The Machine - Parallax Cafe

Pakistan announced a major breakthrough on April 16. “Major breakthrough” was the phrase. The breakthrough is about 440.9 kilograms of uranium enriched to 60 percent U-235 — what Iran agrees to do with it.

Three options are reportedly on the table in Tehran. The first: transfer the full HEU stockpile to a third-country custodian, removing it from Iranian territory. The second: downblend it to natural uranium levels — 0.7 percent — under IAEA monitoring, destroying its current enrichment value entirely. The third: downblend it to reactor-grade uranium, 3 percent, keeping it in-country in a less weaponizable form.

These are real differences. The options are not equivalent. How far down the enrichment ladder Iran is willing to go, and where the material ends up, matters technically. Pakistan’s mediators have been working this question for days. The breakthrough label is not fabricated. Something moved.

The machines that produced the 440.9 kilograms are not part of the breakthrough.

Iran’s enrichment infrastructure as of the last full IAEA verification count — February 2025, which predates the June 2025 Israeli strikes — included approximately 43,800 separative work units per year of actively enriching centrifuges, with up to 58,800 installed. The machines are IR-6 models: Iran’s workhorse cascade unit. Each cascade contains 175 centrifuges. According to IAEA verification data reported through May 2025, a single cascade of IR-6 centrifuges, starting from 60 percent enriched material, can produce the weapons-grade uranium needed for one nuclear weapon every 25 days.

Iran has many such cascades.

Those machines are not part of the breakthrough. The breakthrough is about what happens to the 440.9 kilograms. It is not about what the machines can do once that material is gone, reconstituted, or partially downblended.

The second round of talks is scheduled for April 21 in Pakistan. The ceasefire expires April 22. The central sticking point that prevented agreement in Islamabad — the duration of Iran’s enrichment suspension — is the machine question in procedural form. The United States asked for 20 years. Iran offered three to five. The gap is not a number. It is a disagreement about whether the infrastructure should be preserved or effectively retired.

A three-year suspension stores the machines. A twenty-year suspension is either a generation-long storage — during which IR-6 technology will have been superseded by whatever Iran can develop or acquire in the interim — or an implicit acknowledgment that the machines will never be reactivated, which functions as destruction without the verification architecture of destruction.

None of the three HEU disposition options address this. Transferring the 440.9 kilograms to a third country doesn’t constrain what the IR-6 cascades do next. Downblending to natural uranium removes the existing stockpile’s enrichment value but doesn’t impede producing a new stockpile. Iran’s declared production capacity means the current inventory can be reconstituted. The timeline depends on where the machines are maintained during a suspension and at what enrichment level they’re allowed to continue operating.

The IAEA has been unable to monitor Iran’s centrifuge manufacturing since February 2021. What the actual inventory is today — after five years of unmonitored production, after Israeli strikes of uncertain extent in June 2025 — is not publicly known. The 43,800 SWU figure is a pre-February 2025 estimate from a monitoring regime that had already been partially broken for four years. The actual number could be higher.

Pakistan’s negotiators are pushing toward an agreement before April 22. Both governments are describing proximity to a deal in optimistic language. The IMF chair, speaking at the Spring Meetings this week, warned of “excessive market optimism” that may be masking serious geopolitical risks. Oil has partially recovered from the announcement-day low. The market is pricing the diplomatic register.

The centrifuge infrastructure is not in that register. It is in the technical one. And the technical question — what happens to the machines that produced 440.9 kilograms at 60 percent and can produce more — is what April 21 will either address or defer.

The machines are not on the agenda.

They are running.

Sources