Bitcoin Mining Network Report [Trends/Structure/Marginal Production Cost/Power Consumption and Resources]

Author: CoinShares Research

Release time: end of November 2018

Translation: backlight/Bangkok District Shijun

Review: Yan Po Gong/tinker

Translator introduction:

backlight : Five-year veteran captain of Babbitt Forum, blockchain enthusiast, managing partner of cryptocurrency fund

Bangkok Qu Shijun : A cryptocurrency and blockchain researcher who works for well-known mining pools (PoW and PoS) in Bangkok and Beijing. He also opens a column to explore the Thai cryptocurrency circle while not doing his job.

Thanks to the following communities for their support

Official Account: Cryptocurrency Encyclopedia and Cryptocurrency Explorer Community

Official account: NPC Source Project and NPC Crypto Labss community

Official account: BIB Global Community

Overview

In this report, we investigate the geographic distribution, composition, efficiency, power consumption, and power sources of the Bitcoin mining network. We also examine trends in hashrate, mining machine costs, energy efficiency ratios, and marginal production costs. We find that the average marginal production cost of the market has increased from about $ 6,500 to $ 6,800 since May 2018, based on 5 cents / kWh and 18 months of depreciation. This suggests that at current prices, the average miner is either operating at a loss and unable to cover capital expenditures unless electricity costs are close to 3 cents / kWh and mining machines are depreciated over 24-30 months, or the cost of purchasing mining machines is lower than we expected. In addition, we find that Bitcoin mining farms are mainly distributed in areas with a large excess supply of renewable electricity. We calculate that the proportion of renewable energy used in the Bitcoin mining network is 77.6 %, making Bitcoin mining more environmentally friendly than almost any other large industry in the world.

In this second update to our H2 Mining Report, we continue our ongoing observations and analysis of the state of the Bitcoin mining network. For a brief introduction to mining, its role and an explanation of our network modeling approach, please refer to our May 2018 report. We continue to emphasize the importance of a strong mining network for Bitcoin - harnessing market willingness to redistribute energy is an important guarantee for Bitcoin as a trustless asset. Of all currently deployed consensus algorithms, Proof of Work achieves the highest degree of trust minimization. Under PoW , any user can verify for themselves that the current state of the UTXO database is indeed correct, as calculated by the valid chain containing the most accumulated work (hereinafter referred to as the longest chain). The accumulated work in any given chain is cryptographically and thermodynamically proven, so the validity of the longest chain does not require confirmation from anyone. This has not been proven to be feasible under any other non- PoW consensus mechanism, all of which rely on some form of trust in current or former network participants.

We keep stressing these facts because most of the narratives about the environmental impact of cryptocurrency mining are overwhelmingly negative. The amount of electricity consumed by cryptocurrency mining to produce PoW in a highly competitive market continues to increase. One reputable publication even made the unfounded allegation that the increased CO2 emissions from Bitcoin mining alone could eventually cause a two-degree rise in global temperatures.

We believe these rumors stem from a misunderstanding of the Bitcoin protocol and mining network. Worse, we suspect they stem from the widespread belief that Bitcoin mining is a pointless endeavor that will

It produces negative externalities, environmental impact being one of them, and does little to help the real economy. Our view is that cryptocurrency mining, while costly, does little to no real damage in terms of the environment, and is unlikely to do so in the foreseeable future. We also believe that the cost is worth it for a globally transferable currency that is auditable, backed by a solid monetary policy. In fact, we can even argue that Bitcoin mining could promote the development of renewable energy generation, thanks to the fact that mining farms are moving their operations to areas with cheap and surplus electricity.

Network Development and Marginal Costs

Since our last report in May 2018, the Bitcoin hashrate has increased from about 30 EH/s to about 40 EH/s . The hashrate growth rate is faster than the two-year average [ Figure 3] , but slower than the historical average. At the same time, the price of Bitcoin has fallen from $ 8,500 to about $ 4,000 . This has undoubtedly put pressure on many miners, as revenue has fallen and difficulty has increased significantly.

Several new generation mining machines were also released in the second half of 2018, with significant improvements in both energy efficiency per GH/J and cost per TH/s . On average, the energy efficiency of new mining machines since our last report has been almost exactly the same as the previous trend line [ Figure 1] . However, the cost of mining machine hashrate is significantly lower than the previous trend line [ Figure 2] . Our previous estimate of the market average production cost was based on 5 cents / kW electricity and 18 months of mining machine depreciation, which is now about $ 8,500 , compared to $ 6,500 in May [ Table 3] . In the absence of a reduction in Bitcoin hashrate, we believe this means that ordinary miners are either running at a loss and unable to cover their expenses unless the electricity price is close to 3 cents / kW and the mining machine depreciates for more than 24-30 months, or the cost of the mining machine is lower than our estimate [ Table 1-5] . As can be seen from Table 1-5 , mining is still profitable when the electricity price and mining machine cost are appropriate.

Another trend is that mining farms are preparing to move out of China. After reviewing public literature and talking to industry insiders, it is clear that miners are leaving China, or stopping reinvesting in China. Instead, they are operating in certain areas of Scandinavia, Russia, Canada, and the United States, which have cheap electricity, friendlier regulations, faster Internet access, and to some extent, cooler climates.

Quantifying the proportion of renewable energy used in Bitcoin mining

It is widely believed that the majority of Bitcoin mining farms are still located within China. While we cautiously agree with this view, our estimate is that this proportion is currently no more than 60%. We will conduct this analysis by studying power generation data from different provinces and cross-referencing it with publicly disclosed data from renewable energy power companies whenever possible.

The structure of our argument is a continuation of the hypothesis we proposed in our previous report, which is to analyze China's energy structure and analyze the most likely sources of power supply for miners in the region by cross-validating different provinces with data on power curtailment within the province. What we try to prove is that the vast majority of Chinese Bitcoin mining farms use renewable energy, which was previously not connected to the grid due to overload, known as curtailment.

Figure 6 : Global distribution of Bitcoin mining farms. Blue represents large mining farms, and black represents small mining farms.

China’s mining farms are mainly located in a few provinces – Sichuan accounts for about 80% , with the rest distributed in Guizhou, Tibet, Xinjiang, western Inner Mongolia and Heilongjiang [ Figure 6]

These locations were not randomly selected. The key factors in determining the location of the mines are low-cost electricity, high-speed internet access, and in the north, cold climates to reduce the cost of additional cooling. Although there are some mines located outside the area of ​​our focus, they are too small to be investigated in detail.

Curtailment : Renewables’ long-term problem

A key issue for renewable energy generation in China over the past decade has been the scale of curtailment. The impact of curtailment on the profitability of renewable energy generation companies is significant: on a fixed cost basis, with the depreciation of high initial investments, curtailment of generation output has a huge impact on the net profit and sustainability of these renewable energy generation operators.

For readers who are not familiar with curtailment, the term refers to the consequences of excessive renewable energy generation, where the electricity generated is curtailed by the grid due to concerns about overload and grid shutdown risks. The main reason for curtailment stems from overinvestment and subsequent overcapacity - the Chinese government has a large number of renewable energy generation (mainly wind and solar) in areas with high generation potential but low local population, concentrated in the north and west. The curtailment rate in various regions has previously reached highs of more than 30 % [ Table 6 , Table 7] , creating a huge barrier to the investability and availability of renewable energy projects. The government has been heavily subsidizing the renewable energy sector, with some provinces even having renewable energy funds as high as US$ 25 billion. However, we believe that in order for the industry to continue to grow, independent sustainability must be established in a sustainable and permanent manner, and should not rely on government subsidies.

Of the three main forms of renewable energy, curtailment data is most readily available for wind power, followed by a smaller number of solar plants, but there is almost no public data for hydropower. Much depends on the number of listed companies in the relevant industry, which determines whether there is verifiable and publicly disclosed data.

While publicly available news articles provide sufficient data on the locations of major hydropower plants and dam projects in China to construct Table 8, data on curtailment of hydropower generation is not yet available. After a cursory analysis and synthesis of the data, we believe that this cannot be explained by mere coincidence: most of China’s Bitcoin mining farms are located in provinces with excess wind/solar power or large total hydropower installed capacity.

Quantifying renewable energy usage in Chinese Bitcoin mining farms

It is obviously no accident that mines are located in areas with an oversupply of cheap electricity, especially electricity that would otherwise be wasted.

Electricity is essential for miners because electricity costs account for a large proportion of the total mining costs (just under 50 % based on an 18-month depreciation schedule, and a larger proportion if the depreciation period is longer).

While we cannot determine the percentage of renewable energy and non-renewable energy used by a single mine, we can make an estimate based on the policy targets for renewable energy and non-renewable energy use set out in the latest Renewable Portfolio Standard (RPS) issued by the government. Here are the estimates from the Morgan Stanley Utilities Research Team (Simon Lee and Eva Hou ):

The correlation is clear: provinces with a high concentration of mining farms in China are also provinces with a high share of renewable energy. For example, Sichuan is home to 80% of China’s Bitcoin mining farms ( 48 % of the world’s), and in 2017 , 90 % of its energy mix was renewable energy, and most miners face the same situation.

Based on the chart above, we can reasonably estimate that 43.3% of the world's Bitcoin mining farms are powered by renewable energy in Sichuan. We can then take the average renewable energy ratio of the other six provinces, study the remaining 20 % ​​of Chinese Bitcoin mining farms ( 12 % of the world), and calculate their global mining. The proportion of total energy consumed is 5.7 % renewable energy and 6.3 % fossil fuels and nuclear energy [ Table 10]

Quantifying the proportion of renewable energy used in Western mines

Outside of China, the largest Bitcoin mining farms are located primarily in the Pacific Northwest (Washington, Oregon, and British Columbia), Quebec, northern New York, northern Scandinavia (Norway and Sweden), Iceland, and Georgia. With the exception of New York and Russia, the energy sectors in almost all of these regions are dominated by renewable energy, and publicly available data shows the percentage of renewable energy in each region [Figure 9] . Europe and North America also have the lowest hydropower utilization in the world, with less than 40 % of installed capacity in both regions.

We can see that it is no coincidence that miners choose to open mines in regions with cheap excess renewable energy (primarily hydropower) such as Washington, Oregon, Norway, and Quebec. While we suspect the same is true for miners in New York State and Russia, and there is evidence to support this, we still conservatively assume that these regions have low renewable energy in their energy mix.

While there is no way to unilaterally prove the source of electricity used by a single non-Chinese mining farm, it is reasonable to assume that they use a renewable energy share that is not lower than the regional average. Let's accept the consensus that 40% of mining is done outside of China, and 40 % of mining is in the relevant regions mentioned above. We then assume that mining is evenly distributed - this is not true, but it is useful as an approximation for a lower bound. We can then take the average of the renewable energy mix in these regions ( 79 %) to reasonably estimate that of the 35% of Bitcoin mining farms in the West, at least 27.8 % are powered by renewable energy. Assuming the remaining 5 % of mining farms are evenly distributed worldwide, and the global renewable energy share is 18.2 %, then the remaining mining farms use 0.9 % renewable energy and the fossil / nuclear part uses 4.1 %.

The combined chart shows that Bitcoin's global electricity mix has a minimum of 77.6% renewable energy and a maximum of 22.4 % fossil / nuclear power. This is far from what was previously reported, and the results calculated using a designed method are flawed and completely unconfirmed.

Therefore, we believe that the myth that cryptocurrency mining causes environmental damage is fundamentally fallacious, and many miners, in their quest to find the most cost-effective sources of electricity in the market, have expanded their mining operations to areas around the world with an oversupply of renewable electricity.

This stands in stark contrast to common assumptions cited in the media about Bitcoin mining emissions and its subsequent negative environmental impact, with the argument predicting an exponential increase in demand for electricity, a fundamental misunderstanding of the mining process. Reports even assume that the incremental supply will be powered by traditional energy sources such as coal - which is itself wrong.

While we’ve seen a few studies that come pretty close to using plausible methods when estimating the carbon footprint of Bitcoin mining, even then the authors completely fail to take into account regional differences in the energy mix.

Don’t be evil vs. Be good

Cryptocurrency mining may actually be consuming excess capacity from the grid and even supporting its profitability, thereby supporting the growth of the renewable energy generation industry. While we can certainly say that cryptocurrency mining at least does not cause any incremental harm, and even proves that it is actually beneficial to the renewable energy industry by using stranded and wasted electricity, this is a greater challenge and requires deeper follow-up research work.

However, it is interesting to consider the following idea: Is the rise of cryptocurrency mining really a desirable prospect for the renewable energy industry? It is obvious that using optical fiber is cheaper than building a UHV grid. Therefore, the high mobility and low manpower requirements of mining equipment compared to its total power consumption make it ideal for monetizing distant and idle renewable energy sources, thereby increasing the profitability of such projects. We must also not forget that mining ensures the consensus mechanism of cryptocurrencies, which contain assets worth about $80 billion.

While we also find it rather ridiculous to attack a value-creating industry simply for the electricity consumption that is purchased voluntarily on the open market, if someone is making a false argument attack, we suggest that he would be better off targeting other power applications: for example, there are about 85 million PlayStation 4s , 40 million Xbox Ones , and 15 million Nintendo Wii Us in households around the world . Their gaming power consumption averages about 120W . Assuming they are connected to a 40 -inch LED TV, play for 4 hours a day, and idle for 20 hours with an average power consumption of 10W , these gaming systems ( 4.9GW ) consume more power than the entire

The Bitcoin mining network (4.7GW) is even larger.

in conclusion

Based on China’s energy structure and historical data on cryptocurrency mining farms, we found that, contrary to most opinions, most (at least 77.6%) of the world’s Bitcoin mining farms run on renewable energy. The overlap between regions with high renewable energy usage and Bitcoin mining farms is not accidental.

While data on the nascent industry is currently scarce, further research may prove that, at least in terms of the environment, cryptocurrency is not only not harmful, it may actually be beneficial.

In fact, Bitcoin mining could become the ultimate purchaser of electricity, wherever it is located, creating a highly liquid base demand for any electricity that can be produced at a price below the current cost of production. If demand for Bitcoin mining continues to increase, its power demand could itself help develop highly productive renewable energy sites in remote areas that are economically underdeveloped today.

Building on Nic Carter’s framework, we can visualize this idea by imagining the global energy grid as a 3D world map divided into grids . We set areas with cheap energy to be low in the terrain, while areas with expensive energy to be high in the terrain. Traditional residential and industrial demand can be thought of as blocks pinned to the map, which can only be moved with considerable effort. Bitcoin mining, on the other hand , due to its high liquidity, is more akin to a glass of water poured over the surface of the map, settling in the grooves, smoothing it out. This effect “liberates idle [ generation ] assets and makes new ones available.”