The chart is by Andreas Persbo, who has another interesting blog post expressing skepticism about the IAEA’s assessment that Iran may install 3,000 centrifuges by August and 8,000 centrifuges by the end of the year.

“Although ElBaradei is a very capable guy,” Andreas concludes, “I personally remain sceptical to any claim that the Iranian’s suddenly constructs at a pace much higher than historically proven.

To help, Andreas made that graphic, as well as a table.

Date Historical 1 per week 2 per week
7 January 07 656
16 April 07 1312
13 May 07 1640
2 July 07 2247 2810 3980
21 August 07 2855 3980 6320
10 October 07 3462 5150 8660
29 November 07 4070 6320 11000

Source: Andreas Persbo, “Iranian centrifuge construction,” Verification Matters, June 15, 2007.

Mark Heinrich with Reuters has a very good story that mentions some of the same questions that Andreas raises, particularly whether the question of whether the centrifuges are working as well as expected.

Damn the Torpedoes, Full Speed Ahead

Whether or not Iran keeps this pace, it seems that the Iranians are rushing to build large numbers of centrifuges that don’t work very well. (Paul has a nice roundup on some of the problems.)

That, to me, is very, very telling.

Iran—far from complying with UN Security Council demands for suspension—is racing, either for a bomb or to create “facts on the ground” that will present the United States and the EU-3 with a fait accompli.

Either way, Iran’s centrifuge program now seems driven by politics not technology.

Pakistan Did It, Too

”[T]he amount of centrifuges in there isn’t that important,” Persbo argues, adding that “the size of the facility already exceeds what Pakistan had at the beginning of their nuclear weapons programme.”

That’s true—Pakistan’s centrifuge facility at Kahuta probably had around 1,000 machines at the beginning of Pakistan’s bomb program. (Mark Hibbs reports the number of centrifuges in Kahuta as having now stabilized around 3,000.)

There is another, more apt comparison between Pakistan and Iran that ought to be made. Pakistan, too, assembled large numbers of sub-optimal centrifuges that frequently crashed as part of a, well, crash bomb program.

Hibbs recently quoted a pair of Western officials describing Pakistan’s early centrifuge efforts by noting that a would-be nuclear weapons-state might be …

... willing to take decisions and shortcuts which would mean that the initial failure rate of the machines might be as high as 10% and that ‘’after two or three years of operation, a very large number of machines would crash.’’

This, another Western official said, was ‘’exactly the route’’ followed by Pakistan during the early years of its centrifuge program in the 1970s and 1980s when it set up its first SWU plant at Kahuta. ‘’They built a lot of simple machines, there were lots of holes in the diagnostics, so they could make HEU in a hurry,’’ one official recalled1.

In 1992, US officials confirmed to Mark Hibbs and David Albright that Pakistan may have built as many as 14,000 centrifuges, out of which only 1,000 operated at Kahuta. Pakistan’s “junk pile,” according to one official “is sizeable2.”

What Do We Do?

The difference between Pakistan and Iran, of course, is that inspectors still have (limited) access to the Iran. If Iran starts churning out HEU with the current set of centrifuges, we’re likely to know before they have enough for a bomb.

At this point, Iran seems to be racing to create facts that will improve its bargaining power and, at a later, date give Tehran a bomb option.

Our top diplomatic priority ought to be getting a “warm standby” solution—where the centrifuges spin empty—as the technical basis of an interim pause during which Iran returns to compliance with the Additional Protocol (so we can inspect the centrifuge workshops). We can then negotiate some kind of multinational arrangement that imposes a much more intrusive inspection regime.

Keeping inspections in place, and getting Iran back into compliance with the Additional Protocol, is probably more likely to contribute to our security than toothless demands that for a suspension to which Iran responds by speeding up its program.

I think about policies to constrain and monitor Iran’s program over time, integrated as part of a larger coercive diplomacy to manage Tehran. I’ve written before about a hypothetical policymaker with an “indifference curve that defines the rate at which she is willing to allow Iran access to enrichment technology in exchange for better capabilities to monitor Iran’s activities.”

That, not coincidentally, is the subject of a little workshop that the New America Foundation is hosting with Stanley Foundation and the American Association for the Advancement of Science in mid-July.

Not exactly Led Zeppelin tickets, I know, but it should be quite a show.

footnotes

1 Mark Hibbs, “CIA Assessment On DPRK Presumes Massive Outside Help On Centrifuges,” Nuclear Fuel 27:24, November 25, 2002, 1.

2 David Albright and Mark Hibbs, “Pakistan’s Bomb: Out of the closet,” Bulletin of the Atomic Scientists 48:6, July/August 1992, 38-43.