Underwater Data Centers

exclaimedleech8

ACCESS: Top Secret
Joined
30 September 2024
Messages
823
Reaction score
999
Might be plausible if ocean currents power them. Might be less damaging to place them near the largest rivers but seems destructive to living ecosystems. It has the potential to pool warmed water in the deeper water bodies whereas in a river it flows with the gravity fed current. There aren't a lot of easily accessible shallow cold currents near the northern population centers.
 
Might be plausible if ocean currents power them. Might be less damaging to place them near the largest rivers but seems destructive to living ecosystems. It has the potential to pool warmed water in the deeper water bodies whereas in a river it flows with the gravity fed current. There aren't a lot of easily accessible shallow cold currents near the northern population centers.
I wonder if you could build them next to offshore drilling rigs.
 
If you're doing X amount of calculations, it's going to generate Y amount of heat wherever it is. The core problem with datacentre proliferation is increased energy consumption overall. Space-based datacentres get that energy use offworld, oceanic ones don't.

With the weakening of the Atlantic Meridional Overturning Circulation, pumping heat into the deep ocean is not a good idea.
 
If you're doing X amount of calculations, it's going to generate Y amount of heat wherever it is. The core problem with datacentre proliferation is increased energy consumption overall. Space-based datacentres get that energy use offworld, oceanic ones don't.

With the weakening of the Atlantic Meridional Overturning Circulation, pumping heat into the deep ocean is not a good idea.
You think this could have a serious environmental impact?
 
You think this could have a serious environmental impact?
The current forecasts for the impact of datacentre growth on overall energy consumption suggest it will be between 3 and 9% of global energy consumption by 2030, so yes. Putting it in the deep ocean is probably the worst place I can think of.
 
If you're doing X amount of calculations, it's going to generate Y amount of heat wherever it is. The core problem with datacentre proliferation is increased energy consumption overall. Space-based datacentres get that energy use offworld, oceanic ones don't.

With the weakening of the Atlantic Meridional Overturning Circulation, pumping heat into the deep ocean is not a good idea.
It makes dealing with waste heat much, much simpler. And no, it is highly unlikely that the heat produced will be enough to change the earth's climate.
 
Thinking of each site as the problem is disingenuous. Thinking of how many would be involved will clarify the problem. Satellite use makes more sense imho and could perhaps be done in association with solar and laser tech to take energy production off planet also.
 
It makes dealing with waste heat much, much simpler. And no, it is highly unlikely that the heat produced will be enough to change the earth's climate.
Now Anthropogenic Climate Change is primarily directly manifested through atmospheric composition changes, but heat plays a major part in secondary effects - the ones we notice - via more energetic weather systems and melting of the ice sheets. Ocean current changes are a major driver of weather systems (NOAA just announced we're in an El Nino event, for instance), and we're seeing worrying signs of instability in some of the major currents - AMOC particularly. If we lose AMOC, the forecast is Europe's climate drops 8C, US sea levels rise, and god knows what other currents will shut down or shift as a response. (Remember, London and Berlin are on the same latitude as Calgary, without AMOC our climate turns Canadian).

The one place you really, really don't want to be dumping heat into AMOC is in deep water off Northern Europe where the surface current releases heat to the atmosphere and the cold water sinks and turns south west back to the US. Doubly so when it's already weakening. Where did MS go for their trial? Where has the existing marine infrastructure for this kind of mass offshore engineering?

One trial won't make a difference, but if a significant percentage of datacentres shift offshore? Remember, we're not just talking about the effects of one year's release of heat, we're talking decades. And we're talking about that in combination with other effects from the related power generation, which is often opting for new installations of cheap and dirty technologies - cf X's Memphis datacentres.

Maybe it won't be an issue, but it's a chaotic system that's in a period of instability. That's not somewhere to throw in a new input without researching the consequences.
 
Microsoft went to the Pacific. Also, something of interest to me is that if we’re generating a bunch of power offshore, we can also use it to bring up cool water from the depths and lower surface temperatures.
 
Humans already create lots of waste heat with their activities and the impact on weather patterns is negligible. The heat absorbed emitted by asphalt roads and parking lots is much greater than that dumped out the exhaust and brake pads of the cars that use them.
 
Microsoft went to the Pacific.
"An experimental, shipping-container-size prototype is processing workloads on the seafloor near Scotland’s Orkney Islands, Microsoft announced today.
The deployment of the Northern Isles datacenter at the European Marine Energy Centre ... "
 
That will eventually heat the entire water column.

Bad idea.
It would take centuries, and we’ll have other options in the interim. The ocean is an enormous heat sink.
"An experimental, shipping-container-size prototype is processing workloads on the seafloor near Scotland’s Orkney Islands, Microsoft announced today.
The deployment of the Northern Isles datacenter at the European Marine Energy Centre ...
Thanks. I’m thinking of something else then.
 
would take centuries
Temporary measures have a nasty habit of becoming permanent. In shallow water, the effect is immediate. Interpolate for depths in between.
The one place you really, really don't want to be dumping heat into AMOC is in deep water off Northern Europe where the surface current releases heat to the atmosphere and the cold water sinks and turns south west back to the US. Doubly so when it's already weakening.
^ This.
I am at an age where I will probably escape the worst consequences of terminal optimism.
 
Temporary measures have a nasty habit of becoming permanent. In shallow water, the effect is immediate. Interpolate for depths in between.

^ This.
I am at an age where I will probably escape the worst consequences of terminal optimism.
As energy use increasingly shifts offworld, I think you’ll have far less to worry about than you believe. You should also fear the consequences of terminal pessimism.
 
That will eventually heat the entire water column.

Bad idea.

Well said
https://phys.org/news/2026-06-world-important-climate-threats-image.html
https://arstechnica.com/science/202...min-backs-off-plans-to-kill-ocean-monitoring/

Warming isn't the only threat. Acidification, the threat of leaks...internals poisoning the water...electrocution hazards, etc.

This and aerosol injection type geoengineering are both to be banned.
https://phys.org/news/2026-06-atmospheric-cooling-effects-stratospheric-aerosol.html

Sunshades, data centers and powersats share similarities and savings can be had.

I remember a book by Heppenheimer where large solar electric tugs can move payloads away from Earth farther into space.

I would also like to see all warheads on Luna, so all these plans complement one another.
 
Last edited:
I’ve read that book. It’s a good one. Realism is not pessimism though.

I think sea colonies, like space colonies, are only a matter of time, and it may be they’ll spring up because of seaborne data centers. There’s quite a lot we can do to benefit the environment and ourselves at the same time, if we have the brains to do so.
 
This is getting way off topic, but I’ll reply one more time nonetheless.
The ocean is an enormous battery powering weather events such as hurricanes and is already showing the effects of heating. How much more powerful would you like your hurricanes to be?
You must have a good physics and mathematical background. Care to calculate what trillions of gallons of cold water does to surface temperatures?
Redundancy? Who needs it? Everything will work fine!

Safety regs are what happens when the optimists realise the people they were calling pessimists are actually realists.
It’s telling that you have no idea what I think about redundancy or safety and jump straight to what you feel I must think. It’s almost as if you’re not here to discuss, you want to feel smart. To give you a hint: I deal with safety systems at least weekly. They are important and necessary. The pessimists aren’t the ones pushing for safety regulations. They’re the ones insisting on things like ships staying in harbor because the ocean is scary, and that safety is a binary. The realists acknowledge there are risks but aren’t ruled by fear.
Nice try. But you don’t understand me either. You can further deteriorate the discussion so you can feel smart, or you can discuss floating data centers as the OP posted about. Which do you prefer?
 
The pessimists aren’t the ones pushing for safety regulations. They’re the ones insisting on things like ships staying in harbor because the ocean is scary, and that safety is a binary. The realists acknowledge there are risks but aren’t ruled by fear.
You are talking extremes.

I grew up on the North Sea shore, when I was a teenager fishermen drowning was so commonplace that the local rag never once made it a front page item.
Didn't stop me from going to sea, donning every lifejacket offered.
I will never be as rich as Stockton Rush. As cautious as I am, I might fall victim to ordinary bad luck. I have had a few narrow scrapes on my motorbike, some of them in no way my fault. I have never had sleepless nights over that.

Humanity is deliberately meddling with systems humanity doesn't understand. Systems that, if upended, will cause us great harm.

Because of greed, stupidity, ignorance, shortsightedness. And optimism that a technological solution is just around the corner. In some cases, the knowledge is readily available in the public domain, but people look the other way, fingers in their ears.

You may call me a pessimist.
“Two things are infinite, as far as we know – the universe and human stupidity.” - Frederick S. Perls
Then Einstein proved that the universe is, in fact, finite.
 
Last edited:
You are talking extremes.
I’m responding in kind to people who, at the slightest hint of us trying something, offer apocalyptic visions of the future.
I grew up on the North Sea shore, when I was a teenager fishermen drowning was so commonplace that the local rag never once made it a front page item.
Didn't stop me from going to sea, donning every lifejacket offered.
I will never be as rich as Stockton Rush. As cautious as I am, I might fall victim to ordinary bad luck. I have had a few narrow scrapes on my motorbike, some of them in no way my fault. I have never had sleepless nights over that.
That’s great. So you do understand on some level that life is risk, which is to be understood and managed.
Humanity is deliberately meddling with systems humanity doesn't understand. Systems that, if upended, will cause us great harm.
One could make the same claim with fire, atomic bombs, and much more, and people did. You will have a stronger contention, but not much stronger, if you think we can never understand such systems.
Because of greed, stupidity, ignorance, shortsightedness. And optimism that a technological solution is just around the corner. In some cases, the knowledge is readily available in the public domain, but people look the other way, fingers in their ears.
Yes, there is all of that. There’s also brilliance, invention, self-sacrifice, love, and thoughtfulness. You neglect all these in favor of focusing on the worst in humanity. And the ignorance and stupidity do not go only one way.
You may call me a pessimist.
“Two things are infinite, as far as we know – the universe and human stupidity.” - Frederick S. Perls
Then Einstein proved that the universe is, in fact, finite.
Some caution is warranted. Too much caution and you get repressive, backwards systems that fear life and freedom more than anything else, and squash the very spirit that animates civilization.

If you only want to continue this digression and not discuss the topic, PM me. Otherwise I see no reason to continue cluttering this topic up with an abstraction.
 
I am a lapsed aquatic ecologist, I received my MSc in 1986, then went into computer programming because a biologist's career was not on the cards during the eighties.
I think using the sea - or any other natural body of water - as a heat sink for underwater data centres is a really bad idea.
 
I’m not completely against putting them in the ocean, but I would prefer they go into space.
 
Humanity is deliberately meddling with systems humanity doesn't understand. Systems that, if upended, will cause us great harm.
https://phys.org/news/2026-06-atmospheric-cooling-effects-stratospheric-aerosol.html

This is why I like in-space options. If sunshades--not dust mind you--block too much sunlight, they can be solar electric tugs.

Another maunder minimum? Reflectors to keep us warmer.

Messing with water columns, atmospheric chemistry--that cannot be undone quickly.
 
One can argue over the need for increasingly more numerous data centers in the first place. Let alone the knock on effects of putting these things into our oceans, which we're already trashing more than enough.
 
and we're seeing worrying signs of instability in some of the major currents - AMOC particularly. If we lose AMOC, the forecast is Europe's climate drops 8C, US sea levels rise, and god knows what other currents will shut down or shift as a response. (Remember, London and Berlin are on the same latitude as Calgary, without AMOC our climate turns Canadian).
You are talking about a group that thinks living on mars is not unreasonable.

Frankly, there is two possible outcomes for this group.
1. Death
2. Summon the AI god and beat death

It is absurd they'd care about the fishes and weather of future yurope when their personal noggin is on the line. The scale of things is just too many OOM off.
 
Last edited:
That’s absurd in all respects, shin_getter, but you do you. This thread has completely veered from the original point; perhaps it should be renamed BANANA: build absolutely nothing anywhere near anyone.

Or you can debate seaborne data centers, but I doubt you will.
 
Where in my post did I say not to build stuff? What I am saying that is that according to the valley e/acc belief system, the upside with compute is immense. (and logically "mundane" downside is nothing)

There is a reason why some people is trying to shove trillions of dollars into compute.
 
I’m referring to the entire thread, not just you. And because few people seem interested in debating the topic, it may as well end.

Also, one needn’t be an accelerationist to think data centers are useful. They’re what make the entire modern Internet possible, but people seem to think that services like this web forum magically exist in the ether, rather than inside of a server rack.
 
Can't think why this suddenly popped into my head, but here goes:
"The Internet is like a herd of performing elephants with diarrhoea - massive, difficult to re-direct, awe-inspiring, entertaining, and a source of mind-boggling amounts of excrement when you least expect it" - Gene Spafford

Replace one item in quote with AI.
No, not elephants.
And no, not diarrhoea.
 
http://large.stanford.edu/courses/2016/ph240/makalinao2/docs/UNU-GTP-SC-18-25.pdf

Heated water for the fishes~

The whole "global warming", "environmental damage" story is upside down. In the anthropocene, human action can impact the environment, and it can be done for good or bad and everything in between, and it can also be predicted to a extent.

The real question is how to manage these externalities. Consider the "global weather" market. Who can pay for what weather, and how do all the small projects impact get measured and have transactions work out. The end result shouldn't be no change to the natural, but discovering and engineering optimal climate (for whoever that can pay for it). The economics of climate have not even been properly attempted.

The current issues are just a tiny preview relative to the era where the combination of solar shades and mirrors covers areas comparable to the earth itself and any kind of FUBAR redirection of planetary thermal energy is on the table. Just think about how to build agencies to manage that situation, and when it is justified to shoot up a lot of space infra to disable rogue mirror operators.

I can just imagine someone in the year 3000 complain about grandfathered NIMBY rules that prevents sunbathing weather on the antarctic ...

Also, one needn’t be an accelerationist to think data centers are useful. They’re what make the entire modern Internet possible, but people seem to think that services like this web forum magically exist in the ether, rather than inside of a server rack.
Actual luddism is in the Zeitgeist. On some level technological civilization is inhuman and will perhaps will become completely so now workers dependence is going away soon-ish.

It is not unimaginable that conditions becomes so weird with such massive mismatch to human evolutionary impulses (to eat wrong food and get mind melted by social media) that living like the amish is the best way to maintain sanity.
 

Similar threads

Back
Top Bottom