All 4.5 billion tons can't be extratcted. More we extract, concentration decreases and harder it gets. I keep asking this question: If seawater extraction of metals is practical, why aren't we extracting other costly metals now? https://twitter.com/AchalHP/status/1011661441412337665
https://inis.iaea.org/collection/NCLCollectionStore/_Public/...
If extracting gold is only 300 times harder, that puts the lower bound cost at $108,000 per kilogram of gold. That's significantly more expensive than the market price of gold.
Indeed, extracting uranium itself from seawater is not cost competitive with conventional terrestrial mining at present. But the process has been demonstrated on a technical level. Either terrestrial uranium deposits will have to get closer to exhaustion or seawater extraction will have to be much more cost optimized (or both) before seawater extraction of uranium is economically competitive. I was only addressing your technical claims about the exhaustion of seawater uranium, not making economic claims.
Inefficient technologies will hit some constraints if not others.
We may run out of time: A pressure vessel production cycle is over one year[1]. By the time we reach >1000GW someone may make fusion practical or China's competent nuclear industry may start selling MSRs to everyone.
We may run out of highly skilled people required to build and operate these complex machines. The safety is highly dependent on operations and management throughout the life cycle of complex machines. From manufacturing to operations there is tiny margins for error. (Not a physical constraint.)
LWR technology is associated with military ships and submarines, no one shares it. Few countries control the technology and there are only a few places in the world to make pressure vessels. (Not a physical constraint.)
We may run out of waste disposal sites. A geological repository needs special rock formations away from earthquake zones.
The article is written by a nuclear industry professional who tells that routine things done by the nuclear industry today when applied to MSR is a "disadvantage". They open the lid of an LWR every 18 months and replace/shuffle highly radioactive fuel assemblies and say that it is a disadvantage for them to replace graphite. That is totally unfair.
[1] VVER pressure vessel production cycle needs 3 years. https://www.youtube.com/watch?v=91yVhrSZ5jQ
NuScale pressure vessel R&D: https://www.energy.gov/sites/prod/files/2016/10/f34/02%20-%2...
TL;DR Even small pressure vessels (used by nuscale) have lead times of about a year. A megafactory can build like 1 or 2 vessel per month. At 1.5GW/year/factory, we need like 22 factories to keep up with the production schedule if we need to reach 1000 GW of small-LWR capacity within, say 2050.
Molten salt is one of about a dozen advanced reactor techs that has huge potential.
Hard part is economics. Hazardous coolant has been a pain to maintain cheaply so far.
Also seawater uranium replenishes from erosion and plate tectonics so it will effectively never decrease in concentration, even if we pull it out at world scale.
http://large.stanford.edu/publications/coal/references/docs/...
Once we setup a pebble/advanced fuel making factory, closing it will takes decades (because people may lose jobs) and the new solid-fuel factory will again pause nuclear innovation for another 100 years. The only way to continuously improve nuclear reactors is to go fluid-fuel. No engineered fuel, so no job loses. Reactor innovation is independent of fuel factory. Nuclear fuel becomes a commodity instead of engineered speciality. Example: MSRE ran U235 and U233 without any modification.
Secondly, solid fuel reactors throw away fuel along with heat exchange surfaces (clad). Fuel also undergoes crystal structure degradation along with other solid structures. Maybe there is enough fuel in the seawater, but there may not be enough places suitable for geological repositories.
Solid-fuel reactors always need excess reactivity reserve. Always needs control rods, and if someone (or a bad actor) pulls all the control rods, reactor gets supercritical. Needs highly skilled people and needs security.
For emergency shutdown of solid-fuel reactor, poison is added to coolant, not fuel. In an emergency, poison is added to the liquid-fuel, permanently destroying the fuel. Emergency can be anything, from a natural disaster to terrorist attack. Fluid-fuel reactors offer unbeatable safety features against anything.
Fusion reactor runs hotter and again fuel is in fluid state.(Fluid: Liquid, gas & plasma.). But only runs for 10-100 minutes. Demonstrated fission reactors run for thousands of hours continuously.
We either stop using those transport modes, or revert to prior forms (wind), or wire-based electrical systems (rail, canals, trucks).
A Google X Project tried and failed.
That can't be right. About 200 tonnes of natural uranium is needed to produce 1 GWe per year in conventional reactors [1]. That's 3,200,000 tonnes per year if you mean 16000 GW in the form of electricity, or closer to 1 million tonnes per year if you're referring to primary (thermal) energy. Seawater contains about 4.5 billion tonnes of uranium [2]. That's well over a thousand years' worth of uranium, either way.
[1] https://www.world-nuclear.org/information-library/nuclear-fu...
[2] http://large.stanford.edu/courses/2012/ph241/ferguson2/