Something is going to make the planning of regional power generation and associated markets even harder: the ability of big data users to shift their demand across the globe quickly, based on price. Ira Joseph at CGEP peers into the near future to ask some difficult questions. When such power demand can “at the push of a button” migrate long distances, how do you make predictions for local capacity needs? And today, market design in most countries is based on heavily regulated power markets. That design depends on reasonably predictable demand expectations, with prices determined by regulated tariffs, mixed with government subsidies. These government tools allow them to use its incentives and disincentives in a targeted way. Will these tools cope with such mobile big data consumers? There are some possible positives. Such flexible demand can reduce the need for storage, though figuring how much will be tricky. As this type of migratory demand based on price grows alongside big data’s rise, estimating local peak usage and what investments should be made to meet this peak will become much more difficult, says Joseph.
[This post was originally published by the Center on Global Energy Policy at Columbia University’s SIPA. You can find more insights on their blog, “Energy Explained”]
Incremental power use is moving around the world
The idea of a global electricity market has always been a bit of a misnomer.[1] Fuels such as gas, oil, coal typically travel around the world to fuel power generation; the sun travels ever farther through space and time. In contrast, electricity itself travels on a regional basis. Most transmission systems are local in nature and at most, stretch over several hundred miles.[2]
Two factors are changing power markets with globalisation, emerging at both the supply and demand end of the value chain. At the supply end is China, which is upending the status quo in power generation by exporting copious amounts of renewables infrastructure. What is happening here is transparent and clear: China is using renewables, regardless of profit, to take its seat at the global energy table.[3] At the demand end, the focus on intermittency and reliability risk has been unrelenting.
Much less focused on has been the ever-increasing mobility of the power demand itself. From global data centres and cryptocurrency miners to artificial intelligence (AI) users, incremental power use is increasingly moving around the world on a real-time basis in pursuit of the cheapest kilowatt-hour (KWh). As it does, it is undermining traditional planning on load growth and making it more difficult for policymakers to assess risk to their constituents.
Power demand globalises at the push of a key
Power demand is becoming more fluid in its pursuit of the least expensive place to operate, as the needs for instantaneous computing transcend traditional physical boundaries. Traditional uses of power in the industrial, residential, and commercial sectors had been dictated by the fixed location of the user, while pinning down use in the world of data centres or AI is not geographically anchored.
Herein lies the conundrum. Two worlds in energy are colliding and struggling to find a common language to communicate: on the one side sits commodity-driven energy markets, and on the other sits government- or utility-driven power markets. While traditional energy sources such as oil, gas, and coal remain essential commodities, the rise of renewable energy technologies and what is now migratory power demand is disrupting established market dynamics.
Is electricity a commodity or a utility?
The question of whether electricity is even a commodity, as opposed to a utility, is one worth asking, as a wide range of market designs exist around the world that suggest either definition could apply.
Until recently, it was the fuels for power generation – oil, gas, and coal – that moved around the world as a commodity, while the electricity would be produced at the destination of these fuels and distributed locally. While this structure is still true, the use of the power is now beginning to migrate, not in physical form, but in terms of use cases. If the cost of power becomes too high in the given market, the load tied to all forms of computing is shifting to less expensive areas. Conversely, when power prices drop into negative numbers as is now happening more frequently, these same loads are migrating into markets.
One could argue that the market forces attracting and repelling such types of power use will help counter the riskier aspects of relying on intermittent power by chasing away use at times when prices are highest and peak providers such as batteries on gas units step into the supply void.
Migratory power demand based on price
As this type of migratory power demand based on price becomes a larger and larger portion of power demand, pinning down what will be peak usage and what investments should be made to meet this peak may become much more difficult to assess.
The speed at which demand can appear and disappear from a local grid must now be factored in, along with more extreme weather as conditions adding to the grid’s stability. Record amounts of demand will be appearing and disappearing on a real-time basis when based on price, and generation will have to adjust to balance this increased amount of volatility.
Market design today
In most power markets around the world, the role of power stretches well beyond the marketplace. Market design in most countries in the world operates based on heavily regulated power markets, where the price is determined by regulated tariffs, mixed with government subsidies that allow governments to theoretically use electricity both positively as a social welfare tool and negatively as a political weapon. Market design options include vertically integrated utilities, wholesale power markets, retail electricity markets, capacity markets, spot markets, and most recently renewable energy certificate markets.
Highly flexible demand: an alternative to storage?
As excess solar and wind capacity pops up around the world, so will power use tied to AI and data centres, as it offers flexible demand. The caveat is that it also offers an alternative to storage, which suggests that a balance will need to be struck among use cases to keep the lights on, the EVs running, and the natural language processing.
***
Ira Joseph is a Senior Research Associate at the Center on Global Energy Policy at Columbia University SIPA
This article is published with permission
Notes
[1] The following paper comes out of a keynote presentation delivered at Penn State University on April 15. The conference was “The Geopolitics of Cross-border Electricity Grids.”
[2] While power is capable of flowing long distance, Typical voltages for long distance transmission are in the range of 155,000 to 765,000 volts to reduce line losses. A typical maximum transmission distance is about 300 miles (483 km). The longest high voltage AC transmission line in the United States is the “Path 27” transmission line, which runs from the Pacific Northwest to Southern California. It spans approximately 846 miles and is a crucial part of the Western Interconnection grid, helping to transmit electricity over long distances.