Bitcoin mining as a revenue increasing mechanism for Eskom
This article proposes a solution to Eskom’s revenue shortage problem in the form of a cryptocurrency mining farm. A Bitcoin mining farm has a unique combination of characteristics in that it is a highly profitable, electricity intensive business that scales linearly with capital investment. The lead time for full commercial operation is a matter of months. The ability of Bitcoin mining hardware to start up and shut down in less than 60 seconds can allow Eskom to constantly run its generation fleet at its most profitable supply profile.
The highly liquid Bitcoin market will provide Eskom with the option to convert any surplus electricity to cash in real time, which opens a whole range of additional opportunities, such as short term interest free loans which Eskom has had to revert to in the past
Background: Eskom’s revenue shortage problem
Eskom is South Africa’s electricity public utility. The organisation consists of four business units, namely Generation, Transmission, Distribution and Eskom Enterprises. The cash strapped utility is technically insolvent and in desperate need of additional revenue. This required revenue however, cannot be obtained during peak periods (daily between 18:00 to 22:00), since it already runs at maximum generation capacity. Ideally, the essential revenue should be created during off peak periods, but this proves to be a major challenge since almost no power hungry industry has the ability to shift its electricity demand in favour of such inconvenient times.
Eskom has an installed generation capacity of 47 040 MW, of which 70% is contributed by coal fired power stations, 5% nuclear and the balance by renewables, gas and peak stations.
The figure below illustrates the typical electricity demand profile for South Africa. One of Eskom’s biggest challenges is to meet the winter peak demand of around 34 000 MW, which is 80% of its installed baseload capacity. Since Eskom can only achieve an average Energy Availability Factor (EAF) of between 70% and 80% (due to planned and unplanned maintenance amongst other factors), they are not always able to meet this demand. A series of emergency intervention procedures then follow, of which “load shedding” is used as a last resort to prevent catastrophic failure of the electricity network. South Africans have become used to this since 2009, when the lights first went out due to a shortage in supply.
Part of the issue with such a demand profile actually lies in the off-peak periods. Building enough electricity generation capacity to meet the peak demand is expensive. The available options exclude renewables such as solar power since the sun doesn’t shine between 18:00 and 22:00, and wind energy is also not reliable enough since there is no guarantee that the wind will blow every night. Eskom thus reverted to pumped storage as a mitigating system, but these plants are limited to specific geographical areas, takes minimum 10 years to complete and cost billions of dollars in capital. Another option that Eskom frequently employs during peak times is using gas turbines, but the cost of generating power with these is at least 10 times as much as the selling price, meaning that Eskom can wipe out a week of profit for running just 8 hours of the gas turbine fleet.
There are really only two sustainable solutions to this problem: Lowering power demand during peak periods or building a bigger base load so it can absorb the peaks. Eskom has already shown that it cannot effectively manage the demand side, since political shenanigans prevent them from installing smart meters or cutting power to defaulting customers. Eskom acknowledged this fact and are currently undertaking a massive expansion in base load capacity through the Kusile and Medupi fossil fuel power stations, at a cost of >R300 billion to the South African tax payer.
But expanding base load capacity to accommodate only 4 hours of daily peak demand brings numerous other problems. The electricity is not needed for the other 20 hours of the day, and since base load power stations cannot be “two shifted” (shut down for short periods), they remain on load and effectively produce power at a loss because there’s no market. This completely destroys any financial justification for building these new power stations, but since the term “load shedding” has become such a growth inhibitor to the country, rationalisation can still be found from a macro-economic point of view. Eskom and the South African government now has the responsibility to try and lower peak demand in order to attract investment and/or increase off-peak demand to produce more revenue.
Eskom’s base load generators run at minimum capacity during off-peak periods, but the operational cost to power production relationship is non-linear meaning that they effectively operate at a loss during these periods. This is because all operating costs other than primary fuel stay constant for lower generating loads. If Eskom were able to increase power demand to its optimal generator load point during these periods, it will result in such a revenue surge that all of Eskom’s financial issues will be solved. Furthermore, if the peak demand is too high for the installed base load, Eskom has to add emergency peak generation capacity which produces power at 10x the price of coal and nuclear. Eskom can thus only profit from electricity sales if the total demand is within a certain narrow band roughly between 26 000MW and 31 000MW, as depicted in the image below (not exact numbers). Eskom is most profitable at the upper end of this band since that it also the most efficient point of base load generation. It would therefore ideally want a market which consistently uses about 80% of its installed base load power, 24 hours per day as illustrated by the green line in the figure below. (Note: These numbers were estimated from the author’s experience and he could not find any published research to back the exact power demand band in which Eskom profits)
Increasing power demand for only 20 hours per day is almost an impossible task, since there is virtually no electricity intensive industry that has the ability to shut down operations every day and start it again just before midnight. Since Eskom produces electricity at a loss during these off-peak periods, they may even have an appetite to sell power at cost to such an entity. We have seen government stepping in and adding tax-breaks to such businesses in the past, but the lack of interest shows that companies with such a unique electricity demand profile is a rare phenomenon.
Such an industry will have to be able to meet the following criteria to effectively provide a solution:
- It must have a high power demand. Eskom typically has between 4000MW and 12 000MW of spare electricity available during off-peak periods.
- Eskom must be in partial or full control of the electricity supply, with the ability to increase or decrease supply substantially within a matter of minutes with no warning.
- It must be quick to build the plants. South Africa’s electricity demand can change rapidly, and predicting the profile 10 years in advance has proven to be impossible. Eskom has made the mistake in the past to provide long term incentives to aluminium smelters based on inaccurate predictions and they are now paying the price.
- Ideally, the industry must have no negative environmental effect other than the high electricity production. This includes air, water and visual pollution.
- Ideally, the industry must be labour intensive, create local skills and source all equipment & support locally.
- Ideally, the industry must not be restricted to certain geographical areas such as cities, coastal, remote areas etc. In other words, the industry must be located where Eskom needs it to be (typically where jobs are required and where transmission infrastructure is already available).
- The industry should not be dependent on long-term power supply contracts. Eskom cannot afford to lock itself into another suicidal agreement.
The ideal solution would be for Eskom to own and operate its own electricity intensive business that would be able to convert the surplus electricity in blue as portrayed below into revenue:
This is where cryptocurrency mining can be utilised as a solution. For the purpose of this article, the author focused on Bitcoin mining since numerous studies have reported results of electricity usage and financial gains, but the arguments are valid for any proof-of-work (POW) based crypto-assets.
Background: Bitcoin Mining
Bitcoin can be described as a form of decentralised, peer-to-peer money. This is achieved with a single, distributed global ledger containing all Bitcoin transactions that has ever occurred and gets updated roughly every 10 minutes. It can also act as a platform for building applications which require settlement finality between parties that do not trust each other, without any intermediaries. This trust is provided through the use of a consensus algorithm called proof-of-work (POW), which requires anyone who wishes to add entries to the ledger to stake electricity as a promise to submit accurate information in return for a bitcoin reward after other nodes validated the submission. The act of staking electricity in the form of computational power in return for a reward of a limited commodity is called “mining” because of its close relation to precious metal extraction.
Bitcoin mining shares the following characteristics with gold mining:
- There is finite quantity of gold. Bitcoin also has a limited, known quantity
- It becomes more expensive to extract gold as time goes on, requiring deeper excavation. Bitcoin follows the same pattern, with the supply halving every 4 years.
- Gold mining assets depreciate as new technology is developed and the precious metal is depleted from the ore. The same applies for Bitcoin mining hardware.
The difference between gold and bitcoin mining however, is that a gold mine has major minimum barriers to entry while a Bitcoin “mine” has virtually none. A gold mine takes years to plan, design and construct before revenue is generated, after which the mine needs to be rehabilitated and leaves permanent pollution effects. A bitcoin “mine” requires no permission, can be built within a matter of days and needs no rehabilitation after decommissioning. Whereas a gold mine has a minimum investment of millions of dollars, a small Bitcoin mine can be built with about $800 and effortlessly scales linearly with capital expenditure.
Bitcoin is mined with specialised computer chips called Application Specific Integrated Circuits (ASIC’s). These processing units are designed to be specialists in solving cryptographic hashing algorithms required for mining bitcoin. The hardware cannot be used for any other purpose.
The Bitcoin mining difficulty is a protocol variable that adjusts the mining conditions every two weeks to make it easier/more difficult to mine. This is done to ensure that new bitcoin is minted on average every 10 minutes in order to guarantee a stable, deflationary, predictable monetary supply of the currency. If a large entity such as Eskom would thus build a mining farm and own a significant share of the hash rate (mining power), the difficulty would increase so the Bitcoin reward expressed per computational power would decrease. When that happens, mining farms with high operating costs would become unprofitable and shut down. This results in a balance which guarantees a profitable business model to any entity mining with lower than average operational costs. Currently, the breakeven point is at an electricity price of R1/kWh. Eskom finds itself in the unique situation that it can sell off-peak power at much less than that (<40c/kWh for most power stations), meaning that the mining of Bitcoin will remain profitable for the foreseeable future. The possibility of converting surplus electricity to revenue in real time (less than 12 hours turnaround time), might even incentivise Eskom to mine at a loss.
The table below compares a Bitcoin mine with a precious metal mine in terms of Eskom’s off peak demand levelling issues:
The numbers
Feasibility calculations are based on the following assumptions:
- Total Bitcoin network hashrate equals 51 201 325 148 GH/s
- Mining difficulty equals 7 152 633 351 906
- All of the mining is done with Antminer s9’s
- The average Levelised Cost Of Electricity (LCOE) for Eskom is R0.80/kWh. Note that effective demand profile management will result in higher energy efficiency and lower Operating & Maintenance (non-fuel) costs per kWh, therefore decreasing the LCOE substantially since all generators will run at its most effective point. 80c/kWh is taken as a worst case.
If Eskom were to build a 4000MW Bitcoin mine to operate for 20 hours per day during off-peak periods, it would cost them about R16 billion in CAPEX. If all the network variables remain constant, this mine would produce 1259 BTC per day, or R45.9 billion in annual revenue at R100 000 per BTC. Operating costs would consist almost entirely of electricity consumption, at an annual, higher end rate of R29.2 billion. EBITDA after year one would be R22.6 billion with an annualised ROI of 186% and an IRR of 205%.
Conclusion
Using Bitcoin mining as an off-peak electricity profile leveling mechanism to increase revenue makes financial sense for Eskom. The unique combination of a business which scales effortlessly, consumes vast amounts of electricity, runs at a healthy profit margin and having the ability to start up and shut down within a matter of seconds allows the mining of Bitcoin to be the perfect tool for a utility like Eskom. Such a plant can be constructed within months and require less than 5000m2 of ventilated warehouse space, virtually anywhere in South Africa.
The added advantage of having the ability to convert surplus electricity to revenue in real time as opposed to dealing with the risks of late payments from conventional electricity sales opens a vast range of additional opportunities. Eskom might, for instance be willing to mine bitcoin at a loss (with expensive peak generation for example) in return for generating short term liquidity when it desperately needs it. Electricity can effectively be converted to revenue within 12 hours using this strategy, which can be used as a short term, interest free loan.
A 4000MW plant would cost R16 billion to construct with an ROI of 186%.
Investment risks include bitcoin price volatility and mining difficulty increases. But these risks can be closely monitored and assessed by building a fully scalable pilot plant for as little as R10 000.
This article was researched and written in the author’s personal capacity.