Planning an AI/GPU-ready data hall
AI has broken the assumptions data halls were built on. A rack that used to draw 5–8kW now draws 40, 80, even 130kW; air cooling runs out long before that; and the network between the GPUs matters more than the one to the outside world. This is a technical guide to designing for it — power density, the cooling ladder, the 800GbE fabric, and the physical design that holds it all up — with an interactive density & cooling calculator.
Why AI Breaks the Data Hall
For twenty years, a data hall was a power and cooling problem solved at around 5–10kW per rack. AI training rewrote that overnight. GPUs are power-dense by design — thousands of them, packed tight, running flat-out for weeks — and they need to talk to each other at enormous bandwidth. The result is a facility where power density, liquid cooling and the internal network are the design, and everything else follows. A hall built for yesterday’s densities can’t simply be “upgraded” for AI; the power, cooling and weight assumptions are different from the floor up.
enterprise
/ HCI
/ high-density
cluster
rack
Per-rack power, by workload (indicative). The jump from virtualisation to AI training isn’t incremental — it’s a step change of 5–10× that takes you straight past what air can carry and into liquid cooling, reinforced floors and a fundamentally different power and network design.
The Density Problem — and What It Demands
Everything starts with the power budget per rack, because it sets the cooling, the electrical distribution and even the structural design. Drag the density below to see what a rack at that power actually needs — the cooling that can carry it, the heat it rejects, and what it costs to run.
Set the power per rack. The cooling methods light up where they can carry that load — and the readouts show heat and annual energy.
The Cooling Ladder: Air to Immersion
As density climbs, cooling moves from blowing cold air to bringing liquid ever closer to the silicon — because water carries roughly 3,500× the heat of the same volume of air. Each rung handles more heat, at more cost and complexity. Most AI halls today land on direct-to-chip liquid, often hybrid with air for the parts that don’t need it.
| Method | Density ceiling | How it works | Trade-off |
|---|---|---|---|
| Air + containment | ~30 kW | Hot/cold aisle containment with CRAC/CRAH | Cheap and familiar — but runs out fast for AI |
| Rear-door heat exchanger | ~70 kW | Water-cooled coil on the back of the rack | Big step up, keeps the air model — needs water to the rack |
| Direct-to-chip (DLC) | ~150 kW | Cold plates on the GPUs/CPUs, liquid loop | The AI workhorse — plumbing, CDUs & a leak-management regime |
| Immersion | ~200 kW+ | Servers submerged in dielectric fluid | Highest density & efficiency — biggest operational change |
Ceilings are indicative and depend on coolant temperatures, rack layout and the specific kit. Two numbers matter alongside density: PUE (liquid pushes it toward ~1.1, against ~1.5 for air) and water — liquid cooling needs a facility water loop, CDUs (coolant distribution units) and a maintenance discipline that air never required. Designing the white space, the loop and the containment together is the difference between a hall that works and one that throttles.
The 800G Network That Feeds the GPUs
In a traditional hall, most traffic is north-south — in and out of the building. In an AI cluster it’s overwhelmingly east-west: GPUs exchanging gradients with each other, constantly, at line rate. A slow or congested fabric leaves thousands of pounds of GPU sitting idle, so the network is built as a non-blocking, low-latency leaf-spine — today on 400/800GbE, heading to 1.6T.
East-west is the workload
GPU-to-GPU traffic dominates. The fabric has to be non-blocking — full bandwidth between any two endpoints — or expensive accelerators stall waiting on the network.
Rail-optimised topology
AI fabrics are often wired so each GPU’s NIC connects to its own “rail” of the network, spreading collective traffic evenly and keeping latency predictable across the cluster.
Low loss is non-negotiable
At 800G the optical loss budget is tiny. Ultra-low-loss connectors, clean MPO trunks and certified links are what let the fabric actually run at rate — the cabling guide goes deep on why.
Pre-terminated & dense
Base-8/16 MPO trunks, pre-tested and ready to plug, handle the sheer fibre count of an AI fabric — faster to deploy and far cleaner to certify at scale.
What Holds It All Up
Density doesn’t just stress power and cooling — it changes the room. A fully-populated AI rack can weigh as much as a small car, draw three-phase power through busway, and demand containment and cable pathways planned around the liquid loop. The physical design is where an AI hall is won or lost.
Power distribution
High-density rPDUs fed from overhead busway, three-phase at the rack, and a power chain sized for sustained near-100% load — AI runs flat-out, not bursty.
Structural & weight
A loaded liquid-cooled rack can exceed 1,500kg. Floor loading, plinth design and access routes have to be engineered for it — you can’t retrofit a slab.
The liquid loop
CDUs, manifolds, leak detection and a facility water loop, designed alongside the white space — not bolted on. Serviceability and redundancy of the loop become uptime-critical.
Containment & airflow
Even liquid-cooled halls have an air component. Hot/cold aisle containment, blanking and airflow management still matter for the heat that stays in air.
Structured cabling at scale
Ultra-low-loss OM4/OS2 and pre-terminated MPO trunks, designed for the 800G fabric and the breakouts — with the counts taken straight off the design model.
White-space planning
Rack layout, pathways, the loop and the fabric planned as one. Get the white-space design right and the hall scales; get it wrong and you strand power, cooling or ports.
Builds on the rest of the series
An AI hall is where the whole lifecycle converges: the 800G fabric rests on the fundamentals in our structured cabling guide, it has to be run to the disciplines in data centre operations, and the facility itself is chosen against the criteria in how to choose a data centre.
How Optronix Helps
AI/GPU halls are exactly what we’re built for — high-density, 800G, liquid-ready, designed and delivered end to end. We design, supply, install, certify and migrate across the full data-centre lifecycle, for hyperscale, colocation and enterprise clients across the UK, EMEA and worldwide. Power distribution, containment, the liquid loop and white-space fit-out are our day job, not a sideline.
Model-derived design, fast
We take a project from a single client drawing to a costed, model-derived design — densities, cooling, fabric and cabling counts — in under 48 hours. The counts come straight off the model, so the design you sign off is the one we build.
Certified & warranty-backed cabling
The 800G fabric is installed as ultra-low-loss, pre-terminated, fully tested and certified plant — backed by a 25-year manufacturer system warranty through our accredited installer status with the major cabling manufacturers.
Design & model
Density, power, cooling strategy, fabric and cabling counts — modelled and costed from your drawing, fast, so decisions are made on real numbers.
Supply & install
Power distribution, containment, liquid-ready white space and the 800G cabling plant — supplied and installed by our own engineers, not subcontracted.
Test & certify
Every link tested and certified to standard, documented, and warranty-backed — so the fabric runs at rate from day one.
Migrate & operate
Staged migration of live estates and ongoing operational support — smart hands, moves and changes — across the lifecycle.
Planning an AI or high-density build?
Send us the drawing and the target densities, and we’ll come back with a costed, model-derived design — power, cooling, 800G fabric and cabling — built to scale.
Sources & notes
- Rack densities and accelerator/system power from public GPU and AI-system vendor specifications (e.g. 8-GPU servers ~10kW; GB200 NVL72-class racks ~120–132kW)
- ASHRAE TC 9.9 thermal guidelines and common industry practice for air, rear-door, direct-to-chip and immersion cooling density ceilings and PUE
- IEEE 802.3 (400/800GbE) and leaf-spine / rail-optimised fabric design as used in AI clusters; cabling per TIA-942 / ISO 11801 (see our structured cabling guide)
- Uptime Institute on facility power, water and operations for high-density compute
General guidance, not a design specification. All densities, cooling ceilings, PUE and energy figures are indicative and depend on the specific kit, coolant temperatures, layout and site; calculator figures are rounded for clarity and energy uses representative PUE and a £0.20/kWh illustration. Confirm against vendor specs and an engineered design.