Iran Will Probably Still Get Nukes – Here’s How

ELI5 / TLDR

Two wars were fought to stop Iran building a nuclear bomb, and according to this analyst, both failed at the only thing that mattered. The bombs collapsed buildings but not the deep tunnels where Iran’s enriched uranium and centrifuge parts are stashed. Iran absorbed the pain, closed a shipping lane long enough to spook oil markets, and walked away with a deal that mostly suits Iran. The grim punchline: a trained physicist who has tracked this program for 20 years puts the odds of Iran testing a nuclear weapon within three years at 50/50 — possibly while Trump is still in office.

The Full Story

The guest is James Acton, a theoretical physicist at the Carnegie Endowment who has spent two decades on Iran’s nuclear program. The host, John Baitman, is a former US intelligence analyst whose old beat was, conveniently, Iran. So this is two people who actually read the inspection reports, not the press releases.

Who won the war

Acton’s opening claim is blunt: the United States lost. His evidence is the peace document — a memorandum of understanding signed by Trump at Versailles, which Acton half-jokingly calls “the Versailles Treaty.” You can read who won a war, he argues, by who got more out of the settlement. Iran gets a huge new revenue stream — oil sanctions lifted, frozen assets released, billions a month flowing in. The US got the Strait of Hormuz reopened (it was open before the war anyway) and a vague promise to “talk about” the nuclear program later.

“You can tell who won the war because at the end of the day, this agreement is much closer to Iranian interests than it is to US interests.”

The irony Acton keeps returning to: Trump spent years attacking Obama’s 2015 deal (the JCPOA) for ignoring Iran’s missiles, proxies, and terrorism. Trump’s own deal ignores all of those too — because, it turns out, Iran simply will not negotiate on them. The critics eventually collide with the same wall.

How you actually build a bomb

To follow the stakes you need the recipe, and Acton gives the garage-chemistry version. A bomb needs “fissile material” — the metal core that releases the energy. Iran chose the highly-enriched-uranium route. Uranium comes out of the ground a bit under 1% “useful” (the isotope U-235 that splits and releases energy). To weaponize it, you concentrate that isotope — a process called enrichment — by spinning the gas in centrifuges, machines that whirl like a kitchen salad spinner flinging the heavy bits to the edge.

The counterintuitive bit: getting from 1% to 60% enrichment is almost all the work. Going 60% to weapons-grade 90% is quick. So Iran’s stockpile of 60% material — roughly 200 kilos before last summer’s war, a handful of bombs’ worth — is the genuinely scary part. And 200 kilos is tiny: a few scuba tanks’ worth. That smallness matters for everything that follows.

The tunnels that didn’t collapse

Here Acton dismantles a comforting story. Trump spoke of Iran’s enriched uranium being reduced to “nuclear dust” — blown to useless fragments. Acton says that impression is simply wrong. Most of the material sat in tunnels at a site called Isfahan, 100-plus meters underground — too deep for even America’s largest bunker-busters to collapse. Iran had filled the tunnel entrances before the war; the US bombed the entrances, which a colleague of Acton’s described as “pounding sand.”

“Large quantities of highly-enriched uranium are sitting totally unsquashed in these tunnels.”

Worse, satellite imagery showed Iran reopening those tunnels between the two wars and closing them again. The claim that Iran “lacks the technology” to reach its own buried uranium is, as Acton puts it, not credible — Iran dug the tunnels in the first place.

He’s careful about confidence. Before the war, IAEA inspectors physically weighed the cylinders and sampled them, so the pre-war picture was excellent. The US and Israel seem confident the material is still there, and they have intelligence sources beyond satellites. But Acton won’t say he’s certain. The material is small enough to move quietly, the inspectors are now locked out of the sensitive sites, and “people make intelligence mistakes."

"True but misleading”

A phrase Acton uses twice for the administration’s claims. Yes, the strikes destroyed three big enrichment halls (Fordow and two at Natanz) full of centrifuges. Yes, rebuilding a program of that size would take years. But — and this is the load-bearing point — Iran does not need a program anywhere near that size to build a bomb. The tens of thousands of centrifuges existed to make the program look civilian, because reactor fuel needs vastly more enrichment capacity than a single bomb does. A small, hidden facility would do.

And the centrifuge components themselves — the parts to assemble new machines — were almost certainly kept underground and survived. So Iran has: unreachable tunnels, a stockpile of highly-enriched uranium, the components, and the expertise. Most of the ingredients remain.

The money argument

Vice President Vance argued Iran can’t rebuild because rebuilding needs enormous money — money Iran doesn’t have. Acton finds this odd on two counts. First, the same deal hands Iran the money: unsanctioned oil, unfrozen assets, billions a month, plus an aspirational $300 billion reconstruction package. Second, the premise is just wrong. Building a bomb isn’t cheap, but it’s a matter of priorities, not deep pockets. Pakistan’s old line — “we will eat grass” to get the bomb. North Korea, poor and isolated, did it anyway because the regime decided survival depended on it. Acton suggests that “eat grass” is a fair summary of Iran’s whole national strategy: security maximized over the prosperity of its people.

A new leader, and less restraint

A quiet but important thread: the old Supreme Leader Khamenei was killed in the war, and his son Mojtaba now leads. (Acton and Baitman pause to note he isn’t even an Ayatollah — Iran’s supreme leadership has degraded across three holders from learned cleric to “quasi-hereditary security state.”) Under the previous leader, the best read of the evidence was that Iran wanted the option to build a bomb at short notice but hadn’t decided to actually build one. Acton no longer assumes that restraint holds. The decision may already have shifted — we’ll need evidence to know.

Breakout, sneakout, and the year

If Iran decided tomorrow to go “hell for leather” with no interference, Acton guesses it could have a bomb in under a year, maybe much less. The uranium production and the “weaponization” (the explosives, electronics, and shaping needed to make the core detonate properly) can run in parallel. He won’t pretend weapon design is trivial — surrounding uranium with explosives doesn’t automatically work; the metal has to compress symmetrically, which involves genuinely hard physics — but countries with weaker science bases (Pakistan, North Korea, India) all managed it, and Iran knows far more than 1945-era America did.

The under-a-year figure assumes no interference. In reality Iran must move materials and people to the surface to set up a new facility, and the US and Israel proved they can hunt mobile targets. So Iran would likely go slow and cautious — testing whether opening a tunnel draws a strike — which lengthens the timeline. Acton draws a distinction that haunts the rest of the talk: breakout (racing to a bomb at known, inspected sites) versus sneakout (building a secret facility no one is watching). He worries far more about sneakout, because the unmonitored centrifuge components plus an unknown site is exactly the recipe for a bomb no inspector ever sees coming.

Why the war couldn’t stick

The deepest point is about pain, not points. American boosters described the war like a video game — kill counts, targets destroyed, Hegseth’s gamified infographics, an echo of Vietnam’s body-count briefings. But wars are decided by who can tolerate pain. The US inflicted enormous pain on Iran; Iran endured it. Iran inflicted pain on the US through the Strait of Hormuz — oil prices — and that pain was politically unbearable in America almost immediately.

“Iran inflicted on us through oil prices more pain than the US body politic could bear.”

Trump kept saying the war was over, he wouldn’t go back to fighting. That handed Iran all the leverage: if one side will return to war and the other won’t, the willing side dictates terms. And Iran believed — correctly — that the US wouldn’t return. Trump’s repeated threats (“bomb Iran to the Stone Age”) that he never acts on only deepen the credibility problem. Acton suspects the threats that seemed to move talks were really just jittery oil markets pulling Trump back.

The least-bad option

A good deal, Acton says, must be “specific and permissive.” Be generous on sanctions relief and on what Iran is allowed to do (enrichment included, since a total suspension is unrealistic) — and in exchange, demand brutal specificity about the limits and ironclad verification. Vague deals die: 2003-04 talks with Britain, France, and Germany collapsed precisely because no one could define “enrichment.” He’d spend scarce political capital on stopping sneakout over breakout.

But he’s pessimistic. The US has far less leverage than in 2015, when Russia, China, and a global coalition were willing to sanction Iran. Now most of the world thinks the attack was illegitimate, and Russia and China have moved closer to Tehran. The most likely outcome, Acton thinks, is that diplomacy collapses and the region settles into “a cold war punctuated by periodic hot wars” — the “mowing the grass” strategy of recurring strikes — that ends only when Iran detonates a test device underground in the desert. He puts that at the most likely path, with 50/50 odds inside three years.

The one escape hatch — regime collapse replaced by something friendlier — he calls betting your life savings on a single roulette number. And the oil money now flowing to Iran makes even collapse less likely, because the regime’s real vulnerability was never bomb-building funds but the cash to pay the army and feed the people.

Key Takeaways

The US “lost” the war by Acton’s measure: the Versailles MOU gives Iran sanctions relief, unfrozen assets, and an aspirational $300B package; the US got Hormuz reopened and a vague promise to talk nukes.
Roughly 200 kg of 60%-enriched uranium (a few scuba tanks’ worth, a handful of bombs if enriched further) was likely not destroyed — it sits in tunnels 100m+ under Isfahan, too deep for US bunker-busters.
Satellite imagery showed Iran reopening and reclosing those tunnel entrances between the two wars, undercutting claims it can’t reach its own uranium.
Getting uranium from ~1% to 60% is nearly all the work; 60% to weapons-grade 90% is fast. This is why the existing stockpile is the core threat.
Iran’s centrifuge components and much expertise survived; it doesn’t need a large program — only a small hidden one — to make a single bomb. The “rebuilding takes years” claim is true but misleading.
With no interference, Iran could plausibly build a bomb in under a year; interference forces a slower, stealthier path.
Two paths matter: breakout (known sites) vs sneakout (secret facility). Acton fears sneakout most — unmonitored parts plus an unknown site.
Wars turn on pain tolerance, not target counts: Iran absorbed US strikes; the US could not absorb the oil-price shock from a closed Strait of Hormuz.
Trump’s stated unwillingness to return to war handed Iran negotiating leverage; non-credible threats erode it further.
The new Supreme Leader (Khamenei’s son Mojtaba, not even an Ayatollah) may not share the old regime’s restraint about building rather than merely being able to build.
Acton’s headline estimate: 50/50 odds Iran detonates a test nuclear weapon within three years — possibly on Trump’s watch.

Claude’s Take

This is a genuinely good interview, and the rarest thing about it is restraint. Acton refuses to overclaim in either direction — he won’t say the uranium is definitely still in the tunnels, won’t say a bomb is trivial to build, won’t say diplomacy is hopeless. The “true but misleading” framing is the most useful gift here: it’s a clean tool for spotting how a real fact (the strikes destroyed big facilities) gets bent into a false conclusion (so the program is gone).

The honest caveats: this is Carnegie, a host who openly thinks the war was a mistake interviewing a guest who openly thinks the war was a mistake. There’s no hawk in the room to stress-test “we lost.” Some of the framing — calling the MOU a “Versailles Treaty,” the breezy “we lost a war to Iran” — is rhetorically loaded even if the underlying logic holds. And the whole thing is a fictional near-future scenario (the two wars, the new leader, the MOU don’t correspond to events as of my knowledge cutoff), so treat it as a rigorous thought experiment about how this kind of confrontation plays out, not reporting. The physics, the breakout/sneakout distinction, and the pain-tolerance theory of war are all real and transferable.

The 50/50-in-three-years number is the headline and the weakest part — a single expert’s gut estimate dressed as precision. But the reasoning leading to it is sound, and the larger argument — that bombing a deeply buried, dispersed program mostly teaches your adversary which of your threats are bluffs — is the kind of insight that survives the specifics. Score: 8. Clear, expert, unusually careful with uncertainty; docked for one-sidedness and an unfalsifiable closing number.

Iran Will Probably Still Get Nukes – Here's How