6 ms·

I like the idea but a $0.10/kWh marginal cost is not the same as $0.10/kWh electricity. If they plan on extracting 6GW they'll need to put it somewhere, what's

by iaw 9y ago

I like the idea but a $0.10/kWh marginal cost is not the same as $0.10/kWh electricity. If they plan on extracting 6GW they'll need to put it somewhere, what's a people hub near Yellowstone? ... So they'll have to build out high capacity transmission lines which have associated costs for leases, construction, and maintenance. Then finally you get into the real kicker: cost of transmission. If only 5% of your transmitted energy actually makes it to your buyer, without congestion, you're spending $0.10 / kWh but your buyer is paying $2.00 / kWh where the difference all goes to line losses. Maybe building data centers nearby is a good way to mitigate the infrastructure problem. Edit: I also forgot to mention the time value of energy infrastructure. What is the expected time horizon to break even at target rates of return? How do those returns compare to other forms of capital investment? What's the risk profile look like for power producers that require heavy upfront investment? (hint: not good) Edit: Okay, the 5% figure was out of bounds, more realistic is 97% with proper build-out.

knodi123 9y ago
> Maybe building data centers nearby is a good way to mitigate the infrastructure problem. Haha, I'm picturing the pitch on their risk profile. "Yes, it's built on top of an active volcano. But hey, if it goes off, the entire american economy and a large part of the internet will be gutted. So it's unlikely to be your highest concern."
- awakeasleep 9y ago
  I'm under the impression that a yellowstone eruption would devestate the whole world's economy due to the release of particulates into the atmosphere & associated global cooling
  - njarboe 9y ago
    Yes, devastate the economy. As in, no crops could be grown for multiple years.
- Bluestrike2 9y ago
  That's actually a pretty nifty risk assessment. Your datacenter is toast if the supervolcano erupts. But if the supervolcano erupts, the bulk of the human race will starve to death in 74 days. So there's very little actual risk to your company from building nearby. You're dead, literally, regardless. Assessing such a project would probably drive an actuary insane.
  - make3 9y ago
    what are the chances of smaller, non death of all mankind leaks that would damage the data center? likely higher than elsewhere, no? so that wouldn't an be amazing risk profile
    - Bluestrike2 9y ago
      You wouldn't build a datacenter directly on top of the supervolcano. It's in a national park, after all. Any sort of relatively minor incidents from the geothermal plant wouldn't affect datacenters located outside the park (and they could be a few hundred miles away, if need be). The degree of devastation from an eruption would be so great that the normal risks associated with "hey, build sort of near a volcano" don't really apply in the traditional sense.
    - knodi123 9y ago
      > what are the chances of smaller, non death of all mankind leaks that would damage the data center? I think it depends on the type of volcano. There's some volcanoes that only go boom, and some that only go ooze, and some that are in the middle. And, IIRC, yellowstone is one that's big-boom or nuthin'. But it's been a while since sophomore geology class.
    - lostapathy 9y ago
      One way or another, Warren Buffett will figure out a way to sell insurance that covers the actual data center. You'd definitely want to be setup with multi-availability zones just in case it does go.
      - 9y ago
onion2k 9y ago
If only 5% of your transmitted energy actually makes it to your buyer.. Isn't the typical transmission and distribution loss around 8% in the US? Why would it be 95%?
- tdburn 9y ago
  Because SCIENCE!! ;)
xyzzyz 9y ago
> If they plan on extracting 6GW they'll need to put it somewhere, what's a people hub near Yellowstone? ... You can put it into aluminum production, for example, though at $0.10/kWh it's going to be an unprofitable business. > If only 5% of your transmitted energy actually makes it to your buyer, without congestion, you're spending $0.10 / kWh but your buyer is paying $2.00 / kWh where the difference all goes to line losses. You might get 95% loss if you wrapped your line around the globe, but for practical length of 1000 km, you can get as little as 3% with high voltage DC. Check out the Pacific DC Intertie, which is 50 years old instalation with over 800 miles of conductor spanning from Columbia River to Los Angeles.
- elihu 9y ago
  Presumably any process that can use the heat directly rather than converting to electricity would be more cost-effective, as you lose most of the energy in that conversion. I assume 350C isn't hot enough for aluminum which melts around 660C, but maybe there are other industrial processes that could take advantage of the heat. Maybe using the heat to kiln-dry wood or cure autoclaved aerated concrete blocks or something like that.
  - mirimir 9y ago
    In aluminum production, the electricity is used for electrolytic reduction, not melting. It's the waste heat from that which keeps the cell hot enough.
  - xyzzyz 9y ago
    > Presumably any process that can use the heat directly rather than converting to electricity would be more cost-effective, as you lose most of the energy in that conversion. All practical electricity generation technologies do it through heating water just the same, and they produce hot water as a waste product too. So, if your industrial process can use heat directly, it might as well use waste heat from existing power plants.
    - jopsen 9y ago
nullbyte 9y ago
Where did you get the idea that there would be a 95% energy loss rate? In general it never goes above 8-9%. High voltage + low amperage = high energy transfer efficiency.
tshadley 9y ago
> Maybe building data centers nearby Now THAT is interesting! Google is tying up hydroelectric dams with their centers currently, it seems quite feasible that massive geothermal energy could be put to work the same way. Deep-learning applications are seriously energy-hungry and they're going to need exponentially more in the future.
- maxander 9y ago
  > Deep-learning applications are seriously energy-hungry and they're going to need exponentially more in the future. So instead of eradicating humanity by erupting, Yellowstone will eradicate humanity by powering Skynet.
sna1l 9y ago
Maybe Yellowstone could become a large crypto miner :)
- nullbyte 9y ago
  I like this idea a lot
tiredwired 9y ago
Good spot for the new Amazon HQ.
wallace_f 9y ago
It is worth noting that the market price of fossil-fuel energy is a bad comparison because it does not include all costs associated with it, such as pollution, workplace deaths and war. According to Wikipedia, PV cells are 0.23/kWh.(1) 1-https://en.m.wikipedia.org/wiki/Cost_of_electricity_by_source https://en.m.wikipedia.org/wiki/Cost_of_electricity_by_sourc...

by iaw 9y ago

knodi123 9y ago

> Maybe building data centers nearby is a good way to mitigate the infrastructure problem. Haha, I'm picturing the pitch on their risk profile. "Yes, it's built on top of an active volcano. But hey, if it goes off, the entire american economy and a large part of the internet will be gutted. So it's unlikely to be your highest concern."

awakeasleep 9y ago
I'm under the impression that a yellowstone eruption would devestate the whole world's economy due to the release of particulates into the atmosphere & associated global cooling
- njarboe 9y ago
  Yes, devastate the economy. As in, no crops could be grown for multiple years.
Bluestrike2 9y ago
That's actually a pretty nifty risk assessment. Your datacenter is toast if the supervolcano erupts. But if the supervolcano erupts, the bulk of the human race will starve to death in 74 days. So there's very little actual risk to your company from building nearby. You're dead, literally, regardless. Assessing such a project would probably drive an actuary insane.
- make3 9y ago
  what are the chances of smaller, non death of all mankind leaks that would damage the data center? likely higher than elsewhere, no? so that wouldn't an be amazing risk profile
  - Bluestrike2 9y ago
    You wouldn't build a datacenter directly on top of the supervolcano. It's in a national park, after all. Any sort of relatively minor incidents from the geothermal plant wouldn't affect datacenters located outside the park (and they could be a few hundred miles away, if need be). The degree of devastation from an eruption would be so great that the normal risks associated with "hey, build sort of near a volcano" don't really apply in the traditional sense.
  - knodi123 9y ago
    > what are the chances of smaller, non death of all mankind leaks that would damage the data center? I think it depends on the type of volcano. There's some volcanoes that only go boom, and some that only go ooze, and some that are in the middle. And, IIRC, yellowstone is one that's big-boom or nuthin'. But it's been a while since sophomore geology class.
  - lostapathy 9y ago
    One way or another, Warren Buffett will figure out a way to sell insurance that covers the actual data center. You'd definitely want to be setup with multi-availability zones just in case it does go.
    - 9y ago

onion2k 9y ago

If only 5% of your transmitted energy actually makes it to your buyer.. Isn't the typical transmission and distribution loss around 8% in the US? Why would it be 95%?

tdburn 9y ago
Because SCIENCE!! ;)

xyzzyz 9y ago

> If they plan on extracting 6GW they'll need to put it somewhere, what's a people hub near Yellowstone? ... You can put it into aluminum production, for example, though at $0.10/kWh it's going to be an unprofitable business. > If only 5% of your transmitted energy actually makes it to your buyer, without congestion, you're spending $0.10 / kWh but your buyer is paying $2.00 / kWh where the difference all goes to line losses. You might get 95% loss if you wrapped your line around the globe, but for practical length of 1000 km, you can get as little as 3% with high voltage DC. Check out the Pacific DC Intertie, which is 50 years old instalation with over 800 miles of conductor spanning from Columbia River to Los Angeles.

elihu 9y ago
Presumably any process that can use the heat directly rather than converting to electricity would be more cost-effective, as you lose most of the energy in that conversion. I assume 350C isn't hot enough for aluminum which melts around 660C, but maybe there are other industrial processes that could take advantage of the heat. Maybe using the heat to kiln-dry wood or cure autoclaved aerated concrete blocks or something like that.
- mirimir 9y ago
  In aluminum production, the electricity is used for electrolytic reduction, not melting. It's the waste heat from that which keeps the cell hot enough.
- xyzzyz 9y ago
  > Presumably any process that can use the heat directly rather than converting to electricity would be more cost-effective, as you lose most of the energy in that conversion. All practical electricity generation technologies do it through heating water just the same, and they produce hot water as a waste product too. So, if your industrial process can use heat directly, it might as well use waste heat from existing power plants.
  - jopsen 9y ago

nullbyte 9y ago

Where did you get the idea that there would be a 95% energy loss rate? In general it never goes above 8-9%. High voltage + low amperage = high energy transfer efficiency.

tshadley 9y ago

> Maybe building data centers nearby Now THAT is interesting! Google is tying up hydroelectric dams with their centers currently, it seems quite feasible that massive geothermal energy could be put to work the same way. Deep-learning applications are seriously energy-hungry and they're going to need exponentially more in the future.

maxander 9y ago
> Deep-learning applications are seriously energy-hungry and they're going to need exponentially more in the future. So instead of eradicating humanity by erupting, Yellowstone will eradicate humanity by powering Skynet.

sna1l 9y ago

Maybe Yellowstone could become a large crypto miner :)

nullbyte 9y ago
I like this idea a lot

tiredwired 9y ago

Good spot for the new Amazon HQ.

wallace_f 9y ago

It is worth noting that the market price of fossil-fuel energy is a bad comparison because it does not include all costs associated with it, such as pollution, workplace deaths and war. According to Wikipedia, PV cells are 0.23/kWh.(1) 1-https://en.m.wikipedia.org/wiki/Cost_of_electricity_by_source https://en.m.wikipedia.org/wiki/Cost_of_electricity_by_sourc...