Research note · method

Utilization is not one number.

The AI infrastructure buildout can simultaneously look fully committed and poorly utilized without either observation being false. That is because “utilization” is being used for several different denominators.

Published 5 September 2026 Draft for review
Executive finding

Five measurements, one word.

North American data-centre markets are commercially tight. CBRE reports primary-market supply up 33.7% year over year in H1 2026 while vacancy fell to 1.4%, with 80.4% of capacity under construction preleased. JLL reports 1% vacancy and 95% pre-commitment across a broader universe.

At the same time, Cast AI reports average GPU utilization of 5% across the tens of thousands of Kubernetes clusters it sampled through April 2026. And US grid planners are confronting more than 700 GW of data-centre power requests, some of which regulators now treat as speculative.

These numbers do not describe the same thing. They measure different stages between an intention to build and economically productive compute.

The conclusion

The existing utilization threshold should remain unresolved rather than be force-graded. It should be replaced prospectively with a multi-stage stack that separates planning demand, contracted capacity, delivered capacity, compute occupancy, active workload and economic clearing.

That change makes the instrument stricter, not easier to satisfy.

The missing chain

Six stages between a request and a return.

Public discussion jumps from “data-centre demand” to “AI utilization.” The analytical chain is longer, and an infrastructure cycle usually fails at a conversion between stages rather than at the announcement stage.

01
Planning demandHow much capacity is being requested?
Utility interconnection requests, announced campuses, prospective megawatts. The noisiest layer: it includes optionality and projects that may never be financed or built.
Medium, noisy
02
Contracted capacityHow much future capacity has a financially identified customer committed to take?
Preleased megawatts, signed cloud capacity agreements, take-or-pay contracts. Stronger evidence of demand, but not evidence that infrastructure exists.
Medium-high
03
Delivered capacityHow much contracted capacity has been commissioned and can carry load?
Energized megawatts, delivered halls, functioning cooling and network, live grid connection. This separates construction ambition from physical delivery.
Low-medium
04
Compute occupancyHow much available accelerator capacity is provisioned to customers or workloads?
Allocated accelerator-hours over available accelerator-hours. A commissioned facility can still contain unallocated compute.
Low
05
Active workloadHow intensively is provisioned compute actually being used?
Accelerator busy-time, memory and computation occupancy, serving and training duty cycle. This is the layer the Kubernetes telemetry measures for its own sample.
Low outside selected fleets
06
Economic clearingIs the active infrastructure earning enough to justify its installed cost and cost of capital?
Revenue or gross profit per accelerator-hour or megawatt, contract yield against annualized installed cost, return on invested capital against cost of capital. An asset can be technically utilized and still fail to clear.
Low-medium, company-specific

Public-data quality is assessed as of September 2026 and is expected to improve at stages 3 and 4 before stages 5 and 6. For each layer FP1 publishes the numerator, denominator, geography or fleet scope, observation date, primary source, source independence, known exclusions and confidence.

Measurement rule

Do not aggregate.

The six layers are not summed into a single utilization score.
  • Not until at least three successive sweeps establish stable, comparable public series.
  • Collapsing them creates more apparent precision than the evidence supports.
  • Using CBRE vacancy would redefine utilization as leased physical capacity. Using grid requests would redefine it as planning demand. Using a Kubernetes sample would redefine it as workload activity in one fleet. None is equivalent to the R-012 concept.
  • A longitudinal instrument must not silently switch among those constructs because one happens to be available in September.

Falsifier for this framework. The Utilization Stack should be abandoned or collapsed if a reliable public series emerges that directly and consistently measures delivered accelerator capacity, active workload and economic yield across the relevant hyperscaler and neocloud universe with a stable denominator.

Consequence for the Radar: no new crossing is recorded for the existing utilization dimension until the measurement construction is resolved. The old reading remains part of the record; the revised method begins prospectively. See the September evidence update.