What our numbers are made of

No market currently reaches project-ready depth on all three required fields (recurring cost, connection, reliability), so we do not offer project feasibility conclusions anywhere. Local capacity is not assessed in any market, which is the ceiling on every project claim we could make: a market can look open while the substation you would connect to has nothing available.

Closest to project-ready, one field short: Northern Virginia (PJM) (needs reliability); Great Britain (needs reliability).

Our queue depths, withdrawal rates and typical waits come from generation and storage interconnection data, chiefly LBNL Queued Up. Those figures show how congested the interconnection system is. They are not a measure of how long a new large load waits, and a generation withdrawal rate is not a load withdrawal rate. The Load Pipeline page previously discounted a generation queue to estimate arriving load. It no longer does. Where an operator publishes a genuine large-load queue, such as ERCOT or the Dutch TenneT waiting list, we show that separately.

Every compiled cost band is priced for a single archetype: a 100 MW maximum contracted demand at a 95% annual load factor on high-voltage service at a disclosed point of delivery, where the nominated firm load survives loss of one supply element at the supplying substation with transfer within 30 minutes. Change the archetype and the number changes. The bands are comparable with each other; they are not a quote for a specific project. Customer-funded connection capital is a separate one-off cost and is not included in the recurring figure.

Our capacity read is a market-level estimate of usable capacity against peak demand. It screens system adequacy. It says nothing about whether the substation you would connect to has capacity available, and we do not assess local deliverability. A market can show room while a specific site cannot be served.

We publish estimated average operational carbon intensity from the stated generation mix, which supports location-based comparison between grids. It is not marginal emissions, not a lifecycle assessment, and not by itself a market-based Scope 2 result. Company clean-energy statements are reported as procurement commitments, not as descriptions of the physical grid.

A reliability label reports what the bulk and local outage record showed. It is not a statement about a contracted service level or a designed site, and an absence of identified flags is not an assurance of low risk. Where we have not assessed a market it reads not assessed, rather than clean.

Our archive is append-only in our system with published hashes: once written we do not edit a row, and corrections are added as new dated rows. We previously used a stronger word that implied an external control we do not have. Our data-quality review is a public self-assessment, not an external audit. The performance page is a baseline measurement plan until enough realized outcomes exist to score against.

Its components measure broad market conditions, not how hard it is for a specific new load to obtain power. The old name implied the latter. The recorded series is unchanged and continues without a break.