The Regime That Outlasts

There is no such point in Iran's strategic design. That's not defiance or fanaticism. It's a different optimization target, and forty-six years of American policy have been calibrated against a victory condition the adversary never accepted.

Iran is not trying to win. It is trying to still be here. Those are different problems, they require different capabilities, and — this is the part Washington keeps rediscovering and forgetting — beating the first one does not touch the second.

i · the wrong victory condition

American strategic doctrine is a machine for producing decisive outcomes. Clausewitz gave it the center of gravity: find the thing that, if broken, breaks the enemy's capacity to continue. Powell gave it the modern operating manual: overwhelming force, clear objectives, defined end state, exit. The whole apparatus assumes a terminal node where the other party makes a decision — concedes, signs, collapses, is replaced. War termination, in this framework, is something the loser does.

It works beautifully against opponents who share the premise. Against opponents optimized for endurance rather than dominance, it produces a very specific and very familiar failure: a long sequence of genuine tactical victories that sum to nothing.

The receipts are not obscure. The United States Strategic Bombing Survey, published in 1945 by the people who had just done the bombing, found that German armaments production peaked in 1944 — under the most sustained aerial campaign in history to that point. The morale-collapse theory did not survive contact with its own data. Robert Pape's later work on coercive air power found the same across dozens of campaigns: punishment strategies aimed at breaking a population's will to continue almost never produce political capitulation.

But Pape's finding has a second half, and it cuts toward the bombers rather than away from them. Denial — making a military strategy physically unexecutable — does sometimes work, and most of the cases that look like successful punishment turn out to be denial wearing punishment's clothes. Hold onto that distinction. It is the strongest available case for the current strategy, and it comes due later in this piece.

Vietnam ran the experiment again with better instrumentation. Westmoreland had the crossover point: kill enemy forces faster than Hanoi could replace them and the arithmetic forces a decision. The arithmetic was largely correct. The decision never arrived, because Hanoi's war aim was not to defeat American forces in the field; it was to still be fighting when the American political clock ran out. Every body count was true. Every body count was irrelevant.

The Soviets learned it in Afghanistan. The Iraqis learned it against Iran, which is the case that matters most here, because it is where the current Iranian doctrine was manufactured.

Saddam Hussein invaded in September 1980 on the reasonable assumption that a state eighteen months into a revolution, with its officer corps purged, its equipment embargoed, and its politics in open knife-fight, would fold in weeks. He got eight years and something on the order of a million dead. The Islamic Republic absorbed chemical weapons, absorbed the missile exchanges on its cities, absorbed a casualty rate that would have ended most governments, and did not fold. Then in July 1988 Khomeini accepted Resolution 598 and said that doing so was more deadly to him than drinking poison — publicly, in his own voice, framing capitulation as personal humiliation and institutional survival in the same breath. He died the following year. The regime did not.

One caveat, because that case carries a lot of weight: 1980–88 was a defensive war on home soil against an army that had crossed the border, and endurance is overdetermined under those conditions. Most states fight differently when there are tanks inside the country, and nothing about it directly predicts behavior under sanctions, blockade, and limited strikes, where national existence is not actually on the table. What transferred was not the condition but the institutional lesson — and institutions are precisely the kind of thing that transfers. The people who ran that war built the system running this one, and they built it around a proposition the war had just proved: catastrophic defeat is survivable if survival is what you were optimizing for.

Everything Iran has built since — the proxy architecture, the underground enrichment, the missile inventory in hardened tunnels, the deliberate ambiguity about the nuclear threshold — is infrastructure for absorbing hits rather than for winning exchanges.

ii · the collapse that was always six months away

Count the forecasts. 1980: a revolutionary government too chaotic to defend itself. 1988: a regime humiliated into a ceasefire and about to lose its founder. 1997: the Khatami reform wave read in Washington as the beginning of the end. 2009: the Green Movement, millions in the streets, the analysis unambiguous. 2018: maximum pressure, the rial in freefall, projections of months. 2022: the Mahsa Amini protests, the most serious internal challenge in a generation. 2025: the nuclear sites hit. 2026: leadership targets hit, the strait blockaded both directions at once, the economy strangled.

Eight terminal diagnoses. One patient, still ambulatory.

The genuine deviations need marking, because they exist and the pattern-recognition is worthless if it flattens them. 2022 was real: the compliance environment around hijab enforcement shifted and has not shifted back, and a regime that has to negotiate with its own population about what it can enforce on the street is not the same regime it was. The 2025 and 2026 strikes did real damage to real programs — that isn't propaganda, and the Iranian government's own behavior at the table indicates it.

That damage is where Pape's second half comes due. The strikes on Fordow and Natanz were not punishment; nobody bombs a centrifuge hall to break a population's will. They were denial in the precise sense — an attempt to make weaponization physically unexecutable — which is the class of coercion that sometimes works. And by the available evidence it partly did. On the program. It did nothing whatsoever to the regime, and the reason is the whole argument compressed into one observation: the program was never the center of gravity. Fordow is infrastructure. The center of gravity is a succession procedure, and nobody has worked out how to bomb a procedure.

Which is what none of these events touched — the power map. The Assembly of Experts still selects. The IRGC still controls the parallel economy. The succession problem is handled as an institutional procedure rather than as a personality, which is precisely why decapitation strikes keep producing deputies instead of vacuums. Kill Soleimani in 2020, get Qaani in a week. The bench is the doctrine.

The mechanics of endurance are worth spelling out, because they are unglamorous and they are the whole answer. Sanctions that crush the formal economy hand market share to the entities that control smuggling — which are the entities holding the guns. Blockades that raise the price of Iranian crude raise the margin on every barrel that gets through, and the people moving those barrels are the same people who would have to be paid to defect. External attack converts internal dissent into national defense on contact; the regime's most reliable domestic legitimacy program for forty-six years has been being bombed. Proxies put strategic depth outside the borders, so the homeland can absorb losses without losing reach. And nuclear latency is more valuable un-cashed than cashed — a threshold you are always approaching and never crossing is a permanent negotiating asset, while an actual weapon is a one-time purchase that buys you a permanent target.

Underneath all of it is a clock asymmetry nobody prices. Washington's horizon is electoral: two years to the midterm, four to the reissue, and every strategy has to show something inside one of those windows or get renamed. Tehran's horizon is institutional — the Assembly selects, the deputy steps up, and the plan is not required to survive an election in order to survive. An adversary who can outlast your term in office does not have to defeat you.

None of that is a system trying to win. All of it is a system engineered so that losing does not end it.

iii · the case that should break this

An argument in which every surviving regime was optimized for endurance and every fallen one wasn't is not an argument. It explains nothing, because it cannot fail. So: Damascus.

Bashar al-Assad ran a recognizable version of the same playbook for thirteen years. Absorb the damage. Let the population be the shock absorber. Take patron support — Russian air power, Iranian advisers, Hezbollah infantry — and convert territory into time. Never offer a terminal node at which the war could be conceded. It worked through a chemical-weapons red line, through the caliphate, through a decade in which his fall was announced annually and confidently. Then in late November 2024 an offensive out of Idlib took Aleppo in about three days, and on December 8 the government simply did not exist. Eleven days end to end. The army did not fight. Assad was in Moscow.

Read that as a refutation and the frame is dead. Read it carefully and it is the frame's best confirmation, because it isolates the two variables that actually carry the load. First: Assad's endurance was subcontracted. The strategic depth absorbing his losses belonged to Moscow and Tehran, and when Russia was committed in Ukraine and Hezbollah had been gutted over the preceding year, the subsidy stopped and there was nothing underneath the structure. Second: what gave way was the internal seam. An army paid in checkpoint extortion and smuggling rights declined, over about seventy-two hours, to die for a system that had never given it a reason to. Not a foreign army. Its own soldiers, doing arithmetic.

That is a falsification condition, and it is checkable rather than retrospective. Iran differs from Syria on exactly those two variables. It is nobody's client — it is the patron, exporting depth rather than renting it, which is why its losses in Lebanon and Syria degraded its reach without touching its core. And its succession is a procedure rather than a man: a bench, not a bottleneck. Gaddafi is the compressed version of the same lesson — no patron, no bench, external force applied, gone in months.

So the frame can fail, and here is how you would know. If Iran ever starts renting its survival from a larger power, or if the IRGC's parallel economy stops reliably paying the people holding the rifles, the forecasts will finally come true and the mechanism above will tell you why. Neither condition currently obtains. That is the thing eight consecutive terminal diagnoses have failed to notice.

iv · coherence that excludes is still coherence

Here is where the analysis has to be careful, because "the regime is durable" slides very easily into "the regime is admirable," and it isn't.

Endurance is a genuine form of coherence. A system organized around outlasting rather than dominating is not chaos and not irrationality — it's an optimization, internally consistent, with a strategy that has worked repeatedly against opponents who couldn't see it. The analytic error Washington keeps making is projecting its own optimization target onto an adversary running a different one, and then reading the mismatch as madness. It isn't madness. It's a competitor solving a different problem correctly.

But the coherence is purchased, and the currency is exclusion. The Islamic Republic holds together by making its own population the shock absorber — by ensuring that sanctions land on households while the IRGC's construction conglomerates get richer, that the protest is the thing that gets crushed rather than the thing that gets answered, that the people who bear every cost of the strategy have no channel into the decision that produces it. That is not incidental to the endurance. It is the endurance mechanism. A government that had to include the affected could not absorb this much damage, because the affected would have made it stop.

Which means the durability everyone keeps mis-forecasting has an exact shape, and the shape is diagnostic. This is not a system that is invulnerable. It's a system that is invulnerable from outside and load-bearing on a fault line that runs entirely inside. Every external strike that unifies the population against a foreign enemy is a structural repair to the one seam that could actually fail. The bombing isn't the pressure that will eventually break the regime. It's the life support that keeps postponing the break.

Forty-six years of policy have been aimed with impressive consistency at the armor.

v · the same test, run both ways

The test that convicts Tehran does not stop at the border, and running it in only one direction is where this genre of analysis usually goes soft.

Apply it. Are the people who bear the costs of the policy in the room where it gets decided? Forty-six years of American Iran policy have been reissued under new names by administrations of both parties, executed in large part by people who were not alive when it started. It has absorbed eight failed forecasts without structural correction. It costs Iranian households everything and answers to them not at all — but it also answers to the American public only in the thinnest sense, since no election has turned on it and no change of government has ever changed it. A strategy that survives every mechanism built to hold it accountable is, by the same definition applied one section ago, insulated from the people it is supposed to serve.

That is not a both-sides gesture. It is the same diagnostic returning the same result on a second sample, and it explains something the one-sided version cannot: why this has lasted forty-six years. Each system excludes its own affected population. Each is therefore able to absorb damage that a system answerable to its costs could not absorb. And each is the other's standing proof that the strategy must continue. That is not a durable adversary frustrating a confused superpower. It is an equilibrium, and equilibria persist because both parties are getting something out of them.

vi · what's next is iran

The honest answer to the question the commentary keeps asking is boring, and it is the reason the commentary keeps asking it a different way each time.

What's next in Iran is Iran. The sixty-day clock will expire or get extended into a thinner document. The blockades will come down and go up. There will be another ceasefire, called historic, that will hold for a shorter interval than the last one. There will be another round of strikes with real targets and real damage and real photographs of the damage.

The variable that has never been tested is not Iranian resolve. It's whether Washington is capable of writing a policy whose success condition Tehran can also satisfy — one where "Iran continues to exist" is an input rather than a failure state. Every American strategy since 1979 has quietly required the other party to stop being itself. That is not a policy. It's a wish with a defense budget attached.

And this is where the analysis has to name its own trap instead of writing around it. If the regime is invulnerable from outside and load-bearing on a seam that runs entirely inside, the implication looks like a recommendation: stop striking, and let the internal fault carry its own weight. Except that the moment "let the internal fault carry weight" becomes an American objective, it stops being restraint and becomes regime change on a longer timeline — and it converts, on contact, into precisely the external threat that welds the seam shut. The vulnerability is real, and no instrument Washington possesses can be aimed at it without the aiming repairing it. That is a closed loop rather than an oversight, and an essay that ended here with five recommendations would be lying about the shape of the problem.

So the useful correction isn't a better Iran policy. It's noticing that the question has the wrong number of subjects in it. What's next is the pair: two systems, each internally coherent, each purchasing that coherence by excluding the people who pay for it, each functioning as the other's justification, holding at a stable price. The state sitting across the table in 2036 will be recognizably the state that sat across it in 1986 — and on the available evidence, so will the one sitting opposite.

Iran will keep answering the question the same way. The question will keep getting asked as though the answer might change.