Bitcoin Energy Usage

Intro

Rarely do I see an issue as polarizing as Bitcoin’s energy usage. Scrolling through twitter I will see one person claim Bitcoin will be the reason we boil the oceans, and but a moment later see another claim that Bitcoin will single-handedly be the incentive for transitioning to wind & solar. Such polarization is a huge red flag that reality must be more complex. With this article I’m going to try to present a more nuanced, pragmatic mental model that can assist in proposing balanced regulatory policies.

To be upfront, the one group that may have a problem with the arguments I present here are those who believe the world requires de-growth. I am taking it as an axiom that the de-growth thesis is incorrect; for justification of this axiom I would recommend the book More from Less by Andrew McAffee. Practically what this means is that energy use in and of itself it not a bad thing (in fact much of human flourishing is a direct product of increased energy consumption). The real problems we face today are those of unpriced social costs (ie CO2 emmission), and inefficent consumption (ie unpurchased waste, such as producing products that are not in demand).

With this in mind, let’s first list some of the potential impacts (both good & bad) of Bitcoin mining that are most often dicussed:

• Miners can be used to subsidize the build-out of renewables before consumer demand is available
• Miners can be used to subsidize the prolonged use (or even reactivation) of hydrocarbon-based power plants
• Miners can provide dynamic load balancing to improve the health of a power grid
• Miners can consume so much energy that it causes reduced health of a power grid
• Miners can compete with consumer demand, increasing the prices of home electrical bills

How is it that so many of these points are complete opposites and contradictory of eachother? Are some just made up false-hoods? Well, I think we need to move away from a naive “is this true or false?” type question, and instead ask “when, under which circumstances, are each of these true?”. Being able to answer such a question will let us also discuss policy proposals (such as carbon pricing) and think through how they will actually affect the system. Is there a way we can target a system that gains the good impacts above while minimizing the bad impacts?

Opting Out

A jurisdiction choosing to ban mining as a way of “opting out” of more nuanced policy making is a valid way of saying “hey we don’t think the potential upsides of mining on our electrical systems outweight the risks, we’re just not going to do that here”. However, we must also keep in mind the secondary effects of this decision! Bitcoin is a trans-national autonomous system that will not reduce its purchase of power. By banning mining in one jurisdiction, it will effectively subsize the mining in all other jurisdictions by increasing their profit margins. If a climate-conscious nation with a relatively green grid bans bitcoin mining, global emissions should be expected to increase as hashrate moves off the clean grid to relatively more dirty ones.

Targeting Good Outcomes

The best heuristic I’ve been able to determine for whether a jurisdiction will experience the good impacts of mining vs the bad impacts is the miner’s bid price for power. In general, when a miner is bidding a lower dollar value per kWh it seems to maximize the impact of the “good” outcomes while minimizing the impact of the “bad” outcomes.

• A sufficiently low bid price means mining can only subsidize the cheapest electrical generation sources. Generally renewables are cheaper than hydrocarbons (especially when combined with carbon pricing policies)
• A sufficiently low bid price means miners are quick to turn off as soon as demand on the grid increases very slightly (as compared to delaying the shutdown until prices are very high). This reduces strain on the grid and consumer bills.
• A sufficiently low bid price means miners are easily outbid by households, and will be forced to move elsewhere (ie continually move towards emerging power sources)

A miner’s maximum bid price will be determined by the miner’s revenue (which is structurally similar across all miners), and their costs (which will vary widely from miner to miner).

Miner Revenue

A quick reminder how Bitcoin hashrate purchasing works. With the goal of keeping itself secure, the autonomous Bitcoin system purchases a number of hashes every 10 minutes, funded by currency dilution. Importantly, the BTC/hash market clearing price is determined by way of a target spend rather than a target amount (“I need to spend $10 on candy” rather than “I need to buy 200g of candy”). What this means in mining terms is that the Bitcoin system will purchase as many hashes as the market will provide for a fixed total cost (currently it spends 6.25 BTC every 10 minutes). The important conclusion here is this: more hashrate results in a lower hashprice (and vice-versa)

Miner’s revenue in this model is determined by how much of the total hashrate they provide. Or in equation form: revenue = (hashrate / global_hashrate) * block_reward

Miner Costs

There will be a much wider variety of costs for miners (think real estate, operations, debt load), but one thing they all have in common is of course ASIC cost and electricity cost. Because the Bitcoin system has a fixed spend amount, miners across the board compete with eachother solely on the cost side of the equation. Those with the lowest cost per hash can expand their operations, while those with the highest cost per hash will be forced to downsize or shut down. In particular, the efficiency of newer ASIC hardware will deliver more hashes per second compared to an older machine, while using the same amount of power draw. In a capitalist society it is generally safe to assume that miner costs will continue to be a race to the bottom.

Bid Price

In summary, the ability to place high bids on electricity comes down to either:

1) The hashprice (revenue) is exceptionally high compared to global hashrate. For example a sudden increase in the purchasing power of BTC can be an acute risk.

2) A specific miner has particularily low costs compared to average. For example a vertically integrated monopolistic company that controls the best ASICs.

These are the issues which should be addressed to maximize the best outcomes of mining while minimizing the worst outcomes.

Policy

Below I will analyize a series of policy categories (some I agree with, and some I do not), and how they would affect Bitcoin’s impact on the jurisdictions own grid, the climate, and users of the network.

Tax Miners

Introducing a tax targeted at miners (ie raising their costs) would indeed reduce the maximum bid price of miners within that tax region. However, similar to banning miners (a 100% tax), this creates a global subsidy to other miners by making your own miners less competitive. This can potentially increase emissions, and require other sources of load balancing to fill the gap left by miners leaving the grid.

Tax / Ban Bitcoin

Rather than targetting the miners directly, this policy aims at reducing the demand for Bitcoin the asset, and therefore revenue for all miners. We see this play out naturally during every bear market: when the BTC price drops dramatically, miners are squeezed out of business and efficiency consolidations take place. While this idea works in theory, I have my doubts about its long term success. While a ban would certainly make it difficult for large companies to own Bitcoin, the general public would likely have no issue continuing to use the technology. Banning of the technology itself would likely be a political nightmare as well, as it is widely regarded as a tool for personal freedom, much like encryption. I’d personally find this somewhat unethical.

Set a price limit for Miners

This policy type aims to directly control the bid price by setting a limit on how expensive miners are permitted to bid. I expect this would result in something similar to a tax: miners in this region are sometimes forced to operate sub-optimally and will therefore seek out other jurisdictions to operate under. However I think it can make for a good emergency measure to mitigate against the risk of a sudden increase in the BTC/USD price, while new energy sources are built out.

Carbon Pricing

Whether it is a general policy or targetted to miners, having a tax that is rated based on the carbon intensity of the miner’s power source would certainly shift miners towards cleaner energy usage. However, miners may also shift out to jurisdictions with no tax instead.

Alternatively a “renewable subsidy” can be used to reduce the cost of clean energy in a region. This would attract new miners to the region, likely from more expensive and more dirty power sources. However, a subsidy given to miners may also increase their energy bid.

Ideally, we want renewables to naturally have a lower price than hydrocarbons (which is already true in many regions today), and for carbon pricing to be used only when required.

Democratize ASIC Access

Ensuring that miners have fair access to the latest, most efficient ASIC hardware will help reduce the cases where a single miner has a vastly better cost profile compared to their competitors. The net effect here is margins should be squeezed, reducing maximum energy bids.

Maximizing Competition / Deregulation

The classic way of minimizing the margins of businesses within a capitalist society is to encourage the most free, most competitive markets possible. In practice this will mean more bitcoin mining taking place, and enforcing anti-trust laws.