The Energy Transition Won't Happen

I keep seeing an article in my subscriptions, which asserts that an iron law of our universe is that foundational innovation is followed by an epic increase in energy consumption. The author, Mark P Mills, concludes that recent advances in AI, and its growing demand for energy, requires "a boom in building more natural-gas-fired power plants". Ergo, the energy transition won't happen.

If you've read the piece, how many pivots did you count in the opening paragraphs?

Energy or electricity?

First up, the pivot from energy consumption to electricity consumption. And not just electricity consumption, but local electricity consumption driven by the growth in EVs. Increased electrification of heat and transport will require investment in the distribution grid, but the efficiency of an EV compared to the equivalent ICE vehicle energy requires a fraction of the primary energy to deliver the same mobility.

Combine this efficiency with the increasing capacity for smart charging an EV at times when prices and carbon emissions from electricity generation are both low, and the role of the electrification of transport in the energy transition is evident.

Global or local?

The next pivot is from electricity demand in general, to electricity demand in the USA as a result of repatriating manufacturing. This might well lead to an increase in industrial electricity demand of 50% in the USA, but at the expense of demand in countries where electricity generation is more carbon intensive. And industrial demand in the USA has fallen to a quarter of electricity consumption, having flatlined for decades while other sectors have grown. Further, this consumption is typically found at higher voltages on the grid than the previously mentioned load growth from EVs.

In with the new...

The third cites a Reuters report that “U.S. electric utilities predict a tidal wave of new demand." The author claims that this confirms his 2013 study, forecasting the growth in electricity demand from the digitalisation of the economy. Of its time, the report is not half bad, correctly identifying the growth in electricity consumption from data centers and communications.

Out with the old

However, the study misses the trends and incentives for improving efficiency, both in computing and other sectors. Even before the advent of LEDs, lighting had declined from 38 to 17% of electricity consumption in the commercial sector, offsetting the rise in computing demand from increased digitalisation.

While the assertion in the 2013 study that coal "is the world's largest single current and future source of electricity" is half correct, it is a sizeable leap to the conclusions and title of that study, sponsored by the National Mining Association and American Association for Clean Coal Energy, that "The Cloud begins with Coal."

US electricity from coal has halved in the last decade

Although global electricity generation from coal has risen over the past decade, its share of consumption has fallen by 5%. In the US, coal generation has more than halved since that report, accelerating the decline.

Don't ignore demand response

Future trends in efficiency are likely to challenge the conclusion that more gas-fired generation just as they did the conclusion that more coal would be needed. Improving performance per watt is becoming one of the areas of competition between chip manufacturers.

Datacenters are leading the way in matching their consumption to renewable procurement through granular certificates, Power Purchase Agreements and scheduling tasks for periods and locations of cheap green electricity.

But Mills has been making extravagant claims since at least 1999's Dig more coal, projecting that "half of the electric grid will be powering the digital-Internet economy within the next decade", based on analysis that consumption for chips and computers took 13% of electricity generation. According to a study from EPRI, electricity consumption from data centres is only 4% of generation 25 years later.

Reliability is a system issue

The assertion that power from wind and solar is more costly and less reliable is challenged by the grid instability caused by the trip of a nuclear plant in Spain, or the simultaneous loss of two interconnectors in Ireland, let alone the repeated shut down of gas generation in Texas during extreme cold weather.

And the article also ignores the rapidly growing role of batteries in the energy transition, which are already displacing natural gas in ensuring system reliability in California and Australia.

Abundance not scarcity

Following the exceptional growth in solar generation last year, the abundance of cheap electricity could well be the biggest challenge of the energy transition. But it is also an opportunity that thousands of entrepreneurs are working to address, from grid enhancing technologies such as dynamic line rating, to aggregating the flexibility of small assets, whether smart charging EVs or using their batteries to power the grid, to siting data centers and hydrogen electrolysers in locations where renewable generation would otherwise be curtailed.

The energy transition won't happen with the kind of incumbent thinking that the only solution is to build more fossil generation. But if you're looking in the right places, there are plenty of signs that it is happening.