In June 2021 China started enforcing a complete ban on all Bitcoin mining activities in the country. Up until this point, the Bitcoin mining network had always been dominated by Chinese miners. The mining exodus from China therefore triggered a substantial drop in the total amount of active computational power in the network, which crashed by 50% before stabilizing around 80 exahashes per second.
Update: this content was updated to reflect the network’s situation per the end of July, 2021.
Of course, such a large drop in active equipment also affects the energy consumption of the network. Unfortunately, it’s never possible to say by how much exactly. A 50% drop in the computational power does not necessarily equal a 50% drop in energy consumption. It is impossible to determine what devices make up for the part of the network that went offline. If these mostly belong to the latest most power-efficient generation of Bitcoin mining machines, then the respective energy consumption could represent a lot less than 50% of the total network’s energy demand. For example, it only takes around 727,000 of Bitmain’s (the world’s largest manufacturer of Bitcoin mining devices which was also headquartered in China) Antminer S19 Pro devices to generate 80 exahashes per second. These machines would consume around 25 terawatt-hours of electrical energy annually, which is only a fifth of the network’s total estimated power requirement before China banned Bitcoin mining.
Whatever the precise amount may be, neither the Bitcoin Energy Consumption Index, nor any other model or tracker, is equipped to properly deal with a network disruption of this magnitude. Because of the previous challenges in determining the most likely energy consumption impact, any adjustment would be arbitrary. For this reason, no adjustments were made to reflect immediate impact of the ban.
In any case, the postive effects of China’s ban on the network’s energy consumption have likely been short-lived. The difficulty of mining Bitcoin dropped rapidly in the following weeks, which boosted the income for the remaining miners (as they have less competitors to share the rewards with.
The Bitcoin network hashrate inevitably started rising again afterward, recovering to around 110 exahashes per second at the end of July.
While this may not be the same hashrate as the network had prior to the Chinese mining ban, the energy consumption of the network could be a lot closer to making a complete recovery. One should keep in mind that a Bitcoin miner may have different performance settings and can be overclocked to further boost performance. This type of tweaking can disproportionality boost the energy consumption of the network (relatively adding a lot of energy consumption compared to the amount of hashrate added to the network) while others are relocating their devices. This strategy allows miners that weren’t forced to shut down to maximize the benefit of the Chinese competition going offline. It would also strongly limit the overall impact of the Chinese Bitcoin mining ban on the network’s energy consumption.
The impact on the carbon footprint estimate featured on the Bitcoin Energy Consumption Index is even less significant. As Chinese miners always primarily made use of hydroelectric power in the provinces of Sichuan and Yunnan during the summer rain season, these machines going offline during the summer does little to change the carbon impact of the network. By the end of the summer in September/October these miners would normally migrate to the provinces of Xinjiang and Inner Mongolia to take advantage of an abundance of coal during the rest of the year. These machines are now landing in fossil-fuel dependent countries like Kazakhstan. A recently signed deal would even have most of the machines coming from China end up running on Canadian natural gas. This could mean that, in the short term, the carbon impact of the network could actually increase rather than decrease.
Digiconomist will continue to monitor the developments in the Bitcoin network and provide an update in due course.