Thread by @james_acton32

<THREAD>Concern about China's nuclear build-up (see below) is partly premised on the fear that Beijing seeks to challenge the United States numerically.

This misses a key point: China lacks the fissile material to do so. (1/n)

nytimes.com/2020/07/29/opi…

In 2019, @DefenseIntel Director Ashley estimated that China’s warhead stockpile was in "the low couple of hundreds." The U.S. has ~3,800.

So, how many nuclear warheads could China build given its fissile material stockpile? (2/n)

fas.org/security/2019/…

First off, here's a *worst-case* scenario, i.e. Let's make assumptions tend to underestimate the amount used in each weapon and overestimate the amount available. (3/n)

In 2017, Hui Zhang estimated that China had produced 2.9 +/- 0.6 tonnes of Pu. Note this is *higher* than a leaked DoE estimate from 1999 after China is believed to have ceased Pu production. (4/n)

fissilematerials.org/library/rr17.p…

In 2011, Zhang pegged China's HEU stockpile at 16 +/- 4 tonnes. He later revised this downwards, but in the sprit of worst case thinking, let's use the 2011 estimate. (5/n)

tandfonline.com/doi/abs/10.108…

Finally, let's assume that each Chinese warhead is thermonuclear and contains 5 +/-2 kg of Pu and 20 +/- 5 kg of HEU (which allows for the possibility of modern efficient primaries)... (6/n)

When you run the numbers, you find that China's fissile material stockpile could yield an arsenal of about 840 weapons as an *upper bound*.

Remember, as a comparison, that the US has ~3,800 warheads. (7/n)

Also, these assumptions suggest that the availability of Pu (not HEU) is the factor limiting China's nuclear arsenal size. That's significant because Pu production is much easier to detect than HEU production. (8/n)

Second, let's do a *best estimate* for how many warheads China could produce. We'll use:

--The leaked DoE estimate for plutonium: 1.7-2.8 tonnes.
--Zhang's revised (2017) HEU estimate: 14 +/- 3 tonnes.
--The same assumptions about warheads.

(9/n)

The result...

A best estimate for the number of warheads that could produce is about 450.

(Once again, arsenal size is limited by plutonium availability.)

(10/n)

So, bottom line, it appears that China simply doesn't have enough fissile material to challenge the United States numerically.

But, let's be critical and critique these calculations a bit... (11/n)

There are significant uncertainties in fissile material stockpiles and warhead designs. I've tried to take them into account, but it's difficult. That said, my calculations would have to be *really* wrong to change the basic conclusion--and I doubt they're so far off. (12/n)

Also, I'm assuming that China has ceased fissile material production, as is widely reported. This conclusion seems pretty robust for Pu production, which is relatively easy to detect. (13/n)

It's less robust for HEU production, which is harder to detect. BUT, it appears that Pu availability limits the size of China's arsenal. So, unless China still uses all HEU pits (unlikely), producing more HEU by itself wouldn't change my conclusion. (14/n)

Finally, and this is the big critique, China may restart fissile material production--particularly of Pu.

This would be slow and visible but, if happens, it'd be a huge deal. And, I certainly can't exclude this possibility. (15/n)

(Indeed, China has ambitious plans for civil reprocessing--though, rightly or wrongly, the U.S. hasn't (to date at least) been all that concerned that civil plutonium would be used for weapons purposes.) (16/n)

So, where does this leave us?

If the U.S. is concerned about a Chinese nuclear build-up, fissile material seems the place to start. Specifically, China and the U.S. could work bilaterally on building confidence that they have ceased fissile material production. (17/n)

To be sure, I wouldn't expect China to jump at this. It's still sensitive for Beijing. But it's less sensitive than trying to negotiate limits on weapons. (18/n)

As you may have guessed already, we'll have a publication that develops this idea later in the year. I also want to acknowledge @fissilematerial, which is an amazing resource for this kind of analysis. (19/19)