资源描述
THE BITCOIN MINING NETWORKTrends, Average Creation Costs, Electricity Consumption 2) Deployment of next-generation mining gear at appreciable scale, predominantly in Sichuan, in line with the advent of the wet-season. The efficiency and dollar cost per TH/s both keep improving in line with the five-year trends Figures 1 and 2. These efficiency gains are contributing to a current all-time high of mining efficiency of 11.5 GH/J, up from 10.5 GH/J in November 2018 (+10%). We also note some jurisdictional developments in North America with consequences for local miners. In Oregon, miners have been facing unwelcome treatment both by local governments and electricity providers, leading to a mass exodus of miners from the state. On the other side of the US-Canada border, however, previous negative signals towards the industry from both governments and Hydro Quebec now seem to have reversed with both added clarity from regulators and less hostility towards miners on the part of utility providers. Lastly, we will briefly touch on the never-ending topic of Chinese mining trends. Over the last few years we have observed a steady trend of reduction of Chinese geographical dominance mining era). The 6 months passed can be roughly divided into two major eras: the final drop and subsequent bottoming-out of the hashratecoinciding with the capitulation-phase of the Bitcoin price curve, and; the return to growth and near-full-recovery of the hashratecoinciding with the recent rally in Bitcoin prices and the onset of the wet season in South-Western China. Meanwhile, Bitcoin prices have more than doubled from around $4,000 to $8,500 which has taken some pressure off the highest cost miners. During the same period, we have observed two substantial macro trends, coinciding with the two above inter-period eras: 1) A large number of bankruptcies, liquidations and ownership transfers of mining units, often to more well-situated and capitalised miners whose new capital stock has been acquired at a much lower cost basis than their previous owners, and; 2) The first at-scale deployment of the latest generation mining gear. It is also worth noting that the 40% drop in hashrate observed at the tail-end of 2018 represents the first time we have ever observed a substantial and prolonged drop in hashrate as a result of sustained large-scale corrections in the Bitcoin price. As we explained in our Medium commentary at the time 3, this would not lead to a mining death-spiral. On the contrary, the system acted exactly in accordance with its design, with the difficulty lowering perfectly in-line with the hashrate reduction, decreasing mining costs in accordance with the reduced Figure 1: Hardware Efficiency (GH/J) vs Shipping DateFigure 2: Hardware Cost ($/TH/s) vs Shipping DateFigure 3: Total Estimated Bitcoin Hashrate (EH/s)Sources: Bitcoin Wiki (May 2018), CoinShares Research (May 2019)Sources: blockchain (May 2019), CoinShares Research ( May 2019)Sources: Bitcoin Wiki (May 2018), CoinShares Research (May 2019)31 May 20194and brand new next-generation gear (potentially enabling depreciation over 2-3 years) are able to mine Bitcoin at less than $3,500/btc Table 3. This combination of circumstances becomes even more powerful if the miner has access to preferential pricing on their mining gear, such as miner-manufacturers Tables 1 British Columbia, Alberta, Newfoundland and Labrador, and Quebec Electricity Draw As of the time of writing, we estimate the total electricity draw of the entire Bitcoin mining industry to be approximately 4.7 GW. This is the same estimate as in November 2018, but includes a caveat. In our last report we estimated that miners on average used 20% additional electricityon top of that required by hashingfor cooling. We now understand that estimate to be grossly overstated and have reduced it to 10%. Meanwhile, the current amount of energy required for hashing alone is estimated to be 4.3 GW, up from 3.9 GW in November 2018. This result is also broadly in line with a 25% increase in hashrate and a 10% increase in gear efficiency. On an annualised basis, we estimate that the network currently draws the equivalent of 41 TWh. It is worth mentioning here that as a general principle, the Bitcoin mining network will consume as much electricity as the market is willing to sell it in return for the total value of the Additional Cooling Iceland; Northern Scandinavia (Norway and Sweden); The Caucasus (Georgia and Armenia); Yunnan and most importantly of all regions, Sichuan, provinces of China. There are also minor mining centres found in similar geographies such as Austria, Montana in the United States, Guizhou Province in China and the Siberian Federal District of Russia. The remaining major mining regions which do not fit into the above geographical mould are Iran, and Xinjiang and Inner Mongolia provinces of China. Minor mining regions where the above described geography does not (or where we cannot be certain that it does) fit the above description include: Florida, Texas and Arizona in the United States; Western Australia and New South Wales states of Australia; Belgium; Belarus; the North West Federal District of Russia; Argentina; Venezuela; and Israel. See map for list of sources. Energy Mix Building on our increasing visibility of the mining industry as a whole, we continue our ongoing reporting on the likely energy mix of the input energy in the mining industry. Again, we refer back to our previous reportsin this case page 5 of our November 2018 reportfor a more detailed discussion of the methodology and background of the investigation. For more casual readers, however, we summarise our methodology in simplified terms below. Our main assumption is that miners, wherever they are located, utilise the same mix of power generation (fossil/nuclear or renewable) as the average reported in their region. We then estimate the total percentage of hashrate residing in each relevant region, down to the lowest administrative subdivision for which reports of energy mix are available. Finally, we multiply the percent of renewables penetration in each relevant mining region with the percent of the total global mining industry residing in that region to arrive at a global weighted average estimate of renewables penetration in the Bitcoin mining networks total energy generation Tables 7 - 9. From the previous section (Geographical Distribution of Miners) readers will note that we divided the geographical clusters of Bitcoin miners into two main baskets. We will call the 8 31 May 2019therefore expect our estimates of total renewables penetration in the mining energy mix to vary somewhat with the seasons. Caveats and Uncertainty Factors It is necessary at this point to caveat that while we do our utmost to accurately pinpoint the location of global mining centres, the Bitcoin mining industry remains a highly private and secretive industry. As a result, our estimates may be subject to significant potential uncertainty. While we have made no attempt to formally quantify these uncertainly levels, we intuitively guesstimate that, e.g. our renewables penetration figures should be taken to include a tentative uncertainty of around 10%. That being said, we confidently consider our numbers to be amongst the best available in the industry. For other estimates using survey-based methodologies we refer readers to the following sources 48 49. Conclusion The Bitcoin mining network continues to develop along its five-year trend-lines on metrics of efficiency increase, investment cost reduction and hashrate growth. After having emerged from one of the most challenging price environments ever observed in such those in New York. The renewables estimate is down from 77.8% in our November 2018 report and reflects increased visibility of the industry on our part as well as movements within the industry. For example, we have observed a significant exodus of Oregon miners as well as an influx into natural-gas-dominated areas such as Iran. Seasonality Factors of the Energy Mix As we have alluded to in previous sections we continue to observe moderate seasonal mobility, especially among Chinese miners. We believe this mainly to be a result of the seasonal variability in rainfall, and consequently hydro power prices, in the Yunguichuan region of Southwestern China. As the annual Fengshui, or wet season, period manifests, electricity prices fall as low as 2.5/kWh, and generally to levels that are among the lowest in the world. Multiple sources suggest that more than 100 TWh of electricity could be wasted annually across these three provinces alone 44 45 46. When the dry season returns, electricity prices rise again, causing some miners to migrate to Xinjiang and Inner Mongolia where cheap coal and wind power is available year around. Some sources suggest that as many as 500,000 mining units migrated to and from Xinjiang last year 47. Migration, however, is an expensive endeavour restricted to the most well-capitalised miners with 7-figure (US$) relocation costs and 20% breakage rates reported. Combined, these migration patterns will cause seasonal variability in the renewables penetration of the Bitcoin mining industry. We 9Relevant Chinese RenewablesProvinces PenetrationSichuan (2017) 90%Yunnan (2017) 92%Inner Mongolia (2017) 16%Xinjiang (2017) 23%Average ex. Sichuan 44%Table 7: Chinese Renewables Penetration by ProvinceSource: Morgan Stanley Research (Oct 2018) Relevant Non-Chinese RenewablesCountries/States/Provinces PenetrationWashington (2016) 92%New York (2016) 45%Alberta (2018) 11%British Columbia (2018) 98%Quebec (2018) 100%Newfoundland and Labrador (2018) 95%Norway (2016) 99%Sweden (2016) 65%Iceland (2016) 100%Iran (2017) 0%Armenia (2017) 33%Georgia (2016) 79%Average68%Rest of the World18.2%Table 8: Non-Chinese Renewables Penetration by Country, State or ProvinceSources: EIA (Nov 2018), R2E2 (Jul 2017), Natural Resources Canada (Sep 2018), SATBA (Feb 2017) 31 May 2019
展开阅读全文