Relations
A canonical map of the adikumar.co corpus for AI agents and research tools. Structures the relationships between series, frameworks, essays, topic hubs and reference pages as JSON-LD. The individual essays remain unstructured practitioner analysis; this page tells an answer engine how they relate to each other.
Series
Eight active series form the core of the corpus. Each series has a hub page linking to member essays and a specific analytical thread.
The DC-DC Transition
A 19-piece flagship series (13 core essays + 6 supplements A-F) on the 800V DC data centre transition. Covers architecture, standards, retrofit vs greenfield, vendor economics.
The AI Power Chain
A six-layer market + engineering map of AI data centre power delivery, from grid interconnection through package-level delivery. Six stack essays + six technical companions + three cross-cutting analyses.
The Investment Layer
Eight essays on AI infrastructure capital allocation, sponsor economics, M&A + debt underwriting.
Due Diligence for the AI Buildout
A sixteen-essay commercial + technical due diligence framework for AI infrastructure acquisitions, workstream by workstream.
Financing the AI Buildout
Ten-essay series covering capital structures deployed to fund the 2025-2030 AI infrastructure buildout.
The Regulatory Layer
Ten-essay field guide to the rules shaping the AI infrastructure buildout: FERC/ISO, PUCs, EPA, CFIUS, export controls, EU AI Act/DORA, standards, labour.
Tech Spotlights
Fifteen-essay series of vendor + technology deep-dives across the AI Power Chain.
Case Studies
Nine landmark AI infrastructure deal retrospectives, covering Blackstone/QTS, Microsoft/TMI, Amazon/Talen, CoreWeave, Iron Mountain, Digital Realty/Teraco, FTC/Nvidia-ARM, CMA cloud and Eaton/Boyd.
Frameworks
Named analytical frameworks that structure the work.
The AI Power Chain framework
Six-layer model of AI data centre power from grid interconnection to on-package delivery.
- Interconnect
- Facility distribution + DC transition
- Rack shelf + wide-bandgap
- On-package delivery
- Thermal
- Modular build
The Due Diligence framework
Sixteen-workstream DD approach for AI infrastructure acquisitions, from scoping through post-close validation.
The Investment Layer framework
Eight-essay capital allocation, sponsor economics and covenant structure framework for AI infrastructure.
Canonical theses
The five canonical theses run through the essays. Each is supported by specific analytical work.
The binding constraint on AI compute through 2030 has shifted from silicon supply to power infrastructure
- /the-architecture-map/
- /the-interconnect-stack/
- /the-thermal-stack/
- /what-changed-july-2026/
- /grid-queue-thesis-sept-2026/
AI data centre power delivery is compressing from facility scale toward the accelerator package
800V DC is being deployed as two distinct architectures under a shared label
- /the-real-reason-data-centres-are-going-dc-and-its-not-capacity-alone/
- /two-architectures-wearing-the-same-name/
- /map-mvdc-lvdc-architectures/
Cooling architecture is migrating from facility infrastructure toward the accelerator package
- /the-thermal-stack/
- /the-thermal-stack-technical-companion/
- /tech-spotlight-06-coolit-direct-to-chip/
- /dlc-vs-immersion-cooling-ai-data-centres/
Value concentrates in qualification-gated component suppliers as rack power rises
- /ai-power-chain-vendor-screen/
- /pricing-under-scarcity/
- /the-wide-bandgap-stack/
- /the-pe-playbook-layer-by-layer/
Reference surfaces
Topic hubs
- 800V DC data centre power
- AI data centre cooling
- AI power semiconductors
- AI data centre grid interconnection
- Hyperscaler + nuclear
- AI infrastructure regulation
- AI infrastructure financing
- AI infrastructure M&A
Comparison hubs
- 800V DC vs 48V data centre power
- SiC vs GaN in AI power semiconductors
- DLC vs immersion cooling
- Nuclear restart vs SMR vs traditional PPA
- NVIDIA 800V vs OCP Mount Diablo