I had the fortune (neither good nor bad) of being an energy analyst in 2008 as Lehman Brothers collapsed, in part triggering a global financial crisis. I was also an energy analyst, again, when a pandemic caused the mechanics and circulatory system of much of the global economy to seize up in 2020. Our current moment is equally intense, I think — and now, as then, I look for information that can help me make sense of where we are collectively. There are far better minds than mine for understanding, and explaining, moves in Treasury yields or calculating effective tariff rates (or divining intentions in Washington, for that matter). 

Discerning signal from noise is hard when the noise is really noisy, and important signals start out quiet. Here are those quiet signals I’m looking at this week while I absorb as much broader-economy smarts as I can from those who know it best.

As Halcyon customers (and readers of this blog) surely know well, US electricity demand has returned to growth for the first time in more than a decade. Much of that is artificial intelligence and data center-driven, but US manufacturing buildout has been part of the expectation as well.  And certainly, the data to support a manufacturing boom are there. One of my favorite data sets is the US Census Bureau’s monthly tally of construction put in place for every meaningful sector of the US economy. Manufacturing has been, for most of my career, a relatively staid sector for new build in the US — until the last few years, when it reached an annualized rate of more than $230 billion. 

But, that annualized rate has started to roll over, and I cannot imagine it reversing back to growth at a time of radical trade uncertainty. Planners take note: there’s a huge amount of incipient electricity demand in that annualized construction rate, but actual demand will depend upon factory utilization, not on structures being built. This is a quiet, but strong signal, and one that will continue to broadcast to us as well. 

That does not mean that electricity demand is likely to decline, however! Here’s one quiet signal I picked up this week through Halcyon, in a market which (I confess) I do not monitor closely: western Missouri and eastern Kansas. There, utility Evergy has just filed its latest electricity load forecast which (thanks to Halcyon’s new catalog of Missouri Public Service Commission dockets!) I can easily access and follow. Here’s its forecast for ‘net system input’ (effectively the power it expects to generate) in its mid case for the metro area in its service territory. See if you can spot the move from ‘no historical growth’ to ‘future growth.’

Evergy’s presentation says clearly why it expects this change: “Ongoing assessments of new loads and economic development opportunities for Missouri” and “Building in economic development load profiles to meet future growth.” Particularly noteworthy: this latest analysis is an update to its most recent triennial integrated resource plan published…a year ago! Within the last 12 months, expected growth for the Kansas City metro area has become  discontinuously higher than it has been for decades. 

An example of future ‘development load profiles’ emerged last week in Pennsylvania, where the site of what was once the state’s largest coal-fired power plant is being converted into a 3,200 acre complex of 4.5 gigawatts of gas-fired power plants serving data centers and in particular, “AI and high-performance workloads.” That’s a clear signal, and not a particularly quiet one either.  

And, it makes me think closely upon the White House’s recent executive order on “Reinvigorating America’s Beautiful Clean Coal Industry.” The EO highlights artificial intelligence as a driver of (potentially) keeping coal-fired power on line explicitly. Certainly, robust demand for power, highly concentrated in high-performance, relatively cost-insensitive use cases could support coal-fired power plants. But an executive order is, essentially, nothing more than a memo — a noisy, but not necessarily strong, signal.  Want the quiet, strong signal? You’ll find it in the data, and that data is often unstructured, verbose, and obscured in a public proceeding. 

The ability to distinguish between noisy, weak signals and quiet, strong signals separates those who react to changes from those who anticipate them. Given the rapid pace of change in our economic policies, this is easier said than done. But those willing to do the work (or, at least, have AI help) can tune their economic antenna to the frequencies that truly matter.

Comments or questions? We’d love to hear from you - sayhi@halcyon.eco, or find us on LinkedIn and Twitter