010, that value was 34%. As recently as January 2016, a peak of over 2% was registered (all numbers from the relevant chart).

As bitcoin prices and transaction costs rise, so does the amount of funds moved each day. And since the percentage of costs compared to funds moved goes down, obviously the amount moved rises faster than prices and costs. Recent days have seen more than $3 billion moved on chain. That’s up from $160 million on December 15th of last year, for a gain of about 20x.

Actually paying around 1% in transaction costs is a bargain considering that Bitcoin is always running, is a world wide currency, and has security no centralized system can ever hope to achieve. Since only less than 20 percent of that is fees and more than 80% is mining reward, users get to move their funds for fees of around 0.20% right now.

Anyway, the market obviously thinks that Bitcoin is worth paying transaction costs, or the amount of funds moved on chain wouldn’t go up so massively. I agree with that sentiment.

As transaction costs go up, so does use of electricity by miners. The best estimate I am aware of gives a number of between 5.61 and 10.93 TWh a year for the 6,398 petahash hashrate in July of this year. Since that hashrate has gone up by a factor of about 2 until December 9 (now at 13,760), those numbers would need to be adjusted to between 11 and 20 TWh a year.

That’s a lot of energy. But I happen to think that Bitcoin is worth spending that energy cost as well.

One argument is that Bitcoin actually helps counter global warming. That’s because a large part of miners’ costs are from electricity. Therefore, they need to follow the aluminum industry to put their hardware on locations with low electricity cost.

Since about this year, the lowest electricity cost comes from solar power.

It follows that miners will want to use solar power. That means those 20 TWh a year in the upper estimate above increase demand for solar power. That in turn means that solar power will take over the electricity market even faster than without Bitcoin.

Bitcoin is also a source of electricity demand that does not depend on power lines. You can put a large renewable energy project right in the middle of the Sahara or some other desert and sell your electricity immediately to the Bitcoin network without building an expensive power line first.

It is also a source of electricity demand for time slots where renewable energy produces more than the grid can absorb. With rising renewable penetration rates that happens more often, leading to negative electricity prices. The Bitcoin network is one way to get more flexible demand. Turning your solar panels or wind energy off in time slots with insufficient demand completely wastes that energy. Mining Bitcoin in such time slots leaves value that would otherwise be lost.

That includes time slots where your project is not connected to the grid because the grid operator is allowed to refuse connecting it (like in Japan). You could build a large wind project in the Japanese Tohoku region without asking Tohoku Denryoku for permission and just mine bitcoins until they finally manage to connect the project to the grid.

That said, going forward Bitcoin needs to pay attention to this point. That is one very strong reason to reject the Bcash project. Said project wants to keep all transactions on chain, rejecting more effective level two solutions like Lightning. That approach will lead to more energy use at scale, as a result of accepting more use of network and computing resources (preferably paid by someone else).




  1. Karl-Friedrich Lenz

    That article relies on the estimates at While these are based on much simpler assumptions than Bevand’s model I cited above, they are not completely wrong either. The difference is about a factor of two for the medium assumptions in Bevand’s model.

    But anyway, that’s a lot of electricity. But my point doesn’t depend on either estimate being more in line with reality.

Comments are closed.