According to Congressional testimony from a committee they created to study the problem, a SINGLE nuclear bomb set off to maximize the EMP effect would likely kill about 2/3 of the population of the Continental USA, with possible higher estimates up to 90%. Mostly through slow painful deaths like mass starvation.

Why?

Because it would knock out most of our cars, electronics, payment systems, farm machinery, and everything else that we depend on to keep the supply chains our grocery stores depend on for food distribution.

See https://nationalinterest.org/blog/buzz/millions-americans-co... for verification.

I have no idea where your "worst case scenarios" come from. But they don't exactly seem "worst case" to me!

Those estimates where not from a single nuke. “If the lights stay off for more than a year in this country, the Commission's estimate was the loss of life would run into the tens of millions, perhaps a great deal more.”

Nationwide long term total blackout is unlikely for many reasons.

I think you underestimate the time needed to restore the entire grid if it were to be simultaneously destroyed.

Sure, but the estimate there isn't "we don't restore the grid within a year" it's "the entire grid stays off for more than a year" (the whole thing is completely unrecoverable). As far as I know, your average transmission line or transformer isn't going to be broken, so the issue would be primarily power production and things that use electricity (computers etc.). We wouldn't need to restring powerlines except in places actually hit by blast waves, and the grid wouldn't be "destroyed".

(and it should be further noted that this conclusion was that the US carrying capacity for life would decrease to 30m sustained by like subsistence agriculture. I expect that those deaths of hundreds of millions would be over decades, not quickly, and I really really doubt that we'd be unable to recover more than some of our agricultural capacity in a decade)

There is an open debate about the likely damage done by power surges along long-distance power lines. Industry funded research claims that they should survive both EMP and major solar storms. Real world experience suggests that they are overconfident.

See https://chris-said.io/2020/06/18/everything-ive-learned-abou...

Regardless of which side you believe, it is clear that any "worst case scenario" should include the possibility of an EMP attack that actually does take out the power grid.

That blog post is making a fundamental error, taking down the power grid isn’t the same as taking it out.

Things are setup to fail safe which preserves equipment but defaults to turning everything off. Getting significant parts of the grid up again should take days even if your stuck with rolling blackouts for years.

Further, EMP isn’t a magic wand that instantly destroys all electronics, there is a huge range of sensitivity not just based on type of equipment but also if it’s running and the orientation it’s in. Taking out farm equipment that’s off is really difficult. The actual nukes are a much larger threat.

I hope you're right.

They are enough slbm armed subs stalking the ocean that a clean first strike is impossible. Even if you decapitate the enemy’s command and control the subs still rise and launch a full retaliatory strike.

My numbers included the arsenals of all countries, and casualties for all reasons, including the direct blast, fallout, breakdown of infrastructure, the possibility of a nuclear winter, reduced food production, etc.

Even people in areas with no nukes, little fallout and not affected by nuclear winter could face mass starvation due to lack of fertilizers, diesel, spare parts for machines, etc.

Even in North America, Europe and North Asia, having the electric grids knocked out, central administrations crippled, much industry and agricultural infrastructure knocked out, followed by nuclear winter might see more people die from cold and starvation than from the nukes themselves.

Anyway, I don't think we really want to test this, even if the total numbers do not fully add up to complete destruction.

Exactly. I bet most of us on HN live in large cities and will be killed in the first strike.

I live in a large city, but have reasonable hope not to be targeted in a first strike by either side. (I live in Singapore.)

Why would only 20-40M die after 1000 nukes?

To be glib: Hopium.

To be frank: That number's missing a zero.

Not two zeros?