How many nuclear weapons can India make with its existing fissile material stockpile? Recently, two different sources have produced wildly divergent estimates. In September 2015, the Pakistani newspaper Dawn reported that India possessed “enough fissile material … for more than 2,000 warheads.” In contrast, a report released by the Institute for Science and International Security (ISIS) in November 2015 concluded that India’s stockpile of fissile material was only sufficient to make approximately 100 nuclear weapons. What accounts for the order-of-magnitude difference between these estimates?

Estimates of fissile material have significant real-world policy implications. Pakistan, for example, seems to base requirements for its own nuclear weapons program in no small part on the projected size and composition of India’s nuclear arsenal, as well as perceptions of Indian conventional military superiority. At a time when both India and Pakistan seem increasingly at risk of sliding into an arms race, in spite of their efforts and protestations to the contrary, inflated nuclear weapons projections run the risk of inflaming the public discourse and heightening this competition unnecessarily.

Estimating stockpiles. The estimate reported by Dawn that India has enough fissile material to produce 2,000 nuclear weapons can be traced back to a 2014 assessment by Mansoor Ahmed, a Pakistani nuclear analyst. He estimated that at the end of 2013, India’s fissile material stockpile included 800 to 1,000 kilograms of weapons-grade plutonium, 2 metric tons of highly enriched uranium (HEU), and 15 metric tons of reactor-grade plutonium. Assuming that 4 kilograms of weapons-grade plutonium, 50 kilograms of HEU, or 8 kilograms of reactor-grade plutonium would be necessary to make one nuclear warhead of each type, Ahmed estimated that India could produce 250 warheads from weapons-grade plutonium, 40 from HEU (gun-type implosion devices, not thermonuclear weapons), and 1,875 from reactor-grade plutonium—for a total arsenal of 2,165 nuclear weapons.

In contrast, the recent report from ISIS concluded that at the end of 2014, India likely possessed about 550 kilograms of weapons-grade plutonium, 100 to 200 kilograms of HEU intended for use in thermonuclear weapons, and 2.9 metric tons of separated reactor-grade plutonium. The study assessed that this fissile material was sufficient to produce about 75 to 125 nuclear warheads, with 100 nuclear weapons as the median estimate. ISIS arrived at this number mainly through an appraisal of India’s weapons-grade plutonium stockpile; the authors assume that India would not use reactor-grade plutonium in nuclear warheads, and that HEU would only be used to produce a handful of thermonuclear weapons at most. ISIS also considered that some plutonium is in weapons production pipelines or held in reserve, meaning that only about 70 percent of India’s stockpile is available to be made into weapons. Consequently, assuming that it would take 3 to 5 kilograms of weapons-grade plutonium for each warhead, ISIS calculations yielded an arsenal that could range from 75 to 125 nuclear weapons.

Reactor-grade plutonium. The biggest difference between these two estimates comes from their assessments of, and assumptions about, reactor-grade plutonium. Not only does the ISIS study discount the possibility that India would use reactor-grade plutonium in its nuclear weapons, but its estimate of India’s reactor-grade plutonium stockpile is also significantly lower than Ahmed’s: 2.9 metric tons as opposed to 15 metric tons.

Interestingly, Ahmed himself has given much lower estimates of India’s reactor-grade plutonium stockpile in other instances. In a post on the Stimson Center’s South Asian Voices blog, also from 2014, Ahmed cited India’s reactor-grade plutonium stockpile as just 5 metric tons rather than the 15 metric tons that he posits in the Defense News article. In a private email communication with me, Ahmed explained that this disparity is due to the distinction between separated plutonium and plutonium found in spent fuel. Ahmed clarified that he estimates India possesses 5 metric tons of reactor-grade plutonium that has already been separated, and an additional 10 metric tons in spent fuel that has not yet been separated. The estimate of 2,000 nuclear warheads reported by Dawn included both of these types of reactor-grade plutonium, whereas Ahmed’s lower estimate included only the 5 metric tons of plutonium that he estimates has already been separated.

Other estimates of fissile material stockpiles typically do not include plutonium in spent fuel that has not been reprocessed, for the good reason that such fissile material is not available for use in nuclear weapons. Reprocessing is complicated and expensive, and India in particular has historically had trouble achieving consistent operations in its reprocessing facilities. One of the most highly respected sources on fissile material stockpiles, the International Panel on Fissile Materials (IPFM), recently released a report that sheds useful light on this question. In its Global Fissile Material Report 2015, the IPFM does not include unseparated plutonium as part of its estimate of India’s fissile material stockpile, citing the historically poor performance of India’s reprocessing plants at Tarapur and Kalpakkam.

The report notes that India’s reactor-grade plutonium stockpile is most likely intended as fuel for the country’s Prototype Fast Breeder Reactor—not for nuclear weapons, as Ahmed assumes—but that the reactor’s start date has been pushed back several times. This is likely due to difficulty that India has reportedly experienced in separating sufficient plutonium to fuel the reactor. Considering this historically low rate of separation and the problems it has caused for India’s fast breeder reactor program, it seems difficult to imagine that the country’s reprocessing plants will support a future sprint to rapidly separate its remaining stockpile of reactor-grade plutonium from the spent fuel and turn that fissile material into an exponentially larger nuclear arsenal.

Assessing motivation. Setting aside the question of capability, would India decide to make nuclear weapons from reactor-grade plutonium? As the name would suggest, reactor-grade plutonium is not as suitable for nuclear warheads as weapons-grade plutonium. Weapons-grade plutonium is irradiated for a shorter period of time in order to maximize the proportion of the more desirable plutonium 239 isotope. In contrast, reactor-grade plutonium is irradiated much longer to maximize its energy potential, and consequently contains a lower level of plutonium 239 and a higher concentration of plutonium 240. Weapons made from material that contains a higher amount of plutonium 240 are much more likely to fizzle (to produce a much smaller explosive yield than expected) and require a larger amount of fissile material for critical mass. Although it is possible to make nuclear warheads from reactor-grade plutonium, experts consider it more complicated and risky than using weapons-grade plutonium.

Nuclear weapons tend to be viewed in India more as political symbols than as usable weapons, and they occupy a less salient place in India’s national security strategy than is the case for many other nuclear weapons states. Given that, it seems unlikely that India’s leaders would feel the need to use reactor-grade plutonium that is otherwise intended for fast breeder reactors in order to make lower-quality warheads when they already have the capacity to make 100 or more from superior weapons-grade plutonium alone. India has committed to a doctrine of both “No First Use” and of credible minimum deterrence, and accordingly seems to place much more importance on developing a secure second-strike capability than on the size of its arsenal.

India values its international reputation surrounding nuclear weapons—it has reaped dividends from being perceived internationally as a responsible nuclear power. India’s leaders seem unlikely to risk this carefully maintained image by engaging in a rapid nuclear weapons build-up that might alarm the international community, particularly when substantially greater numbers of warheads are not viewed in India as strategically necessary or even beneficial. In light of these attitudes, it seems doubtful that New Delhi would feel the need to manufacture large numbers of weapons using reactor-grade plutonium.

Revising the math. What is the most plausible estimate of India’s fissile material stockpile and the number of nuclear weapons that it could build? After discounting unseparated plutonium in spent fuel as a source of proliferation in the near future, even Ahmed’s generous estimate from the South Asian Voices blog post is only equivalent to a potential Indian arsenal of 844 nuclear warheads—a significant number to be sure, but nowhere near 2,000 weapons.

If one discounts reactor-grade plutonium entirely, that estimate drops even further to an arsenal of just 219 weapons. In addition, it is likely that much, or even most, of India’s HEU is intended for use in naval reactors rather than in nuclear warheads. It is also clear that some of India’s weapons-grade plutonium was already used in nuclear tests or is contained in process waste. Taking into account those factors, the estimate quickly begins to drop to something much more along the lines of the ISIS estimate of roughly 100 nuclear warheads. This estimate is in the same ballpark as the September 2015 estimate by Hans M. Kristensen and Robert S. Norris in the Bulletin, of 110 to 120 nuclear warheads.

The implications of estimates. Fissile material estimation, particularly when based on open-source information, is an inexact science. The uncertainty in such estimates must be properly contextualized in order to make sound projections of an adversary’s future arsenal. Because nuclear arsenals are the result of political decisions as well as scientific ones, it is important to consider a country’s strategic calculus rather than focusing on technical capabilities alone.

Much like the imaginary missile gap fueled public fears and heightened the arms race between the United States and the Soviet Union during the Cold War, improperly contextualized estimates of India’s fissile material stockpile risk skewing the public discourse and pushing South Asia toward a competition that both countries wish to avoid. The stakes are high: nothing less than a nuclear arms race on the subcontinent could hang in the balance.

This article originally appeared in Bulletin of the Atomic Scientists