LBNL reference case; scenario range 9.5%–15.3% of U.S. use.
Electricity ledger →Public resource-competition ledger
Who actually pays for the AI buildout?
AI is reallocating power, chips, equipment, labor and capital. This site separates measured competition from convenient blame—with a source, date and attribution grade on every claim.
Crowding Index · calibration release
Not published yet
The basket is the intellectual property—and the easiest place to manufacture false precision. A scored value publishes only after series coverage, control matching, weighting and revision rules pass the public tests.
See the provisional basket and release checklist →*The 40% claim figure describes this five-card launch sample, not prevalence in the wild.
DOE’s 2024 distribution-transformer range; large units can reach 3–4 years.
Equipment ledger →Micron’s reported HBM-to-DDR5 production trade ratio.
Memory ledger →GE Vernova backlog plus slot reservations at Q1 2026.
Turbine ledger →Latest resource grades
Where the evidence is strongest
SemiconductorsAI-driven
Memory (DRAM & HBM)
3:1HBM-to-DDR5 capacity trade ratio reported by Micron
AI memory is directly tightening the broader DRAM supply envelope.
Open evidence →SemiconductorsAI-contributing
Storage (NAND & SSDs)
+30.4%BLS storage-device PPI, Jan 2024 to Jun 2026
Storage prices have turned sharply, but AI is not the only buyer in the queue.
Open evidence →Grid equipmentAI-contributing
Transformers & Switchgear
1–2+ yrdistribution-transformer lead time in 2024
Data centers add demand to a transformer shortage they did not create.
Open evidence →Generation equipmentAI-contributing
Gas Turbines
100 GWGE Vernova backlog plus slot reservations, Q1 2026
AI is reserving generation equipment inside a much broader power-build cycle.
Open evidence →Claim report cards
A permalink for the argument
AI-driven“AI is driving the return of U.S. electricity-demand growth.”Supported at the national demand level→Contested“AI raised my household electricity bill.”Possible locally; unsupported as a national blanket claim→Scapegoated“AI caused the transformer shortage.”Overstated: AI adds demand to an older, multi-cause shortage→
How a grade is made
Four steps, one named human
- 1Find the claim
Capture the common wording, not a softened substitute.
- 2Trace the mechanism
Name the market, competing buyer and binding constraint.
- 3Check primary evidence
Prefer agencies, filings and measured series over commentary.
- 4Human approval
Automation drafts and diffs; the editor owns publication.
Changes, not content churn
Every revision is dated and diffed.
Grade flips, source replacements and data revisions live in one public record. A quiet correction is still a correction.
Read the changelog →Alerts
Email opens with the first scored release.
Until the double-opt-in pipeline is live, the Changes page is the canonical release record. No placeholder form collects addresses.
Follow releases →About this project
Pro-growth is not the same as cost-blind.
Maintained by AI architect David Veksler. Automated watchers may draft future updates; every publication requires human review. Affiliations and the capital-markets conflict are disclosed in full. No ads, sponsorships or affiliate links. About & disclosures →