Adam Williams
Articles and tools on Data Center Scope are published under the editorial name Adam Williams.
Adam's editorial focus is data center economics and infrastructure: construction budgets, colocation pricing, power, energy efficiency, cooling, high-density AI infrastructure, reliability and capacity planning.
Data Center Scope does not attribute engineering licenses, certifications, employment history or operating experience to Adam that has not been independently established. The publication's authority is instead built around documented evidence, transparent assumptions and calculations readers can reproduce.
Editorial questions, corrections and source suggestions can be sent to [email protected].
Research scope
Data Center Scope concentrates on questions where engineering design and economics intersect.
Cost per MW, construction budgets, operating costs and project timelines.
Retail and wholesale pricing, committed power, ancillary charges, contract structure and TCO.
Electricity cost, IT load, total facility demand, distribution and power availability.
Air cooling, liquid cooling, heat rejection, water consumption and thermal density.
GPU rack power, high-density design and infrastructure cost implications.
1. Define the boundary before using the number
The first question is usually not “what does it cost?” but “what exactly does this cost include?”
Construction estimates may or may not contain land, utility interconnection, shell and core, electrical infrastructure, cooling plant, IT hardware or financing. Colocation prices may include electricity or bill it separately.
Figures with different boundaries are not treated as directly comparable simply because both are expressed in $/MW or $/kW.
2. Source hierarchy
Source selection depends on the claim being tested. Data Center Scope generally gives greater evidentiary weight to sources closer to the underlying data or engineering specification.
Government statistics, utility information, regulations, standards and public datasets.
Manufacturer specifications, engineering reference designs and original technical publications.
Established data center, construction, real estate and infrastructure research organizations.
Pricing datasets, operator disclosures, procurement evidence and documented project information.
Used for context when useful, but not preferred over stronger evidence for material quantitative claims.
3. Cross-check material claims
When a number is central to an article, the source is checked against other credible evidence where practical.
Disagreement between sources is not automatically a problem. It can reveal different project boundaries, dates, markets, capacity sizes or design assumptions.
4. Derived calculations must be reproducible
Data Center Scope frequently derives useful figures from published inputs rather than quoting a source verbatim.
Examples include converting MW into annual GWh, estimating the electricity impact of PUE, normalizing construction budgets to $/MW, or translating water-use effectiveness into annual gallons.
The assumptions required to reproduce a material derived result should be visible in the article or tool.
5. Calculator methodology
The Data Center Scope calculators are planning and comparison tools. They are not engineering designs, quotations or financial forecasts.
Default values are examples or documented reference points. User-controlled assumptions remain visible so the result can be recalculated for a different facility.
The current tools include the Colocation Cost Calculator and the PUE & Energy Cost Calculator.
6. Precision should match the evidence
Data center projects are highly site-specific. Climate, utility availability, redundancy, equipment lead times, procurement strategy, cooling architecture, density and contract structure can materially change the answer.
Where the evidence supports a range rather than a precise value, Data Center Scope prefers the range.
Apparent precision is not treated as higher-quality analysis.
7. Dates and market context matter
Construction markets, colocation capacity, electricity tariffs, hardware density and equipment lead times can move quickly.
Time-sensitive figures are therefore attached to the relevant period wherever possible rather than presented as permanent facts.
8. Corrections and material updates
If a calculation is wrong, a source was interpreted incorrectly, or stronger evidence materially changes the conclusion, the page should be corrected.
Minor copy edits do not necessarily require a visible correction notice. A change that materially alters a conclusion, calculation or important factual statement should be handled transparently.
Send correction requests to [email protected].
9. Advertising and editorial independence
Advertising can support Data Center Scope financially, but it does not determine which companies are mentioned, which benchmarks are selected or what conclusion an article reaches.
Paid placement should not be presented as independent editorial analysis. If sponsored material is ever published, it should be clearly distinguished from ordinary editorial content.
10. What Data Center Scope is not
Published material is general research and educational information. It is not a substitute for project-specific engineering, procurement, tax, legal, investment or utility advice.
A real project should be evaluated using its own technical design, utility terms, contractual requirements and local conditions.