CASE 02 · NEPAL · SEPTEMBER 2025 · NARRATIVE TO SECURITY

They granted the demand.
It had already been replaced.

On 4 September 2025 the government of Nepal suspended twenty-six social media platforms. On 8 September nineteen protesters were killed. At twenty-five past eleven that night the government lifted the ban. The following day the parliament burned and the prime minister resigned. Every step of this was public. It was read with an instrument that reports a satisfied demand as a closed file.

CONCESSION → FIRE

SAME NIGHT TO NEXT DAY

DOMAINS CROSSED

4

LENS THAT CARRIED

D · HONOUR / STATUS

SECURITY5 / 5POLITICAL4 / 5NARRATIVE4 / 5CYBER3 / 5ECONOMIC2 / 5SUPPLY CHAIN2 / 5ENERGY2 / 5CLIMATE1 / 5STABILITY2 / 53.4REGIONAL STRESSINDEXON A SCALE OF 1 TO 5ELEVATEDELEVATEDSEVERE

THE VERIFIED RECORD

4 SEP26 PLATFORMS SUSPENDED, REGULATORY NONCOMPLIANCE CITED
8 SEP 09:00ASSEMBLY AT MAITIGHAR MANDALA, KATHMANDU
8 SEP 18:0019 DEAD, OVER 100 INJURED
8 SEP 20:00HOME MINISTER RESIGNS
8 SEP 23:25BAN LIFTED
9 SEPPARLIAMENT AND GOVERNMENT OFFICES ALIGHT, PM RESIGNS

Public record. Sources: Britannica, Harvard Chan School, CNN, Reuters.

02 — THE PROPAGATION

CLICK A DATE

TO STEP THROUGH THE CASE — OR USE THE ← → KEYS

SECURITY5 / 5POLITICAL4 / 5NARRATIVE4 / 5CYBER3 / 5ECONOMIC2 / 5SUPPLY CHAIN2 / 5ENERGY2 / 5CLIMATE1 / 5STABILITY2 / 53.4REGIONAL STRESSINDEXON A SCALE OF 1 TO 5ELEVATEDELEVATEDSEVERE

HOW TO READ THIS — Nine domains, each scored 1 to 5 on current stress. The faint web is all 36 connections the system measures continuously. The red line is the route this case actually travelled. Click a step in the timeline to move the ring forward in time.

THREE WEEKS BEFORESIGNAL INNARRATIVE 2/5 → 4/5

Before any ban, corruption content is circulating. It is not complaint. It is inventory — posts contrasting the privileges of political families against young people who cannot find work. The word being used is nepo kids.

L10REGISTER & NARRATIVE VELOCITY

Complaint has become enumeration. A documented precursor register. The grammar changed, not the volume, which is exactly why volume-based monitoring cannot see it.

ESCALATES
L07CLAN & PATRONAGE

This is not economic grievance. It targets patronage transmission: who received what, through whom, by what relation. Nepo kids is a technical description of the target.

ESCALATES
L11AUDIENCE COST & FACE LEDGER

Naming families raises the cost of climbdown on both sides. Neither party can now concede quietly.

CONFIDENCE ONLY

HOW TO READ THIS — 18 layers run on every signal, and each pushes the assessment up or down. Click any row to open that layer and see what it measures. The call is what is left once they stop disagreeing.

STRONGEST MEASURED COUPLINGS — RANKED

NARRATIVECYBER71%
POLITICALSECURITY69%
CYBERPOLITICAL63%

HOW TO READ THIS — Each bar is how hard one domain pushed the next, measured in this case. Solid bars have already fired at the step you are on; faint bars come later in the chain.

03 — THE DIVERGENCE

CONSENSUS READ:The grievance has been identified and removed. The ban is lifted, the channel is restored, the protest decays.

OMEGANOVA READ:The ban was never the grievance. It was the proof. And earlier that same evening it stopped being the largest item on the ledger.

LENS FAN · 8 SEP 23:25

THE SAME FACTSALIBERAL / INSTITUTIONALISTCONSENSUS0%BREALIST / STRUCTURALSILENTSILENTCCONSTRUCTIVISTSUPPORTING0%DHONOUR / STATUSCARRIED0%EPOLITICAL ECONOMYSUPPORTING0%FDOCTRINALSUPPORTING0%

LENS WEIGHTS A .08 B .02 C .25 D .30 E .20 F .15 → WEIGHTED 0%

HOW TO READ THIS — Six schools of analysis, same facts, same day. They do not agree. Click any lens to see its read, its weight and its blind spot.

WHAT THE CONSENSUS LENS COULD NOT SEE

The liberal lens tracks demands, and a demand that has been superseded still looks satisfied when you meet it. Organisers said publicly that this was never only about the ban. The consensus reading treated the ban as the variable anyway, because it was the only variable the instrument had a field for.

04 — THE CALL · ISSUED 8 SEP 23:25

Symbolic state target attacked within a day.

55%
FALSIFIER — Security forces re-establish control of the parliament approach with no further fatalities, and a second, more senior resignation lands before dawn.

Negotiated pause, leadership survives.

24%
FALSIFIER — A named coalition partner publicly re-commits and LII recovers above 0.55.

Broad collapse of state order.

21%
FALSIFIER — State control holds outside the capital through the following day.

ISSUED 8 SEP 23:25 · HASH-STAMPED · GRADED AFTER THE EVENT

QUANTIFIED EXPOSURE

TWO OVERLAND CORRIDORS — TRANSIT SUSPENSION RISK, NEXT TWO DAYS

ONE LISTED TELECOM — REGULATORY AND INFRASTRUCTURE EXPOSURE

THREE DIPLOMATIC DELEGATIONS — POSTURE REVIEW RECOMMENDED ON ISSUE

ONE INTERNATIONAL AIRPORT — DISRUPTION RISK 35%

SOVEREIGN SPREAD — CONVERSION THRESHOLD +180BP

05 — THE GRADE

The parliament and other government offices burned on 9 September. The prime minister resigned the same day. Path one hit at 55 per cent — direction, magnitude and timing correct within the window. The army imposed a nationwide curfew on 10 September. By 22 September the toll stood at 74 dead and more than 2,113 injured.

RECORDED MISS

Path three was over-weighted at 21 per cent. State order did not break beyond the capital — provincial administration held and the security apparatus did not fragment. The collapse-branch prior was carrying a template from a different class of state. Correction applied: reduced by seven points, weight redistributed to path two.

Second limit, stated: the threshold identified the target class. The system does not name individual targets and did not.

06 — THE SECOND AXIS

A country desk is the same failure
as a domain silo.

Everything above moves between domains inside one country. The signal that predicted this one moved between countries, through a shared narrative substrate, and nobody organised it.

BANGLADESHJUL 2024INDONESIAAUG 2025NEPALSEP 2025TIMOR-LESTESEP 2025PHILIPPINESSEP 2025MALDIVESSEP 2025MOROCCOSEP 2025MADAGASCARSEP 2025PERUOCT 2025COUNTRIES 19ELAPSED 14 MONTHS
JUL 24SEP 24NOV 24JAN 25MAR 25MAY 25JUL 25SEP 25OCT 25

BANGLADESH · JUL 2024

INDONESIA · AUG 2025

NEPAL · SEP 2025

TIMOR-LESTE · SEP 2025

PHILIPPINES · SEP 2025

MALDIVES · SEP 2025

MOROCCO · SEP 2025

MADAGASCAR · SEP 2025

PERU · OCT 2025

HOVER A CHIP TO SEE WHAT WAS REUSED.

TRANSMISSION MARKERS — SYMBOL, PLATFORM, TACTIC. NONE OF THESE WERE ORGANISED. ALL OF THEM WERE OBSERVABLE.

The signal that predicted Nepal was generated in Dhaka fourteen months earlier.

Nobody authored the transmission. Everybody could observe it.

120

SIGNIFICANT ANTI-GOVERNMENT PROTESTS IN 12 MONTHS

73

COUNTRIES

2017

LAST TIME THE COUNT WAS THIS HIGH

Carnegie Endowment, Global Protest Tracker. This is not a tail event. It is a base rate the market is still pricing as a series of surprises.

AXIS 1

Moves between domains, inside one country. Measured as coupling percentage. Missed because monitoring is domain-siloed.

AXIS 2

Moves between countries, through a shared narrative substrate. Measured as template adoption latency. Missed because analysis is country-siloed.