An edge states only its declared type. A line never silently means causality.
OBSERVATORY / SYSTEM RELATIONSHIPS
Transition Map
System relationships · linked evidence · explicit unknowns
Relationship definitions
Confidence is semantic and method-bound; line width never encodes synthetic precision.
VARDARK KNEW BY uses the M4 knowledge cursor and never backdates a relationship before review.
Node signals retain their owning module unit and are never normalized across the map.
System, constraints and gaps
Node signals preserve their module units. Edge width carries no analytical meaning. Select a node to trace its immediate system context or select a relationship to inspect its evidence.
The accepted Compute × Energy method derives nominal power and theoretical compute-per-MW from the published ownership series.
- Nameplate board power is not measured wall-plug consumption.
- The result is a hardware proxy, not useful work per joule.
From 2022-Q1 to 2025-Q4, estimated global compute holdings increased approximately 333.8 times, nominal accelerator power 168.7 times and theoretical compute per nominal MW 1.98 times.
epoch-ai-chip-owners-2026-08-10-570f980dc551 · epoch-ai-h100e-methodology-2026-08-112025-12-31 · KNOWN 2026-08-17T23:23:00.134Zcompute-energy/claim/nameplate-energy-envelopeThe world series' nominal accelerator power in 2025-Q4 is equivalent to 117.0 TWh per year if nameplate power is sustained continuously at PUE 1.0.
epoch-ai-chip-owners-2026-08-10-570f980dc5512025-12-31 · KNOWN 2026-08-17T23:23:00.134Zcompute-energy/claim/model-boundaryCompute per nominal watt is a theoretical hardware proxy. Useful AI work per kWh requires workload, precision, system and wall-power measurements that this module does not yet contain.
epoch-ai-chip-owners-2026-08-10-570f980dc551 · epoch-ai-h100e-methodology-2026-08-11Unquantified relationships
Dashed relationships are not weak scores. They are explicit statements that the Observatory cannot yet quantify the bridge.