The Signal Stack

Every one of the six forces leaves a paper trail, and learning to read that trail is the empirical core of this discipline. Think of it as a stack of signal types, each with a characteristic latency, reliability, and failure mode. The art is knowing what each layer can and cannot tell you, and then reading up and down the stack so the layers check each other.

Start with interconnection queues. When a developer wants grid power, they file to interconnect, and that filing enters a public queue with a requested capacity, a location, and a date. Queues are among the earliest hard signals of intent, because you file long before you build. But they are noisy: queues are stuffed with speculative and duplicate requests, many of which will never energize. A queue entry tells you someone is serious enough to file; it does not tell you the project will happen. Read queue depth and withdrawal rates, not just the headline megawatts, and treat a single entry as a hypothesis, not a fact.

Next, permits and zoning actions. A rezoning application, a conditional-use permit, an air permit for backup generation, a water-draw application — these are slower than queue filings but harder to fake, because they cost real money and expose the project to public comment. When a queue entry is corroborated by a matching permit application in the same location and timeframe, two independent layers now agree, and your confidence should rise sharply. That is the corroboration move in miniature.

Then utility and grid telemetry. Real-time and near-real-time grid data — load, headroom, congestion, fuel mix — tells you whether the power a project needs physically exists on the wires it would touch. A gigawatt request into a substation with tens of megawatts of headroom is a request that implies a major transmission build, which implies years and contested cost. Grid telemetry is how you sanity-check whether a stated plan is physically plausible or merely aspirational.

Above that sits corporate and financial disclosure: capital-expenditure guidance, project-finance closings, tenant leases, supply agreements for transformers and chips, and the occasional regulatory filing that names a site. These are the most reliable layer because they carry legal and financial consequences for being wrong, but they are also the latest — by the time a lease is disclosed, the early movers already positioned on the queue and permit signals. Disclosure confirms; it rarely surprises the prepared.

Running through all of it is the people layer: hiring posts, executive moves, contractor headcount, and the quiet reallocation of specialized talent. This layer is the fastest and the least formal, which makes it both the earliest warning and the easiest to over-read. A single job post is weak. A cluster of related posts — substation engineers, commissioning leads, site-acquisition managers, all in one metro, all in one quarter — is a pattern, and patterns are evidence.

The discipline is not to rank these layers from best to worst and use only the best. It is to read them as a stack that must cohere. A real project casts a shadow on every layer at once: a queue filing, then permits, then hiring, then grid upgrades, then disclosure, roughly in that order, roughly on a timeline you can anticipate. When the layers agree, you have evidence. When one layer screams and the others are silent, you have a question, not a conclusion — either you are early and the other layers have not caught up, or the loud signal is noise. Both are possible, and telling them apart is the judgment this course is training.

Two practical cautions. First, each layer has a home-field bias: queue data over-represents speculation, disclosure under-represents the future, people-signals over-represent intent relative to execution. Correct for the known bias of each source rather than trusting or distrusting it wholesale. Second, latency is information. The gap between a queue filing and the matching permit tells you how fast a jurisdiction moves; the gap between hiring and disclosure tells you how far ahead of the market you are reading. Watch not just what the layers say but how far apart in time they speak, because that spacing is itself a signal about which constraint is binding and how long it will take to clear.