AI Power Translation
Where AI power demand becomes financially observable
A filing-based project asking where AI/data-center power demand becomes financially observable: in equipment orders, backlog, margin structure, or only in narrative language.
The first public phase tracks two nodes only: GE Vernova as the primary power-equipment case and Vertiv as the facility-infrastructure comparison. The signature metric is disclosure precision, not a valuation multiple and not an AI-attribution model.
Filing-based power-demand translation research
Summary
Current answer
The current evidence is not that GE Vernova's backlog is “AI-driven.” The filing-backed fact is narrower: GEV quantified orders and gas-power backlog/slot reservations for several periods before it quantified a data-center order figure. In 1Q26, the release disclosed $2.4B of Electrification equipment orders to support data centers.
- GE Vernova: broad power and electrification backlog first; quantified data-center order disclosure later.
- Vertiv: data-center and AI-infrastructure demand is named more consistently, but still not split into a clean AI-only revenue line.
- GEV pre-spinoff financial rows are combined/carve-out history; FY2025 is the first full standalone public-company year.
Core output
Disclosure precision, not attribution
The score increases only when issuer language becomes more specific. A rising score does not prove AI caused the order or backlog change; it shows that the company made the data-center connection more observable in public disclosure.
Disclosure Precision Timeline
The score tracks how precisely each issuer ties AI/data-center demand to orders, backlog, or financial metrics. It measures disclosure precision, not causality.
0 Not mentioned
1 Qualitative
2 Partially quantified
3 Fully quantified split
Latest signals
Where the two nodes stand
GEV
GE Vernova
The 1Q26 release moved from broad backlog visibility to a partially quantified data-center order figure.
- Period
- 1Q26
- Data-center orders
- $2,400.0M
Evidence note: GEV 1Q26 release reported orders of $18.3B, $13.0B sequential backlog growth including $5B from Prolec GE, Gas Power equipment backlog plus slot reservations growing from 83 GW to 100 GW, and $2.4B of Electrification equipment orders to support data centers.
VRT
Vertiv
Vertiv remains the cleaner data-center infrastructure disclosure case, but still at issuer level.
- Period
- 1Q26
- Data-center orders
- —
Evidence note: Vertiv 1Q26 release reported net sales of $2.65B, 23% organic sales growth, and Americas organic growth of 44.3%, with the Americas region described as expanding on strong data-center demand. The release did not provide a data-center-specific order or backlog split.
Company roles
Company evidence
This phase has two public nodes. It does not publish or preview a broader power-market roadmap.
GEV
GE Vernova
primary power-equipment and electrification node
Tests when broad power/electrification backlog becomes a quantified data-center order disclosure.
VRT
Vertiv
existing power and thermal infrastructure comparison
Shows a more direct data-center infrastructure disclosure trail, but still not a pure AI revenue split.
GE Vernova Annual Financial Context
Annual issuer revenue and operating margin from filing-backed rows. For GE Vernova, the basis changes at the spinoff boundary and should not be read as seamless standalone history.
Evidence notes
What the comparison does not claim
It does not claim that GE Vernova's orders or backlog growth are caused by AI. Generation, grid, industrial, electrification, services, and acquisition effects remain mixed.
It does not claim a pure AI revenue line for Vertiv. Vertiv's disclosures are closer to the data-center facility layer, but the published metrics are still issuer-level or broad order/backlog indicators.
It does not smooth over GEV's spinoff accounting boundary. Pre-spinoff rows are combined/carve-out or transition-year rows, and FY2025 is the first full standalone public-company year.
Scope
This is descriptive business analysis based on public filings. It does not provide investment advice, valuation targets, or trading recommendations.
Disclosure-precision definitions and source policy are documented in Analytics Methodology & Data Policy.