Cost on its own is rarely an issue. The thing that matters is payback time.

If one container move generates $10, and the drone moves 12 containers per hour, 12 hours a day, it will only take 6 months to earn the quarter of million mentioned.

Also, it might actually sound like a very cheap option compared to the ship-to-shore container cranes, which cost $20M-$40M each: https://www.freightcourse.com/port-cranes/

As for recharging: if our power train is 28 MW, why not charging it at the same speed as we discharge it? Why settle on 500 kW, which is really low? For comparison, Tesla supercharger is up to 300 kW, and a car is a much smaller EV.

25 MW charging sounds great, but that does restrict you to 4 drones charging at a time per nuclear reactor (comparing to reactors used on carriers), which might be a little bit limiting (not to mention that making a battery that can be charged this quickly without overheating would be a triumph of engineering). Also your example of 12 containers per hour 12 hours a day for 6 months would be 26000 recharges, which is roughly 3x as long as the battery will last given current battery technology. In other words, this workload would require spending somewhere around $1M per year in batteries for a drone that takes roughly 12 minutes to unload a container. For comparison, a port crane, can unload 30 to 50 containers an hour, and will use far less electricity since it is lifting 1/3rd the weight (and isn't hovering).

Yes, the battery lifespan is the big issue (as pointed out in my original post).

As for charging at 28 MW: might work, given that the battery is physically larger, so there's more space to dissipate heat. I agree that it's not proven by the napkin math that it's doable.

What I really suspect is that if Los Angeles refuses to accept nuclear container ships, Mexico would let them in, and then the regular railway will be used to move them to the US. A much simpler solution. :)