EVIDENCE BEFORE VERDICTSPublic edition · Checking pilot
THE CLEARER PICTURESeptember 2026Edition No. 003
Claims Lens 360
Big claims.
Better questions.
The whole picture.
Claims checked so far—Public tally not live yet

AI infrastructure & electricity costs · SOURCE RELEASE November 2, 2026

OpenAI launches grid plan. Savings remain forecasts.

Georgia’s example pairs future demand cuts with future customer savings.

Published November 3, 2026 · Reviewed November 3, 2026 · AI-assisted writer, evidence-editor and publishing passes. Not independent human fact-checking.

The quick read

OpenAI launched a national Grid Protectors Program November 2 for future campuses to cut demand when grids are strained. Its Georgia example is still being developed. Georgia Power describes up to 1,000 megawatts of flexible demand response and says OpenAI will pay its service and infrastructure costs. That is a commitment to managing peak demand, not evidence of bills already reduced. The utility’s projected $950 million in annual customer savings starts in 2029 and covers its large-load portfolio and expected growth, not this campus alone.

The national plan still needs local terms

OpenAI says utilities and operators will need location-specific agreements on how much demand a campus can reduce, how quickly and for how long. It identifies the planned Effingham County, Georgia, campus as a pilot. It also acknowledges that flexibility cannot replace investment in new supply and transmission. This is a program for future campuses, not an operating-performance report. (Sources 1)

Megawatts are a demand commitment, not annual energy saved

Georgia Power’s July 22 announcement describes up to 1,000 MW of flexible demand response under its agreement to serve the planned campus, alongside payment of its infrastructure and electric-service costs. That is a limit on power that could be curtailed, not a measured annual electricity saving. Duration and actual dispatch matter. The two companies’ descriptions of their agreement are not independent tests of its performance. (Sources 1–2)

The savings projection belongs to a portfolio

Georgia Power’s August 26 release projects approximately $950 million in annual customer savings beginning in 2029 from new large-load customers and expected growth. It does not assign the whole benefit to this campus. The utility argues that added revenue and flexibility can spread costs and reduce generation needs. That is the case for the arrangement, not a record of 2026 household bills lowered by an operating OpenAI campus. (Sources 3)

A protective rule is a tool, not the result

The Georgia Public Service Commission’s January 23, 2025 summary describes a rule allowing special cost terms, longer contracts and minimum billing for new large-load customers, with contracts submitted for review. Those are documented choices intended to protect other ratepayers if a project leaves before paying for infrastructure. The summary is not an audit of this signed contract or its future outcomes. The useful test is whether the obligations deliver under real grid conditions, not whether the announcement sounds reassuring. (Sources 4)

What this record cannot settle

Company commitments and utility forecasts, not observed demand cuts or household savings. The full signed contract, dispatch conditions, savings calculation and operating results were not inspected. No matched household-bill comparison or national cost outcome is established by these records.