Ahmadreza Samadi / Home

Technology and proptech, explained

What makes a building ready for AI data centers?

Cloud AI runs on real computers. An empty hall does not establish site readiness. This original learning framework asks what evidence is available before a specialist evaluates a facility.

By ·

A phone connects to a physical building with server racks. The diagram introduces the power, cooling, network and operations questions behind an AI data center.
Original pastel lavender carousel. Simplified explanatory diagrams.

The physical site matters

Power, heat removal, networks and operations shape whether a building can host the intended computing workload.

Documents come before a readiness claim

A dated power document or a cooling plan provides something to examine. It does not certify performance or readiness.

Specialists make the assessment

The fictional comparison below is a learning aid, not engineering signoff or investment advice.

Why is this a current proptech topic?

The public introduction to JLL's 7 October 2026 guide lists power, cooling, structural integrity and staffing as readiness criteria. Meta's 7 and 8 October explainers also put physical AI infrastructure in focus. This post uses JLL's public introduction, not its downloadable scorecard.

Source: JLL: Is your data center ready for AI? A critical infrastructure readiness scorecard · Meta: Debunking the Biggest Data Center Myths · Meta: Why Data Centers Are Such a Big Part of Meta’s AI Approach

Why is floor space only the starting point?

Servers, storage and network equipment occupy physical facilities. Electricity powers the equipment and cooling removes heat. A property description alone does not establish that those systems can support a particular workload.

Source: IBM: What is a data center?

What should the power evidence show?

Our original evidence questions are: what connection and capacity are documented, and what delivery date is supported? Ask who reviewed the information and which operating requirements it covers. A specialist still needs to evaluate the intended workload and site.

Source: JLL: Is your data center ready for AI? A critical infrastructure readiness scorecard · Meta: Debunking the Biggest Data Center Myths

Does a closed cooling loop mean zero site water use?

Meta describes a sealed chip-cooling loop that recirculates liquid while a heat-rejection system removes heat. That is a description of the stated design. It is not a claim that every operator uses it or proof of zero water use across an entire facility.

Source: Meta: Debunking the Biggest Data Center Myths

Where do networks and the building fit?

For a cloud service, data must move between users and computers. Ask about network routes, capacity and service requirements. Also ask about equipment space, structural capacity, maintenance access and the operating team. These are starting questions, not a complete design checklist.

Source: JLL: Is your data center ready for AI? A critical infrastructure readiness scorecard · IBM: What is a data center?

How do the fictional sites differ?

Fictional Site A has an unknown power connection date and an unclear cooling plan. Site B has a dated power document and a written cooling plan. Those documents are still unreviewed; network and building requirements remain unspecified in this example. Neither site is declared ready or certified.

What is the practical takeaway?

Before concluding that a site is suitable, gather evidence for power, cooling, connections and building fit. Then seek a qualified, site-specific assessment of the actual workload and requirements. The diagrams do not predict project outcomes, costs, savings or investment returns.

Source: JLL: Is your data center ready for AI? A critical infrastructure readiness scorecard · IBM: What is a data center?

See the explanation in slides

These original diagrams explain the mechanism. The carousel is in English.

View all 8 English slides
A phone connects to a physical building with server racks. The diagram introduces the power, cooling, network and operations questions behind an AI data center.
Slide 1 / 8
Two source cards identify the JLL facility-readiness guide dated 7 October and the Meta cooling explainer dated 8 October, beside a simplified server rack.
Slide 2 / 8
An electricity symbol points to a document with questions about capacity, connection date and operating needs. No capacity or delivery promise is supplied.
Slide 3 / 8
A qualitative closed liquid loop moves heat from a chip to a heat-rejection unit and returns cooler liquid. An outward heat arrow distinguishes heat release from coolant circulation.
Slide 4 / 8
A phone and a chip exchange data through a network box. Arrows show a simplified request and response path, not measured speed or a guaranteed network service.
Slide 5 / 8
Three cards show available space, structural support and people. A note links these to maintenance access and operations, without certifying a particular building.
Slide 6 / 8
Two fictional hall cards compare Site A with unknown power and unclear cooling against Site B with a dated power document and a written cooling plan. Both still need further review.
Slide 7 / 8
Four cards summarize evidence questions for power, cooling, networks and building fit. The closing note makes clear that the questions are for learning, not engineering signoff.
Slide 8 / 8

Download the English carousel PDF

Sources

Primary sources checked on 2026-10-12. Dated announcements and operator reports are attributed in the explanation; availability and terms can change.