Amazon AI Data Centers To Double as Carbon Capture Machines (semafor.com) 34
Amazon's data centers could soon double as carbon capture machines, offsetting the harmful effects of the massive amounts of energy required to run them. From a report: Amazon Web Services is partnering with startup Orbital Materials, which used artificial intelligence to create a new material specifically designed for separating carbon from hot air exhaust in data centers, the companies announced Monday.
Orbital Materials CEO Jonathan Godwin said he expects AWS to capture enough carbon to exceed the fossil fuel consumption used to power its AI data centers, giving them a net negative impact on climate change. The process will cost less than purchasing captured carbon to offset its climate impact, according to Godwin. The system, part of a pilot program at a to-be-determined data center location, works when outside air is sucked in and used to cool extremely hot semiconductors designed to run or train powerful AI models, such as Anthropic's Claude chatbot.
Orbital Materials CEO Jonathan Godwin said he expects AWS to capture enough carbon to exceed the fossil fuel consumption used to power its AI data centers, giving them a net negative impact on climate change. The process will cost less than purchasing captured carbon to offset its climate impact, according to Godwin. The system, part of a pilot program at a to-be-determined data center location, works when outside air is sucked in and used to cool extremely hot semiconductors designed to run or train powerful AI models, such as Anthropic's Claude chatbot.
which used artificial intelligence (Score:2)
> which used artificial intelligence
Of course they did, because humans are too dumb to do this themselves, right?
Re: (Score:2)
Re: (Score:2, Troll)
Of course they did, because humans are too dumb to do this themselves, right?
Of course we are. Humans are too dumb for a lot of things, which is why we invent tools to help us, be that the abacus, the differential engine, or a trained neural network algorithm (as we know it AI).
Most of us are fucking useless beyond remembering the Times table and how to add two numbers together.
Re: (Score:2)
Most AI is fucking useless beyond some statistics.
Re: (Score:2)
Completely and utterly false. You have fallen into the trap of thinking Most Ai = ChatGPT. It's not. We've been using advanced trained neural models for years and to great effect, we just didn't give the media hyped name AI and somehow recognised that there's a lot we couldn't do without them.
Couple of examples:
Using AI to help sharpen images - something that can now be done in seconds beating the performance of an expert spending many hours perfecting an image deconvolution algorithm.
Using AI to identify s
Re: (Score:2)
What would you say the difference is between ML and AI?
Your examples look like mostly ML.
Here's a tip (Score:4, Insightful)
Anyone discussing a carbon capture "technology" is selling you something.
It's not that we don't need to do it at some point, it's that every single proposal is a joke in terms of both scale and cost-effectiveness compared to reducing fossil fuel use in the first place.
This is supposed to make people feel better about their electric bills going up due to increased electric demand from worthless slop-centric data centers that draw gigawatts.
Re: (Score:3)
This is supposed to make people feel better about their electric bills going up due to increased electric demand from worthless slop-centric data centers that draw gigawatts.
It's more likely something their bought-and-paid-for local politicians can tell voters about after they green light a data center that nobody wants in their back yard.
"Look, it's good for the environment!". Meanwhile their diesel generators run every time the weather gets too warm or too cold.
Re: (Score:2)
Anyone discussing a carbon capture "technology" is selling you something.
Presumably they are selling "technology" in terms of a new technical development. I honestly don't understand what your complaint here is. The word is used throughout most of industry. Simple catalyst manufacturers will sell you "technology" that does exactly what they say it will. Yeah people sell you stuff. News at nine.
Re: (Score:2)
TFA is very light on details, but in principle if they can use the waste heat from a data centre to capture carbon using their new material, it would be something of a game-changer. That's a very big "if" though.
Addresses the "Air Movement" cost (Score:3)
Re: (Score:2)
Re: (Score:2)
The volume of air needed to offset their emissions depends mostly on how they are powering the data centre. If it's renewables or even some nuclear power, those emissions might be fairly low to start with. Presumably they are at least going to cover the roof in solar.
Expectations (Score:4, Funny)
Orbital Materials CEO Jonathan Godwin said he expects AWS to capture enough carbon to exceed the fossil fuel consumption used to power its AI data centers, giving them a net negative impact on climate change.
And I fully expect a harem of supermodels to service me tonight. I wonder if Mr. Godwin has the same degree of confidence in his expectations as I do.
Re: (Score:2)
Lacking in Specifics (Score:4, Interesting)
Wow, there is quite a bit missing here. The article describes their magic dust absorbing the carbon dioxide, and then simply says the carbon is stored underground. No mention of whether the magic dust goes underground, necessitating the purchase of more, or whether it's put through another process (that presumably doesn't cost energy) to pull the carbon out, to be sent underground using yet another process that doesn't cost energy. Good thing they used AI - there are more problems to be solved here than just the creation of magic dust.
Lot to unpack here. (Score:4, Insightful)
Orbital Materials CEO Jonathan Godwin said he expects AWS to capture enough carbon to exceed the fossil fuel consumption used to power its AI data centers, giving them a net negative impact on climate change.
Does this also cover the carbon production of the drilling, pumping, transport, refining, transport again and all the other steps in said fossil fuels used to generate the power? Color me skeptical, but I have my doubts any of that was factored in once they reached some magical, "NET NEGATIVE" theoretical threshold.
The process will cost less than purchasing captured carbon to offset its climate impact, according to Godwin.
Purchasing captured carbon? Purchasing captured carbon? Is that even a real thing you can do? And why would you want to when your primary motivation is using power? To stockpile the carbon for future release? Or is he talking carbon credit shell game bullshit, like what we've been told is the way forward so businesses don't even have to attempt to stifle carbon emissions?
The system, part of a pilot program at a to-be-determined data center location, works when outside air is sucked in and used to cool extremely hot semiconductors designed to run or train powerful AI models, such as Anthropic's Claude chatbot.
Ah. So, not only is it not capturing the carbon from the power generation used in the data center, it's a 'pilot program.' Likely set up in specific ways to make everything look super environmentally friendly for the decision makers so they can approve the real data center, which may or may not implement this little smokescreen on the carbon emissions inherent in almost all datacenters today.
Methinks the emperor's new clothes seem a bit thin and transparent, but perhaps it's just my cynicism at yet another carbon capture scheme that seems dubious on the surface being tossed out as a savior, with AI fairy sprinkles on top for good measure. If only they could find a way to incorporate crypto and blockchain into it.
Greenwashing (Score:2)
I'm sure the only use of AI was in writing the proposal.
Greenwashing at its best (Score:2)
So far, "carbon capture" has proved to be just as real and just as close as fusion powered generating plants. We'll leave the topic of the environmental effects of heat plumes on aquatic ecosystems for another day.
Re: (Score:3)
The difference is there's hope fusion might work someday as something other than to keep the skills for making nuclear weapons available to the government.
Cost effective carbon capture would violate the laws of thermodynamics. We're releasing CO2 as a byproduct releasing energy stored in hydrocarbon form... anything you do to capture that CO2 from the process that released it is reducing your energy efficiency. Anything you do to permanently sequester it is almost certainly going to make the entire proces
Re: (Score:2)
We could technically use solar energy to capture the excess carbon in the atmosphere to fix the current problem, however we need to stop putting carbon into the atmosphere in the first place. If we need energy then we should use the clean forms since that will be more efficient.
Re: (Score:2)
Thanks. I hadn't thought about it in those terms until you laid it out so concisely.
Does not sound very convincing... (Score:2)
Scam? Misdirection?
This is all smoke and mirrors (Score:3)
Data centers don't generate CO2 themselves, the power plants that power them do if they are Coal, Natural Gas ect. So don't put the CO2 capture on the data center for a press release, put it on the power plant. But amazon won't do this for one reason, cost.
Confusing article (Score:2)
I'm familiar with direct air capture systems. And I'm familiar with datacenter HVAC. This article is either confused, or is glossing over important details. Unless my reading comprehension is off, the article is suggesting that the sorbent is going to pull CO2 from the hot aisle of the datacenter, becau
Doctor Evil Pout (Score:2)
"Riiiiiiight"
The Carbon Capture Ponzi scheme (Score:2)
I wonder if we will all scratch our heads 20 years from now about what we were doing back then.
Liquid orAir (Score:1)
Wonder how this would work with the liquid cooling system they announced?
More greenwahing in... (Score:2)
So Amazon will start selling carbon credits? (Score:1)
Politics (Score:1)
I wonder, can this tech be scaled up to the government level - all the hot air used to offset a whole country's carbon emissions? If so, the past 4 years were a huge lost opportunity - but the next 4 years promise to be at least as lucrative.