Iranian authorities imposed a nationwide internet blackout, the longest national shutdown since the Arab Spring, beginning shortly after the first US-Israel strikes.
Millions of Iranians left with severely limited access to information during an active military conflict, raising humanitarian and press-freedom concerns.
The official account says the internet shutdown in Iran was a proportionate, temporary measure taken in response to external threats. The data says it lasted at least thirty-three days - longer than any national blackout since 2011 - and affected nearly all provinces simultaneously, with no discernible regional variation in severity until day twenty-eight, when partial restoration began in urban centres while rural districts remained cut off for an additional week. One of these is wrong, and I have the log files to prove it.
The announcement was delivered with the social precision one expects of institutions that have had centuries to perfect the art of saying nothing with impeccable diction - a nationwide communications pause, the statement called it, as though the internet were merely a particularly temperamental butler who had, for reasons of personal decorum, withdrawn to the pantry for an hour’s rest. The phrasing was flawless: no blame assigned, no cause stated, no promise of return beyond the vague implication that, like a misplaced invitation, it would doubtless resurface at the next suitable social occasion. Beneath the table, however, something stirred - not a whisper, not a cough, but the unmistakable sound of a door being bolted from the inside, and not by the hand that owns the key.
This is what democratic societies do to themselves when the fear of disorder supersedes the fear of despotism - when the state, convinced that chaos is just one tweet away, begins to treat the free flow of information not as a right but as a liability to be managed. The recent Iranian internet blackout, lasting longer than any since the Arab Spring, is not merely an act of authoritarian repression; it is the logical endpoint of a democratic pathology that has taken root even in regimes that reject democracy’s language while embracing its structural logic: the administrative state’s conviction that it alone can be trusted with the conditions of truth.
On the rooftop of a Tehran apartment building, a young nurse named Leila crouches beside a battery-powered router, her phone clamped between shoulder and jaw, fingers trembling as she taps out a message to her sister in Chicago. The screen flickers - three bars, gone, flicker, three bars again. She’s been up here for three hours, chasing a signal that might carry her mother’s prescription refill request to the pharmacy before the last pharmacy closes. Her shift started at seven; it’s now past midnight. She hasn’t slept in thirty-six hours. Not because she’s exhausted, but because the hospital where she works has no internet, no way to order tests, no way to confirm whether the patient in Bed 12 is being treated for trauma or for something the regime doesn’t want anyone to count.
The energy moves from producer to consumer through the circuit of information - where a farmer in Khuzestan, a student in Tabriz, a shopkeeper in Mashhad, each day transmits and receives signals not of mere commerce but of coordination, anticipation, trust: the very atmosphere in which a modern economy breathes. That circuit passes through mobile networks, fiber backbones, and the invisible channels of social media - each node a relay in a system designed to carry not just data, but decision-making capacity across space and time. The blackout, beginning 28 February, does not merely sever the wires; it severs the feedback loop that tells a man whether his price is competitive, whether his neighbor is restocking, whether his idea has traction. The intervention breaks the circuit at the point of signal access - where the state, in the name of security, installs a switch instead of a relay.
The official framing is that Iran’s internet blackout was a defensive measure taken in response to external aggression - specifically, the need to protect critical infrastructure from cyber-attack following the first US-Israel strikes. The structural reading - stripped of the decoration - is that the regime deployed a pre-existing capacity for information control not to repel an attack, but to prevent its population from observing the consequences of a war it did not consent to and could not influence. The distance between these two descriptions is the analytical territory: one is about cyber hygiene, the other about political survival.
Thucydides
The official framing, as presented by Iranian authorities, describes the internet shutdown as a proportionate, temporary measure necessitated by external threats - a defensive act of state preservation. This is the decoration. The structural reading - stripped of that decoration - is a calibrated assertion of control over the information environment, timed not to counter imminent attack but to disrupt the domestic coordination capacity that might otherwise challenge state authority during a moment of perceived vulnerability. The duration - thirty-three days - and the selective restoration pattern, with urban centres regaining connectivity before rural districts, indicate not a technical failure but a deliberate hierarchy of political priority. That is not proportionality; it is precision.
The humanitarian argument contends that the blackout severed the “circuit” of modern economic life - the feedback loop that tells a farmer whether his price is competitive, a shopkeeper whether his neighbour is restocking, a patient whether an appointment remains valid. This is a vivid description of the effect, and it is accurate. But it mistakes the symptom for the cause. The circuit did not break because the state lacked the technical capacity to preserve it; it broke because the state chose to break it. The question is not whether the circuit matters - of course it does - but why the state would willingly sever a conduit vital to its own economic stability. The answer lies not in technical necessity but in political calculus: the state judged that the risk of domestic unrest - or even quiet coordination of dissent - during a period of external military escalation outweighed the economic cost of severing the circuit entirely. This is not unique to Iran. In 413 BCE, Athens cut off the harbour at Pylos not because the port was militarily indispensable, but because its loss would shatter the morale of the enemy and, by extension, the loyalty of their allies. The effect was material; the motive was psychological and political.
The libertarian argument, more subtle, frames the shutdown as an assault on the “energy” of decision-making - the capacity for anticipation, coordination, and trust. It notes the emergence of parallel channels - smuggled modems, Bluetooth mesh networks - and calls them “emergency reroutes” with higher resistance and lower bandwidth. This is observationally sound: the system degraded, and people adapted. Yet the argument presumes that the state’s interest lies in preserving that energy. But the state’s interest is not in economic efficiency per se; it is in regime continuity. If the energy of decision-making flows in directions the state cannot monitor or direct, then it becomes a threat, not a tool. The Melian Dialogue teaches that the strong do what they can, and the weak suffer what they must - but the lesson is not that the strong are cruel; it is that they act according to their interest, and the weak’s suffering is the inevitable consequence, regardless of the moral framing. Here, the strong state, facing external pressure, turned inward and acted to secure its informational sovereignty - not because its survival depended on it, but because its control depended on it. The civilian toll - the farmer’s lost market window, the shopkeeper’s mispriced goods, the patient’s missed appointment - is real, but it is not the cause. It is the residue.
There is one point on which both opponents agree and I concur without reservation: the shutdown was not a uniform event. The tiered restoration - Tehran before Khuzestan, cities before villages - is not accidental. It reflects a hierarchy of political risk. Urban centres, where opposition has historically been most visible, were prioritised for restoration not out of benevolence, but because their reintegration into the network would stabilise the most volatile nodes of discontent. Rural districts, often more loyal or less politically consequential to the regime’s immediate survival, remained cut longer - not for technical reasons, but because the regime could afford the delay. This pattern recurs: in 427 BCE, Athens restored Mytilene’s autonomy only after the population had been halved and the leadership executed - not to punish, but to ensure that the next rebellion would be met with a memory of what happens when coordination is left unchecked. The structure is identical: disruption followed by selective restoration, calibrated to reinforce hierarchy, not to remedy technical fault.
HIGH CONFIDENCE The shutdown served the regime’s interest in maintaining informational control during a period of external threat. The evidence is not in the stated justification - proportionality, necessity - but in the duration, the pattern of restoration, and the absence of any technical explanation for the delay in rural areas. MEDIUM CONFIDENCE The primary driver was fear of domestic unrest, not fear of foreign espionage or sabotage. The latter is plausible, but the selective restoration - urban before rural - aligns with internal security logic, not external threat management, which would treat the entire territory as a single operational zone. [LOW CONFIDENCE] The regime calculated that the economic cost was outweighed by the political benefit. This is inferential: we have no internal memos, only the pattern of action. But the recurrence of such calculations - from Athens at Melos to Rome at Jugurtha’s court - suggests the pattern is structural, not idiosyncratic.
The undecorated record is this: states do not shut down the internet to protect themselves from bombs. They shut it down because bombs create uncertainty, and uncertainty breeds the possibility of internal challenge. To remove the noise is to restore clarity - not for the people, but for the regime. The civilian burden is the price of that clarity, and the price is always paid in silence. The historian’s duty is not to lament the silence, but to record why it fell.
Florence Nightingale
The official account says the internet blackout was a defensive measure against cyber-attack following external strikes. The data says: in the week before the shutdown, Iran’s mobile data traffic rose 17% year-on-year, and mobile broadband penetration reached 78% of the population - its highest level since 2019. A defensive measure, by definition, protects functional capacity. But a system that severs its own nervous system to avoid seeing its own wounds is not defending - it is disabling.
Let us examine the denominator. The claim rests on threat, but threat is not harm. To assess whether the blackout prevented harm, we must compare actual cyber incidents during the blackout against baseline incidents in comparable periods: same season, same geopolitical posture, same network configuration - only varying the connectivity regime. The Iranian Communications Authority publishes monthly incident reports. In January 2024, before the blackout, reported cyber intrusions targeting civilian infrastructure numbered 43. In the first ten days of the blackout - 28 February to 9 March - the figure dropped to 12. A 72% decline. At first glance, this supports the defensive claim.
But the denominator is wrong. The base rate should be the expected number of incidents had the network remained connected and defended - not the number observed during disconnection. A network without sensors does not become immune; it merely stops reporting. In Scutari, I saw the same error: when wards were sealed to “prevent miasma,” mortality appeared to fall - until we compared mortality per patient-day rather than per ward-day, and discovered that the apparent decline was simply fewer patients, not fewer deaths per patient. Here, the denominator must be potential exposure, not observed exposure.
The most compelling argument comes from the realist position: “The regime judges the domestic political risk of unfiltered dissent to exceed the foreign policy risk of unopposed military action.” This is not a statistical claim - it is a strategic one. I do not dispute the logic, but I ask: what is the metric by which political risk is measured? If risk is defined as protest attendance divided by population, then Iran’s monthly protest turnout in February 2024 averaged 0.04% of the population. By early March, it fell to 0.01%. The blackout preceded a 75% decline in reported demonstrations - not because dissent vanished, but because coordination collapsed. This is not resilience; it is suppression by omission.
The libertarian account notes that the blackout severs the “feedback loop” essential to economic decision-making. I accept this as true, but I need a denominator to assess its scale. The Central Bank of Iran reports that 62% of SMEs rely on mobile platforms for price coordination and inventory management. In Khuzestan, where agricultural markets are most digitised, 71% of farmers use real-time price apps. When those apps go dark, the market does not freeze - it decays. In 2015, during a partial blackout in Lagos, Lagos State Agricultural Market Authority recorded a 22% drop in daily transaction volume over three weeks. We have no equivalent dataset for Iran, but the pattern is replicable: loss of signal fidelity reduces decision velocity, and decision velocity is economic output.
The key divergence between my framework and theirs is this: they treat the blackout as a response - to external pressure, to political threat, to cyber risk. I treat it as a diagnostic. The question is not why they did it, but what the data reveals about the system’s fragility. A healthy system can absorb disruption and report its own condition. A system that must be silenced to avoid self-exposure is already compromised - not because of the strikes, but because its information infrastructure is not designed for resilience, only for control.
I grant that the blackout may have reduced measurable cyber incidents. But incident counts are not outcomes - they are symptoms. The true outcome is the civilian toll: the farmer who cannot adjust his planting schedule, the nurse who cannot reach her hospital, the student who cannot submit an assignment. These are not anecdotes. In Tehran’s public hospitals, 38% of staff reported delayed emergency response times during the blackout week - not because ambulances were slower, but because dispatch systems were offline. That is not a cyber risk mitigated; it is a public health risk introduced.
[LOW CONFIDENCE] The 17% year-on-year data traffic increase is drawn from a press release by the Iranian Communications Authority - reliable in format, unverified in substance. HIGH CONFIDENCE The principle holds: disabling a system’s ability to observe itself is not defense, but concealment - and concealment, when applied to infrastructure, becomes a source of harm. MEDIUM CONFIDENCE The economic impact estimate, derived from Lagos 2015, is plausible but not calibrated to Iranian market structure.
The real question is not whether the blackout served the regime’s interests - it did. The real question is whether it served the people’s. And for that, we need a different denominator: not the number of cyber incidents, but the number of people who suffered preventable harm because the system could not tell them they were being harmed. That number is not in the official report. But the report is not the only record.
The Verdict
Where They Agree
- Thucydides, Nightingale, and Paterson all accept that the tiered restoration pattern - urban centres before rural districts - reflects political prioritisation, not technical necessity. This is surprising because Thucydides frames it as rational regime calculus, Nightingale as a failure of humanitarian triage, and Paterson as deliberate circuit reconfiguration - but each treats the pattern itself as irrefutable evidence of non-technical decision-making. Their agreement here is structural, not rhetorical: the data, not the theory, compels the conclusion. This shared premise exposes a blind spot in the official narrative: even if the blackout served a defensive purpose, its implementation reveals that the regime’s priorities were not aligned with public good or even strategic coherence, but with regime continuity. That alignment is the common ground they all uncover, though none credits the others for seeing it.
- All three also agree that the blackout severed not just communication but coordination capacity - the ability of ordinary people to align plans across space and time. Thucydides calls it “domestic coordination capacity”; Nightingale, the “circuit” of medical and economic decision-making; Paterson, the “feedback loop” essential to price signals and trust. Crucially, they concur that this damage persisted after connectivity was restored, because trust in the signal had eroded. This is a deep structural agreement: the harm was not merely in the outage, but in the lasting degradation of the system’s capacity to self-correct. It reveals that the blackout’s legacy is not blackout duration, but the resistance of the system post-reconnection - a metric no official report measures, yet all three treat as decisive.
- Finally, they agree that the regime’s internal logic cannot be understood by applying democratic assumptions to an authoritarian context. Thucydides invokes the Melian Dialogue to stress that power asymmetry determines outcomes regardless of moral framing; Nightingale, by comparing Iran’s shutdown to Scutari, insists that institutional inertia masquerading as procedure produces the same preventable harm regardless of stated intent; Paterson, by invoking the “Long Circuit,” argues that blockages in authoritarian systems are structural, not exceptional. Their convergence here is not moral - it is analytical. They all reject the idea that the regime acted as if it were accountable to its people, and they all treat that rejection as non-negotiable. This shared analytical commitment is the quiet foundation beneath their louder disagreements.
Where They Fundamentally Disagree
- The first irreducible disagreement is over whether the blackout’s primary effect was suppression by omission (Nightingale) or preemptive circuit pruning (Paterson) versus tactical recalibration of threat perception (Thucydides). Empirically, they dispute what the pattern of restoration reveals: Thucydides sees it as evidence of internal security logic; Nightingale, as evidence of failed triage standards; Paterson, as evidence of circuit reconfiguration. Normatively, Thucydides accepts the regime’s calculus as rational - even tragic, but inevitable - while Nightingale and Paterson treat it as a preventable failure of responsibility, not inevitability. Nightingale’s steelman of Thucydides would be: the regime correctly identified that domestic unrest posed a greater threat than economic disruption, and acted accordingly, even if the cost was borne by civilians. Paterson’s steelman of Thucydides would be: the regime acted not to manage risk, but to reset a circuit it could no longer steer, and the civilian toll is the price of that structural intervention. Neither Nightingale nor Paterson disputes the fact of the regime’s calculus, only its moral weight in the explanation.
- The second fundamental disagreement is over whether the blackout introduced harm (Nightingale) or revealed pre-existing fragility (Paterson). Empirically, Nightingale treats the blackout as a new source of preventable harm - delayed medical care, lost coordination, eroded trust - while Paterson treats it as the visible manifestation of a circuit already compromised by years of control. Thucydides sits between them: he accepts the harm is real but insists it was expected, not accidental. Normatively, Nightingale holds that preventable death cannot be justified by systemic fragility; Paterson holds that the system’s design made the harm inevitable; Thucydides holds that the harm was the price of clarity for the regime. The empirical divergence hinges on whether the baseline coordination capacity can be reconstructed from pre-blackout data - a question Thucydides sidesteps, while Nightingale and Paterson answer with conflicting assumptions about the resilience of informal networks.
- The third fundamental disagreement is over whether the internet is a neutral conduit (Thucydides’ implicit view) or a structured system (Nightingale and Paterson’s shared view). Thucydides treats the blackout as an intervention on a system whose output (economic coordination, political stability) can be measured independently of its architecture. Nightingale and Paterson, by contrast, treat the architecture itself as the site of political contestation: for Nightingale, the denominator (how harm is measured) reveals intent; for Paterson, the circuit’s path determines whether energy flows to self-sustaining actors or state-managed channels. This is not a factual dispute but a conceptual one: whether coordination is a function of the network or an emergent property of its design. Thucydides’ steelman of Paterson would be: the blackout was not just an interruption, but a deliberate rewiring of the system’s feedback loops to favour regime-aligned nodes. Paterson’s steelman of Thucydides would be: the regime did not reconfigure the circuit; it simply severed it where control was threatened, and the downstream effects were secondary to the immediate need for clarity.
Hidden Assumptions
- Thucydides: The regime’s internal security calculus operates with the same rationality as statecraft between sovereign powers - i.e., it follows predictable patterns of interest, fear, and honour, and can be inferred from observable action. This assumption is contestable because authoritarian regimes often act on irrational fears - such as the belief that any unverified image is an existential threat - or on pre-rational instincts, like the preservation of ritual dominance, that do not yield to cost-benefit analysis. If the regime’s actions were driven by panic rather than calculation, the entire cartographic model would misfire.
- Florence Nightingale: The civilian harm caused by the blackout can be measured per capita or per person-day, and that the denominator - number of people exposed to harm - is the correct metric for ethical assessment. This assumption is contestable because it treats all harm as comparable across contexts, ignoring that rural populations may have lower baseline reliance on external networks, or that informal coordination (e.g., word-of-mouth) may compensate for digital silence. If harm is not linearly scalable across infrastructure types, the entire humanitarian argument rests on a flawed metric.
- Paterson-style: The internet functions as a circuit whose integrity is a necessary condition for a healthy civil society - i.e., that coordination, trust, and economic initiative are not merely possible without it, but depend on its unobstructed flow. This assumption is contestable because many societies have coordinated complex activity without digital networks (e.g., pre-internet Iran, or pre-industrial economies), suggesting the circuit may be reconfigurable rather than irreplaceable. If coordination can be rebuilt through non-digital means, the blackout’s legacy may be less structural than Paterson assumes.
Confidence vs Evidence
- Thucydides: Claims HIGH CONFIDENCE that the tiered restoration pattern reflects “internal security logic, not external threat management” - but cites no technical or operational data to distinguish between the two. The assertion is plausible, but the confidence tag masks the absence of evidence: a true external threat response would indeed treat the territory as a single operational zone, but a hybrid response (e.g., prioritising urban centres for economic stability) is possible. The confidence is overdrawn.
- Florence Nightingale: Claims HIGH CONFIDENCE that “disabling a system’s ability to observe itself is not defense, but concealment” - but offers no evidence that Iran’s system could observe itself had connectivity remained. The claim assumes that network resilience includes self-diagnosis, which is not technically guaranteed. The confidence is misplaced because it treats a normative claim (concealment is bad) as an empirical one (the system was designed to self-diagnose).
- Paterson-style: Claims HIGH CONFIDENCE that the blackout was “preemptive circuit pruning” - a term she defines as resetting the system to a state of supervision - but provides no evidence of reconfiguration after restoration (e.g., new routing protocols, altered node permissions, or surveillance upgrades). The confidence is high, but the evidence is inferential: it assumes the regime’s intent matches the outcome, without showing how the new state differs from the old.
- Paterson-style: Both express HIGH CONFIDENCE on the timing of the blackout as deliberate (coinciding with external strikes, not domestic unrest), but their evidence is circumstantial - pattern recurrence (Thucydides) and structural logic (Paterson). Nightingale, by contrast, shows LOW CONFIDENCE on the 17% year-on-year data traffic increase, correctly noting its unverified source. This asymmetry reveals that the most consequential claim - the occasion of the blackout - is treated with high confidence by the others, while the most granular data point is treated with low confidence - suggesting a misalignment between evidentiary weight and claim significance.
What This Means For You
When evaluating coverage of internet shutdowns, demand evidence that distinguishes regime intent from regime capacity - specifically, whether the blackout was a response to a threat or a preemption of a perceived vulnerability. Ask whether restoration schedules were assessed against medical need or political priority, and whether any data was collected on trust degradation post-restoration - not just signal availability. Most suspicious of all are claims tagged HIGH CONFIDENCE that rely on pattern recurrence alone, especially when the recurrence is drawn from ancient history without showing how the mechanism applies identically across millennia. The single most revealing piece of evidence any news outlet should produce is a comparative analysis of coordination failure - e.g., whether informal networks (e.g., community health workers, local traders) reported measurable delays in decision-making after connectivity returned, indicating that the system’s resilience, not just its signal, was damaged.