METAXIS Intelligence Workbench (Operator)

AI_ENERGY_INFRA -- Report report_intel_87210a61730c22b9_v3

version=3 readiness=CONDITIONALLY_READY generated_at=2026-10-01T11:50:11.491350+00:00

Intelligence Object: intel_87210a61730c22b9 (IO version 6, readiness CONDITIONALLY_READY)

Provenance Inspector for this report →

Current State

Hypotheses (canonical status)

Claims (Key Claims) AVAILABLE

• {"type": "INTERPRETATION", "text": "현재 확인되지 않았다.", "claim_id": "claim_1d9458f4e7e2cf69", "claim_status": "OPEN", "claim_type": "PREDICTIVE"}
• {"type": "FACT", "text": "현재 확인되지 않았다.", "claim_id": "claim_5787125ab5f0110b", "claim_status": "OPEN", "claim_type": "DESCRIPTIVE"}
• {"type": "FACT", "text": "현재 확인되지 않았다.", "claim_id": "claim_7f1bc452c4ffc128", "claim_status": "OPEN", "claim_type": "DESCRIPTIVE"}
• {"type": "FACT", "text": "현재 확인되지 않았다.", "claim_id": "claim_afe7cc7b19317ee0", "claim_status": "OPEN", "claim_type": "DESCRIPTIVE"}
• {"type": "INTERPRETATION", "text": "현재 확인되지 않았다.", "claim_id": "claim_d170f52f567053e9", "claim_status": "OPEN", "claim_type": "TREND"}
• {"type": "FACT", "text": "현재 확인되지 않았다.", "claim_id": "claim_o1_lbnl_us_dc_2023", "claim_status": "OPEN", "claim_type": "DESCRIPTIVE"}
• {"type": "FACT", "text": "현재 확인되지 않았다.", "claim_id": "claim_o1_korea_capital_region_dc_demand", "claim_status": "OPEN", "claim_type": "DESCRIPTIVE"}
• {"type": "FACT", "text": "현재 확인되지 않았다.", "claim_id": "claim_3ab07077d739a9a4", "claim_status": "OPEN", "claim_type": "DESCRIPTIVE"}
• {"type": "FACT", "text": "현재 확인되지 않았다.", "claim_id": "claim_931b08c7ebfc9604", "claim_status": "OPEN", "claim_type": "DESCRIPTIVE"}
• {"type": "FACT", "text": "현재 확인되지 않았다.", "claim_id": "claim_afe484e45964c605", "claim_status": "OPEN", "claim_type": "DESCRIPTIVE"}

근거 / Research Evidence

empty / NOT_AVAILABLE

Statistics

Observation

• {"type": "EVIDENCE_SUMMARY", "indicator": "IEA_GLOBAL_DC_ELECTRICITY_TWH_FORECAST", "source": "IEA_ENERGY_AND_AI_EXEC_SUMMARY", "geography": "GLOBAL", "period": "2030-2035", "observation_count": 2, "trend_status": "INSUFFICIENT_SERIES", "limitations": []}
• {"type": "EVIDENCE_SUMMARY", "indicator": "EG.USE.ELEC.KH.PC", "source": "src_worldbank_api", "geography": "USA", "period": "2010-2022", "observation_count": 13, "trend_status": "NO_CLEAR_SIGNAL", "limitations": []}
• {"type": "EVIDENCE_SUMMARY", "indicator": "IEA_GLOBAL_DC_ELECTRICITY_TWH_OBSERVED", "source": "IEA_ENERGY_AND_AI_EXEC_SUMMARY", "geography": "GLOBAL", "period": "2024-2024", "observation_count": 1, "trend_status": "INSUFFICIENT_SERIES", "limitations": []}

Forecast

none

반증 자료 / Counterevidence AVAILABLE

• {"direction": "manufacturing/reshoring", "status": "EVIDENCE_FOUND", "source": "Grid Strategies LLC reports"}
• {"direction": "weather/climate peak-load", "status": "EVIDENCE_FOUND", "source": "EIA/Ember peak-load spikes, ERCOT/Eastern Interconnection", "note": "flagged as a PEAK-LOAD timescale finding, distinct from annual TWh/yr trend claims -- not merged with them"}
• {"direction": "DC growth without power problems (region search)", "status": "SEARCH_RUN_NO_CLEAR_EVIDENCE", "note": "honest negative -- search was run, no clear supporting evidence found either way"}

Alternative Explanations AVAILABLE

• GENERAL_CLOUD_COMPUTING_GROWTH

Geographic Applicability AVAILABLE

• {"geography": "USA", "basis": "statistic geography field (World Bank country code), not inferred from publisher or language", "scope": "event_country=UNKNOWN (per M.5E-F geography model discipline)"}
• {"geography": "KR", "basis": "source_registry_verified_publisher (connected event's documents)", "scope": "publisher_country only for event evt_a6a4027d9878143d; event_country stays UNKNOWN (no inference from publisher to event location)"}
• {"geography": "KR", "basis": "Korea capital-region grid-connection bottleneck claim (secondary reporting of National Assembly audit material, via Rep. 이소영; KEPCO 9th T&D plan)", "scope": "DERIVATIVE_REPORTING/NEWS-SECONDARY, not promoted to POLICY_RESEARCH -- primary 국정감사 document not independently fetched"}
• {"geography": "IE", "basis": "Ireland CSO official statistics (national statistical office, PRIMARY_SOURCE)", "scope": "data centers = 22% of Ireland's metered electricity 2024, up from 5% in 2015; AI_ATTRIBUTION=INDIRECT"}
• {"geography": "CN", "basis": "IEA vs Goldman Sachs/Bloomberg estimates via Carbon Brief / Oxford Institute for Energy Studies (DERIVATIVE_REPORTING/ACADEMIC_ANALYSIS)", "scope": "~100 TWh (IEA) vs ~200 TWh (Goldman/Bloomberg) 2024, both kept side by side; East Data West Computing (东数西算) program described, not independently verified as primary source"}

Temporal Applicability AVAILABLE

• {"period_start": "2010", "period_end": "2022", "source": "series_af37f1207e7ff042"}
• {"period_start": "2026-09-29T10:00:04+09:00", "period_end": "2026-09-29T10:00:04+09:00", "source": "evt_a6a4027d9878143d", "note": "1 real event, 2 documents reporting it -- document count != event count"}

Known Gaps AVAILABLE

• {"gap_type": "ATTRIBUTION_GAP", "reason": "partially addressed: IEA/GridLab evidence now connects AI/data-center growth to electricity demand at PARTIAL attribution (named as one driver among several -- general digital services, electrification -- with no numeric AI-only split given by any source); the earlier World Bank per-capita series still cannot be decomposed by end-use sector."}
• {"gap_type": "TEMPORAL_GAP", "reason": "series ends 2022, before the 2023+ AI data-center buildout commonly cited in AI_ENERGY_INFRA reporting -- cannot speak to the most recent period"}
• {"gap_type": "COUNTEREVIDENCE_GAP", "reason": "partially addressed: live counterevidence acquisition (o0_live_counterevidence_supplement.json) found real evidence for efficiency-improvements and non-AI-cloud-growth directions via the Jevons paper; manufacturing/reshoring, EV/electrification, population growth, and weather/climate directions remain SEARCH_NOT_RUN or evidence-thin."}
• {"gap_type": "POLICY_RESEARCH_GAP", "reason": "carried over from M.5E-F -- still NOT_READY, unresolved"}
• {"gap_type": "HYPOTHESIS_MODEL_QUALITY_WEIGHTING_GAP", "reason": "intel/hypothesis/hypothesis_model.py classify_status() is purely evidence-COUNT-based with zero quality/independence weighting (if supporting_evidence and not contradicting_evidence: return SUPPORTED). As of this update it marks H1 (hyp_b0a4bbbc9b601728, 'AI data centers are a significant driver of electricity demand growth') and H3 (hyp_o0_h3_ai_energy_infra, 'recent demand growth results from multiple non-AI factors') BOTH as SUPPORTED simultaneously, even though they represent competing explanations -- because each merely has >=1 supporting claim and 0 directly-contradicting claims in the model's narrow count-based sense. This is a disclosed structural limitation of the current hypothesis model, not a data error and not silently corrected here: classify_status() was intentionally left unmodified this phase per explicit instruction, since rewriting its logic would be an unscoped architecture change. Any report or synthesis reading H1/H3 status must present both as SUPPORTED-under-this-model while explicitly flagging that the model cannot currently adjudicate between competing hypotheses by evidence quality, independence, or directness of contradiction."}

Sources (출처) -- 4 actually used

Reports (version history)

Version Diffs (via report_engine.diff_reports())

Provenance

Full Report → IO → Hypothesis → Claim → Evidence → Source chain →