JOHOR COMPUTE-GRID INSTRUMENT:

WEEK 1 COMPLETE · WEEK 2 READY

Active Mini-Research Investigation

At current approval rates, when does Johor's data centre pipeline cross Peninsular Malaysia's physical grid limit?

This monitor is a live analytical instrument tracing RM164.45B+ of approved digital infrastructure projects in Johor. By synthesizing world developmental baselines with local electricity grids and hourly carbon indices, we translate abstract sovereign AI policies into immediate constraint timelines.

Vector A

Compute Grid Headroom

Will Johor's projected 2029 pipeline (~5,000 MW) cross 25% of total Peninsular Malaysia grid capacity? We track and visualize the physical over-allocation index.

Vector B

The Real Grid Carbon Gap

What is the divergence between marketing paper-green claims (GET/CRESS) and the real hourly fossil-fuel intensity of Malaysia's physical energy mix?

Vector C

Capital Reshaping Metrics

How TNB's massive RM42.8B capex allocation (2025-2027) is driven exclusively by compute nodes, and whether it succeeds in preventing localized outages.

Active Pipeline

Empirical Pull Instructions & Data Pipeline Schema

Supports 14-compute-grid.html files

To power this instrument's structural predictions, three distinct layers of telemetry are pulled sequentially. Review the execution guide below:

1. Macroeconomic FDI

Pull: World Bank BoP current USD indicator BX.KLT.DINV.CD.WD to compare compute investments against structural GDP.

2. Grid Limits (MW)

Pull: Manual parse of ST (Suruhanjaya Tenaga) peak loads & TNB capex files. Feeds the grid_capacity.csv.

3. Hourly Carbon

Pull: Realtime hourly gCO2eq/kWh for zone MY via Electricity Maps API v3 or local open-source csv files.

EXECUTION PIPELINE IN THE CLIS:
Step A: Run python src/fetch_wb.py to pull FDI indicators.
Step B: Run python src/fetch_emaps.py to extract physical carbon intensity indices.
Step C: Run python src/chi.py to calculate Compute Headroom metrics and write outputs.

Telemetry Checklists

✔ World Bank FDI indicators (Live)
✔ Johor DCTF approvals (Curated)
⏳ Live Electricity Maps token
⚠ Southern Grid Segment Share

Telemetry synchronizer compliant with Substrate spec rules.

Johor DC capacity (2021)

10 MW

Fitch baseline

Johor DC capacity (2024)

1,400 MW

140× growth in 3 years

Fitch 2029 projection

5,000 MW

~19% of Peninsular MY grid

Approved investment (Q2 2025)

RM164.45B

MITI/MIDA baseline

Live Compute-Grid Monitor Sandbox

Simulate grid-level adjustments to calculate Johor's headroom in real-time.

Active Simulation

Johor Compute-Grid Headroom Tool · data as of July 2026 · Week 1 active

0.61

WARNING

Compute Headroom Index · Johor Southern Grid Allocation

0.0 — No Headroom (Deficit) 1.0 — Max Headroom

Southern Grid Segment
39% Available

Johor DC Pipeline (MW)

~2,600 MW

Source: DCTF Approvals

Peninsular installed capacity

~25,872 MW

Source: Suruhanjaya Tenaga

🔒 Carbon Intensity (gCO₂eq/kWh)
unlocks Week 2 or via live query agent

🔒 Pipeline vs Grid Chart
unlocks Week 3 · Chart.js

Interactive Parameter Tuning Console

DC Pipeline Demand (MW): 2,600 MW

Approved & committed data center load in Johor segment.

Peninsular Grid Capacity (MW): 25,872 MW

Total Peninsular Malaysia installed electricity generation capacity.

Southern Grid Share (WACC-equiv):

Assumed segment share of grid allocated to the Southern zone.

⚠ R sign-off req.

Existing Non-DC Grid Load (MW): 3,800 MW

Domestic, industrial, and commercial baseline loads in the South.

Active API Connect Hub

Connect directly to live public data nodes. Retrieve macroeconomic metrics from the World Bank API or trigger the grounded search engine of Gemini to scrape actual grid capacities.

1. World Bank API (CORS-enabled)

Target MYS (Malaysia) • Endpoint: api.worldbank.org/v2/country/mys/...

2. Electricity Maps API (Zone MY)

Zone MY • Endpoint: 127.0.0.1:5050/api/carbon (needs ELECTRICITY_MAPS_TOKEN)

2. Gemini Grounding Assistant Search Tool Active

Toggle Custom prompt Model: gemini-3-flash-preview

3. Local Python Pipeline (src/api.py) OFFLINE

Calls your actual fetch_wb.py + process.py over HTTP — same code run.py uses, no JS reimplementation. Sliders above feed the request params.

Start it once per session: python src/api.py (needs venv with flask + flask-cors, see requirements.txt)

Telemetry State: STANDBY

johor-grid-pipeline-stdout

[01:23:16] $ initialize --version=1.0.5 --env=production
[01:23:16] [OK] System pipeline compiled successfully.
[01:23:16] [WARN] SOUTHERN_GRID_SHARE set to 0.25 (verification required).
[01:23:16] $ awaiting telemetry instructions...

Target Region: Southern Johor State: Active

CSV Analyzer

CHI Trend — src/analyze.py

Reads data/external/grid_capacity.csv + data/external/dctf_approvals.csv and runs the same compute_headroom() formula against every curated year — the "CSV data analyzer" milestone project, applied to this dataset instead of a throwaway exercise. 2024/2025 rows are sourced (PETRA grid capacity, Energy Commission DC demand — see data/external/SOURCES.md); other years are still unsourced placeholders, marked as such in the Source column below.

Year	Pipeline MW	Available MW	CHI	Status	Source
Click "Load CHI History" (needs `python src/api.py` running).

Unified Empirical Source Registry (Week 1 Audit)

[01] World Bank FDI Net Inflows (Malaysia) API SYNC READY • GRADE A

[02] Suruhanjaya Tenaga Total Peninsular Grid Capacity ST.GOV.MY DIRECTIVE • GRADE A

[03] MIDA Digital Investment & Approved Projects (Q2 2025) CURATED • GRADE B

[04] Electricity Maps Zone MY Grid Carbon Intensity LOCK STATE • GRADIENT UNLOCKED ON EVENT

01 / CONCEPT

Why combine 14 and 15 — The unified physical question

Project 14 isolated

Tracks FDI approvals and MW commitments. It captures how much compute is promised, but remains incomplete without knowing what the physical grid can safely support.

Project 15 isolated

Tracks grid capacity, power allocations, and clean grid trends. An abstract calculation without a concrete case study load or localized investment tracking.

Combined, they produce one falsifiable claim: At current approval rates, when does Johor's data centre pipeline cross 50% / 75% / 100% of available southern grid capacity? That constraint date is the Compute Pause Button made empirical — a real number instead of a policy abstraction.

CORE EMPIRICAL FINDING: Johor's 2029 pipeline (~5,000 MW) is ~19% of Peninsular Malaysia's entire installed generation capacity (~25,872 MW) — concentrated in one state, on one grid segment, with TNB's RM42.8B capex expansion (2025–2027) explicitly triggered by data centre load. That number is 2× the previous capex period. The grid is being reshaped around compute demand with no public governance framework for compute concentration. That is the compute governance argument, made empirical.

02 / VISUALS

Design direction — Substrate system, instrument aesthetic

This is a Substrate product, not an Irori product. The aesthetic signals "we handle uncertainty rigorously" — high-precision instrument, not warm editorial. No dark backgrounds. No terracotta. No serif. Every number in JetBrains Mono.

PALETTE SPEC: #F5F5F3 — Page background
#FFFFFF — Card ground
#EEEEED — Ghost states
#C8C8C4 — Structural borders
#B03040 — Alert & Constraints

TYPOGRAPHY RULES:

JetBrains Mono — every number, metric, label, code, grade badge, coordinate readout.
Inter — prose, navigation, section headers.

03 / TELEMETRY

Data architecture — three layers, three APIs

Layer & Source	Extracted Indicators	Access Method	Data Quality
World Bank (BoP Indicators)	FDI Net Inflows for MYS comparison	api.worldbank.org/v2	GRADE A
Suruhanjaya Tenaga Annuals	TNB installed capabilities, peak demands	Manual Extraction (PDF)	GRADE A
Electricity Maps (Zone MY)	Hourly grid carbon intensities (gCO2eq/kWh)	api.electricitymap.org/v3	GRADE A
MIDA Press Releases	Approved regional investment milestones	Manual Curation	GRADE B

04 / FORMULA

The Compute Headroom Index — formula & parameters

CHI = 1 − (DC_pipeline_MW / grid_available_MW)

where:
DC_pipeline_MW = operational + under_construction + committed MW in Johor
grid_available_MW = (peninsular_installed_capacity × SOUTHERN_GRID_SHARE) − existing_non_dc_load

CHI = 1.0 → no data centre load, full headroom
CHI = 0.5 → 50% of southern grid spoken for by DCs
CHI ≤ 0.0 → pipeline exceeds available capacity → constraint crossed

Adjusted CHI = CHI × (1 − carbon_intensity / 1000)
Penalises dirty grids: "green sovereign AI" needs both headroom AND low-carbon supply.

05 / INTEGRITY

Pre-registration — committing hypotheses before calculations

H₀ (compute headroom):
Johor's approved DC pipeline does not exceed 25% of available southern grid capacity in any single year between 2024 and 2027.

H₁ (research):
The pipeline crosses the 25% threshold before end-2026, and 50% before end-2028, implying structural grid stress without additional generation capacity.

Support condition: CHI < 0.75 before 2027 based on confirmed DCTF approvals.

06 / STRUCTURE

Project scaffold — folder structure

johor-compute-grid/
├── data/
│ ├── raw/ # API responses
│ ├── processed/ # tidy CSVs
│ └── external/ # curated grid_capacity.csv
├── src/
│ ├── fetch_wb.py # World Bank FDI scripts
│ ├── fetch_emaps.py # Carbon intensity triggers
│ └── run.py # Pipeline coordinator
└── dashboard/
└── johor-cgm-v1.html [YOU ARE VIEWING THIS INDEX/HOME FILE]

08 / MILESTONES

5-week build plan — MVP first, active visibility

WEEK 1

Skeleton UI

Static headroom equations & WB Live Fetch

WEEK 2

Carbon Unlock

Realtime fetch integration of gCO2eq indices

WEEK 3

Chart Unlocked

Dynamic Chart.js rendering within sandbox

WEEK 4

Scenarios

Alternative simulation paths

WEEK 5

Pipeline Run

Reprod. orchestration pipeline