I mean, the two weeks of darkness is the real killer.
I mean look, let’s say our datacenter requires 100 kw of power. That means about an 800 square m solar array. If you need to store enough power to run the system through a two week night, you’ll need a lot of batteries and a much larger array. You’ll only be collecting power half the time, so the array needs to be twice as large. But there are about 30% losses in transferring energy to the battery, drain over time, and then transferring back, so some quick math says you now need somewhere around 230kw, for a 1800 square m solar array.
Currently, an array large enough to generate 100kw weighs between 1-2 tons. If we’re optimistic and say 1 ton, that’s 2.25 tons for our equivalent moon array
Now the batteries. The capacity needs to be enough to store 130kw over the course of two weeks. 130kw x 24hours x 14days = 43,680 kw hours of storage. Lithium ion batteries weigh about 5kg per kwh of storage. 43,680 kwh of storage would weigh 218 tons. And that’s just the base weight of the batteries, I didn’t factor in whatever structures or systems the batteries would need to be usable in space.
Here’s where I point out that it costs about $5 mil per ton sent to orbit. It costs somewhere between $100 mil and $1 bil to send 1 ton to the moon. Personally, I think it’s a lot easier to send 1 ton to low earth orbit, than to send 220 tons all the way to the moon. I think the only way to make data centers on the moon make sense, would be if you’re already living and working on the moon in a functioning settlement.
As crazy as orbital data centers are, if your counting in dollars, lunar data centers are literally 10,000 times crazier.
First things first, you aren’t shielding solar panels from radiation, you’re shielding the computing components from the radiations. I don’t know how much you know about microprocessors and silicon engineering, but I can promise you they do NOT like ionizing radiation or most other high energy sources of EM radiation, especially while energized. In space, this is a lot of heavy plating that is effectively dead weight and at best can be used as a secondary heat sink, but that’s highly unlikely since it would also be absorbing heat energy while in sunlight.
Additionally, as I pointed out, you’d want to or at least be highly encouraged to pair these lunar data centers with long term colonies, at least as long term as you’d want the data center active. Any unmanned data center is going to fail within 6 months, be it in space, on the moon or some remote valley here on Earth. And while it’s not cheap to sustain life on the moon, it’s even less cheap to do so in orbit when you examine the effects of micro gravity on the human body as well as what it would cost to build a suitable shelter in orbit. There is a reason scientists talk about building long term colonization projects on the moon or Mars, not just putting up orbital space habitats like some gundam series. Earth Orbit may be closer but it’s probably the least hospitable long term to any form of human habitation.
I’m not sure the moon makes a lot of sense.
I mean, the two weeks of darkness is the real killer.
I mean look, let’s say our datacenter requires 100 kw of power. That means about an 800 square m solar array. If you need to store enough power to run the system through a two week night, you’ll need a lot of batteries and a much larger array. You’ll only be collecting power half the time, so the array needs to be twice as large. But there are about 30% losses in transferring energy to the battery, drain over time, and then transferring back, so some quick math says you now need somewhere around 230kw, for a 1800 square m solar array.
Currently, an array large enough to generate 100kw weighs between 1-2 tons. If we’re optimistic and say 1 ton, that’s 2.25 tons for our equivalent moon array
Now the batteries. The capacity needs to be enough to store 130kw over the course of two weeks. 130kw x 24hours x 14days = 43,680 kw hours of storage. Lithium ion batteries weigh about 5kg per kwh of storage. 43,680 kwh of storage would weigh 218 tons. And that’s just the base weight of the batteries, I didn’t factor in whatever structures or systems the batteries would need to be usable in space.
Here’s where I point out that it costs about $5 mil per ton sent to orbit. It costs somewhere between $100 mil and $1 bil to send 1 ton to the moon. Personally, I think it’s a lot easier to send 1 ton to low earth orbit, than to send 220 tons all the way to the moon. I think the only way to make data centers on the moon make sense, would be if you’re already living and working on the moon in a functioning settlement.
As crazy as orbital data centers are, if your counting in dollars, lunar data centers are literally 10,000 times crazier.
Oh man so much to unpack here.
First things first, you aren’t shielding solar panels from radiation, you’re shielding the computing components from the radiations. I don’t know how much you know about microprocessors and silicon engineering, but I can promise you they do NOT like ionizing radiation or most other high energy sources of EM radiation, especially while energized. In space, this is a lot of heavy plating that is effectively dead weight and at best can be used as a secondary heat sink, but that’s highly unlikely since it would also be absorbing heat energy while in sunlight.
Additionally, as I pointed out, you’d want to or at least be highly encouraged to pair these lunar data centers with long term colonies, at least as long term as you’d want the data center active. Any unmanned data center is going to fail within 6 months, be it in space, on the moon or some remote valley here on Earth. And while it’s not cheap to sustain life on the moon, it’s even less cheap to do so in orbit when you examine the effects of micro gravity on the human body as well as what it would cost to build a suitable shelter in orbit. There is a reason scientists talk about building long term colonization projects on the moon or Mars, not just putting up orbital space habitats like some gundam series. Earth Orbit may be closer but it’s probably the least hospitable long term to any form of human habitation.