AI data and Theories of Change
No. 10 | December 9, 2024
A busy year approaches its close, and with it, the inevitable urge to look backwards and forwards.
So, a quick look back at my recent writing, much of which looks forward in its own way.
Halcyon alerts
First, Halcyon has launched its alerts service, which ingests and organizes document feeds from a set of US energy regulatory entities such as the California Public Utility Commission, the Virginia State Corporation Commission, the New York Public Service Commission, and the Federal Energy Regulatory Commission. It’s a start on what will soon be a much larger, LLM-enabled interface with what I consider to be a major rate-limiting step on decarbonization: reading paperwork.
You can sign up for alerts here.
Here is a post on why I think alerts are so important:
Any energy professional knows what sort of information will inform their decision making. They often know where it will come from: in the first instances, from company documents or more frequently, regulatory bodies such as commissions and agencies. But, that same professional may not know when it is coming (commission upload schedules set that tune). They also may not necessarily know who is providing the most important information (it could be a company, it could be a regulator, it could be a law firm, or even an engaged citizen). Awareness of an information challenge is high; discoverability of information is very low.
That frontier, though, is vast. Halcyon processes thousands of documents weekly just from institutions we already crawl and ingest, and we expect to scale that figure by orders of magnitude just to cover one part (electricity) of one country (the United States). Volume is a known known, but that doesn’t make it any less daunting.
Data center theory and practice
Also for Halcyon: a look at the theory and practice of data center electricity demand growth. I took time last week to dive into one regulatory proceeding in Virginia, where the Rappahannock Electric Cooperative is planning a new commercial structure for its potential data center customers. REC currently has just over 1.2 gigawatts of load in its entire service territory; it says that it has multiple projects under development in its territory that are individually larger than that figure. It currently has zero data center load; by 2040, it could have over three gigawatts.
So many discussions about future electricity demand are airless and abstract, which is why querying the source is so important and offers immediate insights:
An electric cooperative that currently meets a 1.2 gigawatt electric load, including zero megawatts of data centers, is planning to meet almost three times that much demand from data centers alone in 15 years. In order to do that, it is creating a new commercial structure to serve these customers in a way that both meets their needs, and protects other customers. It’s all in the filing, so to speak — just not all in one place, or in one document, or at one point in time.
Announcements and pronouncements are one thing, but technically informed, clearly labeled forecasts are another. The truest sense of Rappahannock Electric Cooperative’s future data center demand is not on its website, or in the plans of its future customers. It is buried deep in a filing, public but not hidden, but also not easily discoverable.
Theoretical discussions of data center power demand help drive awareness, and awareness drives policy. Energy supply initiatives from the world’s biggest data center operators expand the sense of what can be built, by whom, and by when. But, much of the real work is here: in filings and testimonies, in hard-to-find forecasts, in dockets and proceedings, and in legalistic-but-essential contracts and entity proposals. I think of it as a cross-referencing of theory and practice, with the practical result of allowing builders and investors to better understand what could work fastest, and in highest volume, and where.
The forecasters’ gap
In late September I received an email from my friend Azeem Azhar of Exponential View, asking for my take on current solar forecasters. I received it late, after a day on foot at New York Climate Week; cloistered in my tiny hotel room I wrote back him a few hours later with detailed notes. So detailed, apparently, that they became the germ of an essay exclusively for Exponential View readers.
I am happy to share it here. It’s a bit philosophical but nonetheless practical: a disentanglement of forecasts and projections, of analytical needs in different time frames, of how stretching from one realm (a short-term forecast) into another (a long-term projection) is flawed, and of theories of change. My conclusion:
I do not believe that analytical skills are the missing ingredient in thinking about the future. Rather, I believe that imagination about the future and a theory of change that helps to describe it clearly, are what is needed to look ahead in a more compelling way. It is not a lack of computing power, so to speak, but a lack of programming. The institutional frameworks and organisations responsible for the energy transition (the IEA and the EIA) are yet to adapt to the reality of rapid, exponential change. They have a responsibility to provide an accurate picture of the future so that society can plan for it.
This disconnect has real consequences. It affects policymakers planning infrastructure projects, investors allocating capital and companies devising strategic initiatives. Until we develop better ways to bridge this gap – aligning what analysts understand with what is formally forecast – we’ll continue to see the same pattern: reality consistently outperforming conservative predictions.
Decarbonization 2025
As the year ends, the data collection for my annual decarbonization deck ramps up. This year’s edition has 200 slides and I have more than that many drafted already - my challenge, and my duty, is be as comprehensive as possible without creating a laundry list. Perhaps a better analogy for the challenge is in motorsport, and the lovely phrase of Colin Chapman of Lotus: Simplify, then add lightness.
A preview slide: in one day in October, the Spanish city of Valencia received three years of rainfall.
Subscribers to this newsletter will get first notice when I publish.
If this not was forwarded to you, please sign up. I look forward to having you as a reader.
I am also setting my speaking schedule for the first half of the year.
If you are interested in a keynote drawn from my 2025 decarbonization deck, I would love to hear from you.
