Research Methodology
How SecondWatt sources, verifies, and publishes equipment pricing, technical specs, data center intelligence, and editorial research.
Sources and Citations
SecondWatt operates under a zero-hallucination rule: every published data point traces to a verifiable public source or to our own marketplace transaction data. Primary sources rank in this order — U.S. federal regulators (FERC, EPA, NERC, EIA, DOE), grid operators (PJM, ERCOT, CAISO, MISO, NYISO, ISO-NE), national equivalents in Europe and Asia (Ofgem, ENTSO-E, RTE, National Grid ESO), and OEM technical documentation. Secondary sources include peer-reviewed journals, national lab publications (NREL, LBNL, ORNL), and analyst reports with disclosed methodology. Anonymous sources and vendor-sponsored whitepapers are not eligible as primary citations. Citations render as inline prose links; a claim without a source link blocks publication.
Pricing Methodology
Price benchmarks combine public secondary-market price points (mostly dealer asking prices, plus filed sales where public), public OEM list pricing where disclosed, and procurement filings including FERC Form 1 and EIA Form 860. The Four-Tier Framework segments every category into New OEM, Gray Market, Refurbished, and Used, each reported as a midpoint with a 10–90 percentile band rather than a single number. Refresh cadence is quarterly for high-volume categories and semi-annual for low-volume categories. Seller identities, exact locations, and re-identifying condition filters are never published. Live tables sit in the Pricing Index.
Equipment Research
Model pages are built from OEM datasheets, certified test reports (ISO 8528, IEC 60076, IEEE C37), and field service bulletins, with every spec field citing a datasheet URL or the governing standard. Each model runs through a structured enrichment pass covering technical specs, common configurations, lifecycle notes, and a source URL verified before publication. Confidence levels are recorded per field, and no model is indexed until an editor approves the research. Only official manufacturer photography is used; AI-generated imagery of specific products is not permitted. See gas turbines, generators, and switchgear.
Modeling and Market Data
Technology comparisons in the Power Intelligence Tool score five generation options against identical inputs, discount cash flows at a stated cost of capital, price equipment lead time explicitly, and compute fleet availability from unit count and forced-outage rate rather than vendor claims. Data center market and operator coverage is documented in the DC Library, and published analysis lives in articles and reports.
Data center market and project evidence
Our publication policy requires market observations to be recorded separately from facility and project records. Each published observation must state its definition, unit, geographic or commercial scope, reporting period, publisher and source location. Historical observations must retain their dates when a later report is added.
Keep reporting-date snapshots, such as inventory and vacancy, separate from period activity, such as take-up and net absorption. Preserve the source’s definitions. A project’s announced design capacity, planned generation and operating capacity must retain their own labels. These different measures must not be combined into a single capacity total.
A source-check date records when the cited document was reviewed. It does not establish that a project is still in the same stage. Forecasts, roadmaps and vendor statements must remain attributed to their publishers. A forecast date passing does not change a project to operational.
Facility ratings, floor areas, commissioning dates, operator identities and power arrangements require evidence specific to the field. Where the evidence is incomplete, omit the field. Do not use a campus name, operator relationship or market assignment to fill an unknown rating.
Label SecondWatt procurement analysis separately from reported market observations. This analysis identifies questions and interpretations for further investigation. Establish site requirements and the offered equipment configuration for the actual sourcing inquiry.
SecondWatt holds no inventory. Buyers submit requests against equipment dossiers; sourcing partners register availability privately; introductions are anonymised. Public research must exclude seller identities, SecondWatt sourcing partners, private equipment stock quantities and internal pricing. A reference to an operator or project does not imply that its equipment is for sale.
Read the data center intelligence hub, the project tracker and the equipment dossiers as distinct parts of that research process.