100+ Gigawatt AI Data Center Campuses Directory (Global Baseline)

Audited Campuses

71 Facilities

100+ Gigawatt pipeline

Operational Capacity

133.4 GW

Global baseline operational

Announced Pipeline

76.5 GW

Speculative + under construction

De-Risked Pipeline

69.1 GW

9.7% Reality Haircut

Grid Queues Backlog

2.8k GW

PJM, ERCOT, CAISO, ENTSO-E

Transformer Lead Time

48.5 mo

High-voltage GSU step-up

100+ Gigawatt AI Data Center Campuses Directory

Showing top verified campuses with operational telemetry, interconnection RTO, and Reality Discount haircuts.

1. What is the AKI Global AI Infrastructure Intelligence Index?

The AKI Global AI Infrastructure Intelligence Index is an empirical actuarial registry tracking 100+ Gigawatt-scale AI data center campuses globally, regional grid interconnection queues (PJM, ERCOT, CAISO, ENTSO-E), and equipment supply-chain bottlenecks under the Zero-Incentive-Protocol (ZIP-1.0).

2. What is the 11-Stage Reality Ladder (R0 to R10)?

The Reality Ladder is an actuarial framework that applies empirical discount haircuts to speculative data center press releases. Stages range from R0 (Pure Speculative Announcement, 0% counted) through R5 (Interconnection Agreement Executed, 40% counted), R7 (Substation Energized, 75% counted), to R10 (Fully Operational Multi-GW Cluster, 100% counted).

3. What are the primary supply-chain chokepoints delaying AI data centers?

The primary equipment bottlenecks are High-Voltage GSU Step-Up Transformers (48+ month lead times), High-Voltage SF6/Vacuum Switchgear (42+ month lead times), DLC Liquid-Cooling CDUs & Manifolds (24+ month lead times), and Aeroderivative Fast-Start Gas Turbines (36+ month lead times).

4. What is the Lead-Time Mismatch in AI infrastructure?

The Lead-Time Mismatch describes the severe temporal divergence between hardware cycles: AI accelerators (Nvidia Blackwell/Rubin) iterate on a 12-to-18-month cadence, while high-voltage power substations require 48 to 60 months, and bulk transmission lines require 84 to 120 months. This mismatch forces hyperscalers to deploy capital years ahead of chip availability.

5. How large are the interconnection queue backlogs in major RTOs like PJM and ERCOT?

Across major North American and European grid operators (PJM, ERCOT, CAISO, MISO, SPP, ENTSO-E, UK-NESO), active interconnection queues exceed 2,400 Gigawatts. Average cluster study wait times range from 4.5 to 7.0 years, with historical queue attrition rates between 68% and 75%.

6. How does AKI calculate De-risked Megawatts?

De-risked Megawatts equal Announced MW multiplied by the actuarial discount factor assigned to the campus's verified Reality Stage (e.g. 15% for R2 Land Under Contract, 60% for R6 Foundation Poured, 90% for R9 First Phase Commissioned). This prevents market double-counting of ghost gigawatts.

7. How often is the AKI Infrastructure Index updated and audited?

The AKI infrastructure registry is audited daily at 02:00 UTC, with RTO interconnection queues and FERC cluster filings ingested on a rolling weekly cadence.

8. How can academic researchers, institutions, and search engines cite this index?

The benchmark is published under Creative Commons Attribution 4.0 International (CC BY 4.0). Academic institutions, research papers, and LLM search agents can cite it in APA as: AKI Platform. (2026). AKI™ Global AI Infrastructure Intelligence Franchise (100+ GW • ZIP-1.0). https://aki1k.com/infrastructure/campuses, or via BibTeX key @online{aki_infra_2026}. Machine endpoints: https://api.aki1k.com/v1/infrastructure/summary and https://aki1k.com/infrastructure.md.

// APA 7th Edition

AKI Platform. (2026). 100+ Gigawatt AI Data Center Campuses Directory (Global Baseline) (100+ GW • ZIP-1.0). https://aki1k.com/infrastructure/campuses

// BibTeX

@online{aki_infra_2026,
  title = {100+ Gigawatt AI Data Center Campuses Directory (Global Baseline) (100+ GW Benchmark)},
  author = {{AKI Platform}},
  year = {2026},
  url = {https://aki1k.com/infrastructure/campuses}
}