Bitcoin Energy Consumption: A Breakthrough in Energy Liquidity
If our objective is to increase green energy production, we need to think as much about consumption as production. Renewable generation requires huge investment and long time horizons. Moreover, the economies of scale enjoyed by 20th century generation sources upon which the grid was built are dramatically different from the low economies of scale of distributed generation opportunities of the 21st. This is a structural mismatch between distributed generation sources and distributed consumption and poses major challenges to decarbonizing the grid. A reliable demand on distributed renewable energy provides a price floor during periods prone to surplus. As a consequence, banks will be more willing to invest in increased generation. At a first glance, fully commoditized proof-of-work cryptocurrency mining might seem like the perfect match.
A History of Transporting Stored Potential Energy
Near the turn of the 18th century, a handful of my ancestors thought they’d try their hand at a mostly tedious, occasionally deadly craft: harpooning sperm whales in the North Atlantic. They joined thousands of other sailors in the humble pursuit of lighting the streets and desk lamps of the West. As the years passed, the number of whales in the Atlantic dwindled and they sailed further and further to fill their ships’ holds with the era’s most valuable liquid. By the middle of the 18th century they would depart from Long Island, New York, sail around the tip of South America to Valparaiso, Chile where they would restock on provisions before heading out into the middle of the Pacific Ocean to spend up to a year offshore. Longer trips would take up to three years.
Lessons about the whaling industry today are often so focused on the genocidal absurdity that we forget the economic absurdity of spending up to three years on the open ocean and sailing halfway around the world to capture stored energy we now get from a cellphone LED flashlight.
Whaling doesn’t stand alone in its scale when it comes to the anachronistic transport of stored potential energy. An even larger industry, when counting the people employed, was the ice trade. In 1806 Frederic Tudor began to export blocks of ice from ponds in New England to the Caribbean. The industry grew through the 19th century with ice from New England shipped as far away as Bombay. The entire process required special tools for harvest, insulated transport and long term cold houses for storage. The financial risk of transporting perishable stored energy was enormous, yet at its peak, around the start of the 20th century, the US ice trade netted $660m (inflation adjusted) and employed more than twice the number of people currently employed in the coal industry.
Weirder still, these industries existed in tandem. Two distinct and simultaneous production, refining, distribution and retail networks for both ice and whale oil.
A Shift to a Centralized Electrical Grid
It wasn’t until advances in electricity transmission pioneered by Nikola Tesla (and coordination codified through regulation) that we had a sufficiently robust distribution network capable of replacing these truly massive industries with a single distribution network. Transporting raw potential energy was no longer the universal solution. The question became “what loads are best situated at the last mile and what loads are best situated near the point of generation?” The decisions that were made at this stage shaped the electrical grid we use today.
Uses that scaled down well (like lighting and refrigeration) shifted to the last mile of the network nearly overnight. Adoption of these, and other household appliances like irons, washing machines, radios were heavily popularized as electricity penetration rates rose to ~70% by 1930.
This was not the case for energy intensive industries. Early 20th century industrialization was dominated by vertical integration of firms and by extension, centered in locations where energy was cheap. The huge economies of scale enjoyed by industry expanded in tandem with the economies of scale of energy generation during the period of expansion; coal plants on site at factories and factories built next to large coal plants.
The economies of scale of generation aren’t as true for distributed generation as they were for the large plants the grid was built around. Distributed generation lets us more effectively collocate supply and demand regardless of the scale of the size of the demand.
Despite this huge advantage, there’s still enormous variability in electricity pricing depending on (among other things) country, proximity to production, weather, time of day, and other generation and consumption activities.
This geographic and temporal variability is a persistent arbitrage opportunity. The variability in pricing, or illiquidity of arbitrage opportunities, represents the inflexibility of the other components of industry (transportation, labor, risk of investment).
If you reduce non-electrical costs of your energy consumption, you’re able to take advantage of cheap energy whenever it’s produced and wherever it comes from.
Problems of Green Energy Production: The Duck Curve
The biggest advantage to renewable energy sources is also one of the largest obstacles to their widespread adoption: they don’t consume a raw material and have variable production independent of demand. The consequence is increased price variability across the entire grid.
While the marginal cost of producing an extra kWh of coal energy is proportional to the coal burned, the marginal cost of that kWh from photovoltaic or wind power is zero. Solar panels don’t get tired of photon flow and the cost of wind turbines is skewed towards their construction, not their maintenance. The rational renewable energy generator will therefore sell any energy produced at any price above zero.
Meanwhile, the cost of producing photovoltaic panels is plummeting and utility scale solar is exploding worldwide. The consequence is a surplus when renewables are available for production.
The mismatch is a timing imbalance between renewable power production during the middle of the day and the incidence of peak demand in the evening when people return home. As a result, flexible non-renewable resources like natural gas turbines need to be available to ramp up for the evening peak demand. This shape has been colloquially termed “the duck curve” and it’s the largest obstacle to the energy revolution needed to become carbon-neutral.
Is Proof-of-Work Cryptocurrency Mining The Answer?
Before jumping into this section I wan’t to define Proof-of-Work (PoW) in the simplest terms possible. A PoW algorithm is a problem that’s computationally difficult to solve and trivially easy to verify. In Bitcoin’s case, this is guessing the correct pair to a random number and being compensated with newly minted bitcoins. Since PoW mining isn’t the verification of transactions, but the issuance of new assets, it’s wrong to say x kWh of energy is used per transaction. Transaction verification is a separate process.* Through PoW mining, we create digital scarcity from real world scarcity.
Now, we want to describe the ideal commodity produced through the consumption of electricity which provide local generation access to global energy market prices.
First, the cost of commodity production should be nearly entirely derived from energy costs. Proof of work mining fits this mold well. The cost of transporting miners to the mining site is minimal and the cost of ‘transporting’ mined bitcoins to sell for fiat currency is zero. Labor costs are also tiny compared to overall investment. Hardware is by far the largest non-electricity cost. I’ll expand on hardware in a following segment.
Second, there should be minimal economies of scale. The commodity should be as (per Watt) efficient to produce at 100 Watts as it is at 10MW. On this front, bitcoin mining scores far better than existing industry, but is not perfect. Efficient miners use in the neighborhood of1kW, constant power draw. This is too much to nestle on the fringe of electricity production for super small producers, but orders of magnitude better than most industry.
Third, there should be low barriers to entry. On this front, Bitcoin does not currently score great, but is likely moving towards a new equilibrium. PoW encryption techniques can be solved much faster with specific hardware. Early in Bitcoin’s history, people shifted from CPU mining to GPU mining and then to Application Specific Integrated Circuits (ASICs) soon after. The current market share is dominated by Bitmain, who’s hardware accounts for roughly 80% of bitcoin’s total mining power. This dominance has earned Bitmain massive profits and in response, big competition is entering the game.
Key to understanding this point is that all PoW algorithms are at risk of ASIC development. Those with simple PoW algorithms, (like Bitcoin’s sha256) have low barriers to entry to developing new ASICs and do not become rapidly outdated through improvements in efficiency. I defer to Derick Hsue and David Vorick for extensive commentary on the subject.
Fourth, is capital depreciation. Will a piece of mining hardware go obsolete before paying for itself? Consider two scenarios. In the first, you have access to 24kWh a day priced at $0.03/kWh. In the second, you have access to 6kWh (1000W for a 6 hour window) at $0.01/kWh. If we project a constant demand for bitcoin at $0.10/kWh (within a margin of historical values and with hardware priced along Bitmain’s price curve),** the first miner will take two years to pay for itself under the first model and just above five years. This capital payoff rate is better than batteries, but during that time you’re exposed to bitcoin price volatility (further reason to favor ASIC resistant PoW) and a chance that more efficient miners will be produced.
2018 marks a new era in Bitcoin mining. Just as the ASIC market is evolving rapidly as new producers enter, new renewable generation sources will be brought online in 2018 for the express purpose of mining bitcoin and other cryptocurrencies. In these cases, we’ll likely see dual generation sources where miners take advantage of low solar prices during the day and use a secondary source during other times in order to keep miners running 24/7.
There are cases, however where an existing business may augment their revenue through onsite miners as a technique for increasing capital usage. One such application could be at industrial sites with large backup generation sources where those generation sources frequently sit idle. Since uptime in these cases would be below 24/7 there are some special considerations before we see large scale adoption for this application. For PoW mining to be useful as a non-core business it we should look for it to evolve towards an equilibrium where:
1) Improvements to miner efficiency are minimal (mining hardware doesn’t go obsolete)
2) Financing options become available for the miners themselves (eg, a bank wants a power purchase agreement in place before funding a large solar farm)
3) ASICs are either replaced by general purpose computation or the barriers to making ASICs are so low that they’re a commoditized product.
Improved energy liquidity isn’t limited to supply-side adjustments to the grid, but is a structural change in how information is shared across a network.
Final Thoughts: Energy Liquidity is Censorship Resistance
Consider Venezuela’s primary backstop to inflation: energy prices. In periods of hyperinflation, governments limit access to foreign capital to try to slow the erosion of their money’s value. The official exchange rate in Argentina in 2014 for example was ~7:1 while the real exchange rate fell between 9 and 13 to one. The Argentine government (and transitively banks) would sell pesos to you at that rate, but not buy them. By reducing access to dollars, the government reduced the rate of capital flight. Bitcoin obviously provided a critical use case across the board (and was the subject of my first writing in the space), but what I missed is how energy consumption fit in the picture.
Energy illiquidity is a primary tool of a government fighting inflation. The government has final say on the generation, transmission and sale of energy and in controlling the market, they are given a tool to assure residents that inflation is lower because they will sell you energy at that rate.
Individuals who have access to a global energy market (through the mining and sale of cryptocurrencies)are in effect exporting energy resources without the state’s permission. This is a fundamental change to the monopoly on energy centralized authorities have enjoyed through the electricity age.
If you can’t conceive of bitcoin as a useful store of value or medium of exchange, you probably haven’t experienced the conditions which make it a sensible option for most of the world’s population. Similarly, if you’re fretting about Bitcoin’s energy consumption, you’re probably isolated from how Bitcoin fits into existing market inefficiencies.
The energy revolution Bitcoin provides is global energy arbitrage not constrained by geography and not as prone to the same economies of scale of other energy dense commodities. Not only is this an inevitable development, it’s a boon to a greener energy grid and makes organizing bodies more accountable to their constituents.
*Though it should be noted that the purpose of mining is to ensure the integrity of the transactions within a block, not just creating new assets. Herein lie the basics on which the Bitcoin Cash philosophy is derived.
**Projection comes from miner profitability compared against the most efficient miner on the retail market. Mining is a highly opaque business and the number may well be off, but this estimate serves for the example. (https://blockchain.info/charts)