Ring the equator with solar panels and then run cables north/south to wherever it is needed, providing consistent supply 24/7. It works well enough in Dyson Sphere Program.
Only takes a few hundred billion USD of aluminium for that ring to have a cross section of a square meter, which is enough to lower the resistance to 1 Ω.
Of course, only China actually produces enough aluminium for this, and you'd want to divide it between several cables both for redundancy and because putting 2 TW though one cable at any plausible voltage will give it a surface magnetic field strong enough to prevent most tools working on it, and the political issues are going to be huge, but on paper it's fine.
There are many alternatives, but I have a 10kWh LiFePO4 pack that is sufficient for my house 95% of the time.
There have been great strides recently in battery recycling and new chemistries. Not to mention alternatives to batteries.
To me this seems like one of those problems that seems impossible until the economics start driving innovation. I think we are heading in the right direction.
I've never played DSP but I'm thinking of Factorio. Night falls equally across the world so you have to provide power from somewhere else at night or let your stuff go dark (which can actually be sane--I've put up radars with solar without any backup--it will show you the spread of the nests fine.) You have to start with burning wood or coal for power, solar comes before adequate storage. Thus if you go the solar route there's a range where you have solar, some storage and some steam. Unfortunately, while the game correctly prioritizes solar over steam it puts steam over storage. Thus the efficient approach is to put your steam behind a switch that keeps it isolated until your storage drops low enough. I've had a couple hundred megawatts flowing through one of the starter tech poles--it doesn't melt?!
Also, why, on earth, would anyone in Kenya need solar energy from Australia? It has some of the best wind resources on the planet in the Rift Valley alone. Not to mention its own ridiculously high levels of solar available. And all that before we've even talked about the expense and difficulty of laying such a line. Add in unparalleled mineral availability in Africa and the Chinese building out manufacturing in Kenya. (Heck, all over the EAC actually.) I just have a hard time even seeing storage batteries as much of a problem for Kenya in the near future.
You need places like Singapore. Where the resources are unequal to the need for this whole thing to make sense.
Was thinking more that Kenya and Australia would sell electricity to Singapore during Singapore's night.
But Kenya <-> Australia also works, for nighttime coverage.
Wind may be better for each though, I've not looked at that kind of specific.
I'd assume Kenya would also be great for geothermal, having been past (IIRC) Olkaria V in Hell's Gate.
2) That equation ignores the time value of money, which is absolutely relevant. A dollar spent today is not the same as a dollar saved 100 years from now.
And if you have skyscrapers you have lots of side surface area to capture solar energy as well
Some of the cables that start forest fires in CA have been there for a century by that point.
Even at 17e9 USD, if this lasts a century, and is in full use for that time, it's adding 1.1¢/kWh to the price of delivery:
http://www.wolframalpha.com/input/?i=1.75%20GW%2A100%20years...
I could easily imagine them doing this, and another similar one to India or twice as far to Kenya, and getting a decent amount of PV at night as a consequence.
(OTOH, as I'm not a grid or marine engineer, I am ready to be mocked for suggesting an underwater cable may last a century).