Public discussion often collapses Iran’s internet crisis into a single event. The monitoring timelines suggest a more complicated reality: two major shutdown phases, separated by partial restoration and degraded service.

### Shutdown #1: January’s protest blackout

Internet Society Pulse documents a near-total nationwide shutdown beginning January 8, 2026, amid protests. By January 28, service was partially and patchily restored. The recovery was not clean. Pulse notes that through mid-February, connectivity remained degraded and unreliable, roughly 50–60% of normal traffic in that period.

That partial restoration matters, because it shapes what can responsibly be called “continuous.” Some posts and counters combine the phases; careful reporting separates them.

### Shutdown #2: February 28 and the start of the wartime phase

The wartime shutdown is widely pegged to February 28, 2026, after U.S.–Israel strikes and the onset of open hostilities, as described in The Guardian’s reporting. Internet Society Pulse notes traffic dropping to near zero again on February 28, with Cloudflare Radar confirming the collapse.

From a reader’s perspective, the distinction is not academic. The protest shutdown and the wartime shutdown reflect different state calculations: one aimed at controlling internal unrest, the other at managing information in conflict. Treating them as identical blurs the “why,” not just the “when.”

When headlines say a country “went dark,” the phrase can sound metaphorical. The monitoring groups referenced in current coverage treat it as measurable: the sharp reduction of internet traffic and global reachability.

### NetBlocks: near-total loss at national scale

NetBlocks is repeatedly cited as an independent monitor of national connectivity. In reporting summarized by Al Jazeera, NetBlocks assessed that Iran’s connectivity hovered around ~1% of pre-war levels during the wartime shutdown. That figure offers both magnitude and context: not a localized outage, not a throttling episode, but something close to a switch-off.

NetBlocks’ claim—also echoed via AFP-distributed reporting—is that Iran’s blackout has become the longest nationwide shutdown on record by its metrics. Yet even NetBlocks, as quoted by Al Jazeera, acknowledges a crucial qualifier: some longer blackouts have occurred elsewhere, but not in a way that matches Iran’s duration at this scale.

### Cloudflare Radar: the cliff edge

Internet Society Pulse notes Cloudflare Radar confirming the February 28 crash: traffic dropping to nearly zero. Cloudflare’s value here is corroboration—an independent, widely used network observation source showing the same “cliff edge” pattern consistent with a coordinated national restriction.

### Kentik and Doug Madory: the mechanics behind the blackout

The Guardian’s reporting points readers toward analysis associated with Kentik, where internet analyst Doug Madory is frequently cited in wider shutdown coverage. The significance is not celebrity expertise; it’s method. These analysts look for the structural signs of deliberate shutdowns—patterns across networks, routing behavior, and the scope of loss—rather than relying on anecdotal reports.

Taken together, the monitoring ecosystem tells a coherent story: Iran’s connectivity did not merely degrade. It collapsed—and stayed collapsed.

Key statistics (with context):
- ~1% of pre-war connectivity during the wartime shutdown (NetBlocks via Al Jazeera).
- Feb 28, 2026: traffic drops to near zero (Cloudflare Radar, cited by Internet Society Pulse).
- 37th consecutive day / 864 hours by April 5, 2026 (NetBlocks, via AFP-distributed report).
- 50–60% of normal traffic in mid-February during the degraded recovery between shutdown phases (Internet Society Pulse).

The duration is the headline hook, but the details matter. The most commonly cited count begins at the start of the wartime shutdown.

### The start date: February 28, 2026

Multiple sources place the beginning of the current nationwide shutdown at February 28, 2026, linked to the escalation into open hostilities after U.S.–Israel strikes. Internet Society Pulse’s timeline aligns with that date, and Cloudflare Radar data cited there supports the abruptness.

### The count: 37 days, 38 days, “more than 38 days”

By April 5, 2026, an AFP-distributed item quotes NetBlocks saying Iran had entered the 37th consecutive day—864 hours—of blackout. By April 6, 2026, The Guardian described the outage as “more than 38 days.”

The one-day difference is not a contradiction. It is the normal friction of reporting cutoffs: one publication time-stamps to April 5, another to April 6, and both describe an ongoing event.

### The confusion: January’s shutdown bleeding into the wartime count

Internet Society Pulse also lists the earlier January shutdown and the partial restoration. Some readers will see a running “duration” counter and assume uninterrupted darkness since early January. The more defensible interpretation is:

- Jan 8 → Jan 28: near-total protest shutdown, then partial easing
- Mid-Feb: degraded recovery at 50–60%
- Feb 28 → early April: near-total wartime shutdown, now 38+ days

That separation doesn’t soften the story. It sharpens it. Iran has experienced not one extraordinary cutoff, but repeated ones—culminating in a prolonged wartime blackout.

The record claim is where responsible writing becomes necessary, not optional. NetBlocks is being cited for a superlative, but comparisons depend on definitions.

### NetBlocks’ “longest on record” claim—and its own qualifier

Al Jazeera reports NetBlocks describing Iran’s wartime switch-off as the longest nationwide shutdown on record, while also acknowledging that some countries/regions have experienced longer blackouts, but not at comparable national scale.

That’s a subtle but meaningful distinction. A months-long outage in part of a country, or in a smaller network environment, is not the same as a near-total cutoff across a large nation’s connectivity footprint.

### The Guardian’s framing: “since the Arab spring”

The Guardian uses a more bounded comparison: calling it the longest national-scale blackout since the Arab spring, rather than the longest in absolute global history.

The Guardian also points to historical comparators:
- Sudan (2019): a 37-day shutdown, which Iran has now surpassed in the current wartime phase.
- Libya (Arab Spring): reported to have lost connectivity for nearly six months, though context and comparability differ.

Those examples show why absolutist headlines are risky. If Libya’s disruption is treated as a national-scale shutdown comparable to Iran’s, it challenges “longest ever.” If Libya’s case is treated as different in structure, scope, or measurement, NetBlocks’ “at this scale” argument strengthens.

### What readers should take away

A fair synthesis, based on the reporting in hand, looks like this:

- Iran’s wartime shutdown is among the longest nation-scale blackouts documented.
- NetBlocks is cited as calling it the longest nationwide shutdown on record by its metrics, while still acknowledging longer outages exist in other contexts.
- Some outlets, notably The Guardian, prefer the more conservative claim: longest since the Arab spring.

The strategic logic of an internet shutdown is usually presented as security: disrupt coordination, slow the spread of sensitive information, and manage public narratives during crisis. The timeline here—February 28, linked to the onset of hostilities—fits that pattern.

Yet the social cost scales faster than most policymakers admit. When connectivity collapses to near zero, the blackout is not experienced as a single policy decision. It is experienced as a thousand daily failures.

### Information collapse: from journalism to rumor

A prolonged shutdown changes what counts as “news.” Without broad access, public knowledge becomes localized and uneven. Even when a state controls broadcasting, digital disconnection adds a different kind of control: it limits what can be verified externally and what can be shared internally.

### Economic paralysis: the quiet catastrophe

A nation-scale cutoff hits:
- Payments and commerce that rely on online verification
- Small businesses dependent on messaging, ordering, and logistics
- Remote work and cross-border contracting
- Diaspora connections that support families and medical decisions

Even readers outside Iran should recognize the pattern. Modern economies aren’t merely “online.” They are synchronized by the internet. Break the synchronization for weeks, and the losses become structural.

### Human relationships under strain

Prolonged disconnection is also intimate. Families lose touch. People abroad cannot reliably check on relatives. For those inside the country, the absence of contact becomes its own form of pressure—one that compounds wartime anxiety.

Iran’s blackout is not happening in a vacuum. Recent history provides case studies that show how shutdowns become normalized—and how duration turns a “temporary measure” into a political technology.

### Sudan 2019: 37 days as a benchmark

The Guardian cites Sudan’s 37-day shutdown in 2019 as a comparator that Iran has now exceeded in the wartime phase. The significance of Sudan in this story is that “weeks-long blackout” is no longer hypothetical. It is a proven tactic, with known consequences for civil society and economic activity.

### Iran January 2026: a rehearsal that became a prelude

TechCrunch reported in mid-January 2026 that Iran’s protest shutdown had already become one of its longest ever, with historical comparisons to shutdowns such as Sudan (mid-2021, ~35 days) and Mauritania (July 2024, ~22 days).

That earlier reporting matters because it shows momentum: a country that has already executed a major shutdown early in the year is more capable—technically and politically—of doing it again during war.

### The hard lesson of repetition

A single blackout can be described as exceptional. Multiple blackouts within months begin to look like infrastructure policy: an implicit claim that the state can, when it chooses, treat nationwide connectivity as discretionary.

For readers watching other governments flirt with shutdown powers, Iran’s case offers a warning: once the switch exists, using it becomes easier.

For audiences outside Iran, the question is often, “How do we know what’s true?” The answer lies in triangulation: cross-checking independent measurement sources and reading record claims with precision.

### How to read the next headline responsibly

When you see “longest ever,” ask:
- Does it specify nationwide vs regional?
- Does it clarify continuous vs intermittent/partial restoration?
- Is the claim attributed to a specific monitor (NetBlocks) and a specific metric (connectivity level)?

A careful version of the headline is not less urgent. It is more credible.

### What to monitor (even as a non-expert)

You don’t need to be a network engineer to follow the evidence. Look for:
- NetBlocks updates on national connectivity levels (often expressed relative to baseline)
- Cloudflare Radar traffic graphs showing sharp drops and recoveries
- Timeline aggregation from Internet Society Pulse, which logs major changes and sources

### Why the “1%” statistic matters

“~1% of pre-war levels” (NetBlocks via Al Jazeera) is not just a number for policy wonks. It signals that connectivity is not merely restricted; it is close to functionally absent. That difference shapes everything from humanitarian communication to basic commerce.

The longer the blackout runs, the more likely it is to produce second-order effects: business closures, broken supply chains, and the normalization of alternative information channels—often less reliable, sometimes more extreme.