ISIS Technical Assessment: Pakistan's Stock of Weapon-Grade Uranium
David Albright, President
Kevin O'Neill, Deputy Director

June 1, 1998

On May 28 and May 30, 1998 Pakistan conducted up to six nuclear tests. A Pakistani government statement said that the nuclear test devices contained "uranium-235," commonly referred to as "weapon-grade uranium" (WGU; uranium enriched to 90 percent or more of uranium-235). The statement said that this uranium was produced at the Khan Research Laboratories (KRL), called the Kahuta facility, near Islamabad. While Pakistan's tests remove any doubt about its ability to produce nuclear weapons, significant uncertainties surround estimates of Pakistan's WGU stockpile.

This analysis provides information about the size of Pakistan's WGU stock and the number of nuclear weapons that it could make from this stock. The estimates presented here draw from and expand upon Plutonium and Highly Enriched Uranium 1996: World Inventories, Capabilities and Policies, authored by David Albright, Frans Berkhout and William Walker and published by the Stockholm International Peace Research Institute and Oxford University Press in March 1997. Readers are referred to this source for additional information.

Because of uncertainties about the number of tests Pakistan conducted, and when it resumed WGU production, these estimates are preliminary. However, ISIS has concluded that Pakistan resumed WGU production at least several months ago.

How much WGU did Pakistan produce before its 1991 moratorium on further production?

Pakistan started producing WGU at Kahuta in the early to mid-1980s. U.S. and former Pakistani government officials have said on many occasions that Pakistan voluntarily halted production of highly-enriched uranium (HEU; uranium enriched to 20 percent or more of uranium-235) at Kahuta in 1991. Pakistan is estimated to have produced 160 kg to 260 kg of WGU, with a central estimate of 210 kg, at Kahuta through the end of 1991.

How much WGU is in a Pakistani nuclear weapon?

Pakistan's nuclear weapons are believed to be based on a Chinese design that has a solid core of WGU. This design need not require more than 15 kg of WGU, but a number of factors might increase this amount. A greater amount of WGU could be used to increase the weapon's reliability or yield. It is also likely that some WGU would be lost during the weapon-production process. Therefore, ISIS has concluded that an average of 20 kg of WGU per Pakistani weapon is a reasonable estimate.

Up to the end of the moratorium, therefore, Pakistan had sufficient WGU for 10 nuclear weapons, based on a central estimate of 210 kg of WGU. If Pakistan had small losses during the production process, and each weapon used an average of 15 kg of WGU, Pakistan would have had enough WGU for 14 weapons.

Has Pakistan resumed WGU production?

It is likely that Pakistan resumed WGU production at Kahuta many months ago. Some speculate that WGU production resumed several years ago. Information about operations at Kahuta since 1991 is uncertain, since Pakistan has not allowed the international community to verify its moratorium.

U.S. government efforts to verify Pakistan's moratorium on HEU production have involved intelligence operations and are therefore classified. However, U.S. officials periodically stated in the early and mid-1990s that the moratorium was in place.

If WGU production was resumed, how much WGU could Pakistan have possessed at the time of its nuclear tests?

The amount of WGU that Pakistan could have produced at Kahuta depends upon the "feed stock" used in the uranium enrichment process. Feeding low-enriched uranium (LEU) into Kahuta's enrichment cascade would significantly increase Pakistan's ability to produce WGU. Pakistan produced LEU at Kahuta during the moratorium, but uncertainties exist about the quantity and enrichment of this LEU. Taking into account production inefficiencies and plant renovation, this LEU stock is estimated to be equivalent to about 4.5 to 5.5 years of total Kahuta production, and to have an enrichment of 5 to 20 percent uranium-235.

If Pakistan resumed WGU production in January 1998 using LEU as feed, it could have more than doubled its WGU stock within the last five months to a central estimate of roughly 490 kg. If Pakistan resumed production in March, after the Bharatiya Janata Party (BJP) won a plurality in India's most recent national election, Pakistan could have a stock of 425 kg of WGU.

Using natural uranium as feed, the Kahuta plant is estimated to be able to produce about 80 kg to 140 kg of WGU annually, with a central estimate of 110 kg. Based on this central estimate, if Pakistan resumed WGU production in January 1998 using natural uranium as feed, it could have increased its stock to 260 kg through May 1998.

If Pakistan produced WGU after 1991, contrary to Pakistani government statements, ISIS estimates that Kahuta could have produced a total of 680 kg of WGU from the mid-1980s through May 1998. This estimate assumes that WGU production did not occur for 1 to 2 years during the period of mid-1991 through May 1998.


How many nuclear weapons does Pakistan have following its May 28 and 30 nuclear tests?

Pakistan's nuclear tests consumed WGU that would otherwise be available for nuclear weapons. Uncertainties about the number and type of Pakistan's tests make it difficult to estimate the quantity of WGU consumed in these tests. U.S. officials are skeptical about Pakistan's announcement of six tests, and believe that fewer tests may have been conducted. Media reports, some citing anonymous Pakistani government sources, attribute different yields to the tests. At least one Pakistani media report has indicated that three of the tests were in the sub-kiloton range. Until better information about Pakistan's tests is made available, ISIS has decided to accept that Pakistan conducted six tests.

The quantity of WGU used in Pakistan's nuclear tests is unknown, and may differ from the quantity of material used to assemble nuclear weapons. The sub-kiloton test devices may have contained significantly less WGU than a Pakistani nuclear weapon; the reported 25 to 40 kiloton test may have consumed more material.

ISIS has decided preliminarily that the test devices contained an average of 15 kg of WGU, or 90 kg for all six tests. Therefore, through the end of May 1998, ISIS estimates:

· if, contrary to expectations, Pakistan has not resumed WGU production, the nuclear tests have lowered its WGU stockpile to 120 kg, roughly enough for 6 nuclear weapons. This provides a lower estimate of Pakistan's capability to produce nuclear weapons;

· if Pakistan resumed production of WGU in January 1998 using LEU feed, the nuclear tests have lowered its WGU stockpile to approximately 400 kg, roughly enough for 20 nuclear weapons;

· if Pakistan resumed production of WGU in March 1998, following India's national election, using LEU feed, the nuclear tests have lowered its WGU stockpile to approximately 335 kg, roughly enough for 16 nuclear weapons;

· if Pakistan resumed production of WGU in January 1998 using natural uranium feed, the nuclear tests have lowered its WGU stockpile to approximately 170 kg, roughly enough for 8 nuclear weapons; and

· if Pakistan produced WGU after 1991, then the nuclear tests have lowered its WGU stockpile to approximately 590 kg, roughly enough for 29 nuclear weapons. This provides an upper estimate of Pakistan's capability to produce nuclear weapons.

Best Estimate

Based on these scenarios, ISIS's estimates that Pakistan possesses sufficient WGU to assemble 6 to 29 nuclear weapons as of the end of May 1998. ISIS's best estimate is that Pakistan possesses 335 to 400 kg of WGU, sufficient to produce 16 to 20 nuclear weapons. This estimate assumes that Pakistan resumed WGU production at Kahuta since the beginning of the year.

Based on this best estimate, Pakistan's stockpile could grow to 600 kg by the end of 1998, enough for 30 weapons. This stock is nearly as large as if Pakistan had never halted WGU production in 1991. In the future, Pakistan's stock will grow at an annual rate of about 110 kg, or enough for 5 weapons per year.

# # #