Research, analysis, and perspectives on the built world.
A developer-focused guide to data center cooling systems in the AI era. Covers air cooling, containment, rear-door heat exchangers, direct-to-chip liquid cooling, immersion cooling and how cooling strategy affects site selection, water risk and phasing.
May 10, 2026 ·5 min read
This post explains what powered shell data centers are, why the model is gaining traction and what developers need to diligence before pursuing one. It covers the boundary between base building and tenant fit-out, power, cooling pathways, leasing structure, capital risk and AI-assisted delivery workflows.
May 9, 2026 ·5 min read
GPU data centers require different development assumptions than conventional colocation facilities. This post covers site criteria, power density, cooling requirements and where AI helps development teams underwrite the asset class.
May 8, 2026 ·5 min read
This post defines AI factory data centers as a distinct development format shaped by high-density compute, power deliverability, cooling complexity and network requirements. It explains the site criteria and underwriting implications for institutional developers.
May 7, 2026 ·5 min read
This post explains how data center campus development differs from single-building delivery. It covers site criteria, power strategy, phasing, fiber, water, permitting and where AI can compress diligence without replacing engineering judgment.
May 6, 2026 ·5 min read
N+1, 2N, and the Uptime Institute tier framework govern data center uptime guarantees, hyperscale tenant specifications, and a significant portion of MEP construction cost. This guide explains how each redundancy level works, how it maps to facility tiers, the capex implications for developers, and how to match architecture to tenant profile and return targets.
May 3, 2026 ·4 min read
The Uptime Institute's four-tier classification system defines data center reliability requirements with direct implications for site requirements, MEP capex, and tenant targeting. This guide breaks down what each tier demands in practice and how developers should use tier specification as a development decision variable.
May 2, 2026 ·5 min read
Prefabricated and modular data center construction compresses delivery timelines by 30-50% versus traditional stick-built development, but imposes real site constraints that standard feasibility often underweights. Covers the modular construction spectrum, speed and cost tradeoffs, site access requirements, and the scenarios where modular makes the strongest development case.
May 1, 2026 ·4 min read
Edge data centers are a distinct product type from hyperscale, driven by AI inference demand, 5G densification, and enterprise private cloud growth. This post covers site criteria, tenant profiles, underwriting distinctions, and where AI compresses the development workflow for metro and regional edge facilities.
April 29, 2026 ·5 min read