HUMANITYVECTOR
Menu

Why it matters, technology

AI compute

Installed capacity to train and run machine-learning models: accelerators times their throughput, housed in data centers.

Depends on
2
Depended on by
1
Metrics
0

Neighbourhood: what depends on it (left), what it depends on (right)

Dependencies around AI computeeconomic-output affects ai-compute: Candidateai-compute requires input ai-accelerator: Established relationai-compute requires input data-center: Established relationEconomic outputAI computeAI acceleratorData center
  • Candidate
  • Hypothesized
  • Mechanism supported
  • Causally supported
  • Established relation

Systems that would feel a shortage

Downstream: what depends on it

weakest edge Candidate
  1. AI compute affects Economic output0 evidenceCandidate

    Economic output depends on AI compute.

    Mechanism

    Proposed: productivity effects of AI compute on output. No mechanism reviewed and no evidence attached.

    Evidence

    No evidence attached.

    Status
    candidate
    Reviewer confidence
    0.30
    Stated strength
    not stated
    Lag
    not stated
    Version
    curated_seed_v1
    Reviewed
    2026-09-07

    Confidence and strength are curated judgements recorded with the edge, not measured quantities.

    evidence and analysis for this edge

Inputs it cannot do without

Upstream: what it depends on

weakest edge Established relation
  1. AI compute requires input AI accelerator1 evidenceEstablished relation

    AI compute depends on AI accelerator.

    Mechanism

    Training and serving large models runs on accelerators. Compute capacity is the number of accelerators installed times their throughput.

    Evidence

    Status
    established_identity_or_engineering_relation
    Reviewer confidence
    0.99
    Stated strength
    1.00
    Lag
    not stated
    Version
    curated_seed_v1
    Reviewed
    2026-09-07

    Confidence and strength are curated judgements recorded with the edge, not measured quantities.

    evidence and analysis for this edge

  2. AI compute requires input Data center2 evidenceEstablished relation

    AI compute depends on Data center.

    Mechanism

    Accelerators run in racks that need power, cooling and network at data-center scale.

    Evidence

    Status
    established_identity_or_engineering_relation
    Reviewer confidence
    0.99
    Stated strength
    1.00
    Lag
    not stated
    Version
    curated_seed_v1
    Reviewed
    2026-09-07

    Confidence and strength are curated judgements recorded with the edge, not measured quantities.

    evidence and analysis for this edge

Measured, from the catalog

Linked metrics

Current value for the metric's default geography. Values come from the metric pages, not from the graph.

No catalog metric is linked to this entity yet.

Named on the edges that depend on it

Substitutes

No substitute is named for this entity on any stored edge.

Derived from curated edges

Bottleneck centrality

centrality_v1Derived
Score
0.75
Downstream entities
2
Direct dependents
1

Sum over every entity downstream (within six hops) of the strongest path weight to it, where each edge weighs its stated strength divided by one plus its named substitutes. A graph statistic over curated edges, not an observation. Computed 2026-09-07 22:42 UTC.

Curated paths this entity is part of

Chains

Names that resolve to this node

Aliases

  • AI compute demand
  • AI computing capacity
  • AI training compute
  • accelerated computing

slug ai-compute, updated 2026-09-07 22:42 UTC