Anson

41 posts

Anson

Anson

@EM203Anson

Katılım Eylül 2024
19 Takip Edilen6 Takipçiler
Sabitlenmiş Tweet
Anson
Anson@EM203Anson·
A thread on AI and it's energy consumption crisis. 🧵
GIF
English
1
0
0
92
Anson
Anson@EM203Anson·
A thread on AI and it's energy consumption crisis. 🧵
GIF
English
1
0
0
92
Danielle Fong 🔆
Danielle Fong 🔆@DanielleFong·
i've been going on and on about this, but basically, AI flips electricity on its head. previously, nearly half the cost of electricity to the consumer was allocated towards distributing it where (and when) you want it. AI data centers want upwards of hundreds of megawatts of power in one place, most of the time. (and for training they don't care about latency) The most constrained thing isn't even the energy itself -- there's a surplus during the solar peak -- it's often the sheer availability & the interconnection. It now can take multiple years to get a grid interconnect to draw power of this magnitude, if the location can even handle it. The capital cost of the power infrastructure is just a tiny fraction (like 3%!) of the capex of the compute, and even just the depreciation of the compute exceeds the cost of even premium power. Hence, AI hyperscalers and those that aspire to be in their class are traveling to where the power is, are building where power has been (and there's legacy transmission to support it, like old nuclear plants) and are getting into the business of actually building powerplants and reactors. Utilities and transmission and distribution companies, interconnection queues, all are used to react much slower -- over many years -- unlike the top technology companies, now vying to compete at the highest levels of AI performance. Since ~99% of energy technologies previously died withering while waiting for utilities to consider them bankable, this represents an extremely fertile, attractive new state of play if you're bringing a new energy technology to market (💁🏻‍♀️🔆). Whereas before you'd have to brave what was once called a "green valley of death", now you have teracap companies like microsoft bidding on: - Conventional, large nuclear fission - Small modular nuclear reactors - Engineered Geothermal - Fusion! - Stirling Engines 😬 - Who knows what else - Maybe you, Anon! Even Tesla is standing up gas generator arrays to burn fossil fuels for their AI power supply. Shit is getting real
Mark Nelson@energybants

BREAKING: BLOCKBUSTER MICROSOFT DATACENTER DEAL RESURRECTING THREE MILE ISLAND NUCLEAR PLANT Microsoft and nuclear plant owner Constellation have agreed to a massive, unprecedented deal to restart the closed Three Mile Island by 2028 to power its datacenters, per the New York Times. The deal? Microsoft purchasing as much power as possible from its 880 MW reactor over 20 years for prices rumored to be above $100 per MWh. Most famous for its 1979 meltdown, TMI closed in 2019 because of cheap fossil fuels and tech companies refusing at the time to consider buying its electricity to meet clean energy goals.

English
114
352
2.1K
712.9K
ChainGPT
ChainGPT@Chain_GPT·
AI is powerful, but at what cost? 🔌 Here's a look at the current trends and projections for AI's energy use. 🧵 [1/5]
ChainGPT tweet media
English
99
93
2K
110.3K
Ethan Mollick
Ethan Mollick@emollick·
I think it is completely reasonable to be worried about the growing electrical demand of AI (I am), but I often see people assuming AI is a much bigger energy user than it actually is right now This is from the International Energy Agency's 2024 report, which gives some context
Ethan Mollick tweet media
English
26
50
397
41.4K
Alain.l812
Alain.l812@AlainL812·
What should we do then? I feel like we cant avoid the integration of AI in our lives given we are entering the era of tech, so what is a sustainable way to actually use it/implement it?
Lakyn Thee Stylist@OgLakyn

Not only that, but, the computers require so much computing power, which generates so much heat, they use a LOT of water to keep them from overheating. Training ONE AI program for three weeks can require over 250,000 gallons. It's harmful RIGHT NOW, let alone long-term.

English
1
0
0
13
Lakyn Thee Stylist
Lakyn Thee Stylist@OgLakyn·
Not only that, but, the computers require so much computing power, which generates so much heat, they use a LOT of water to keep them from overheating. Training ONE AI program for three weeks can require over 250,000 gallons. It's harmful RIGHT NOW, let alone long-term.
trapper/rapper@lonthecelebrity

I was in a workshop a few weeks ago and they talked about how harmful Chat GPT will be longterm. One search uses as much electricity as leaving the lights on for like 20 minutes. I never even thought about that.

English
63
9.2K
26K
775.8K
Reid Southen
Reid Southen@Rahll·
AI data centers are on track to demand so much energy as to cause shortages, which will then drive up the cost of electricity.
Reid Southen tweet media
English
9
83
262
9K
Reuters Business
Reuters Business@ReutersBiz·
Artificial intelligence may seem to generate content from the cloud, but its vast energy use brings it back down to earth. According to the International Energy Agency, data centers could be using as much energy as Japan by 2026. Industry experts explain AI's energy problem
English
4
8
5
9K
The Future Investors
The Future Investors@ftr_investors·
$MSFT $GOOGL $AMZN Big Tech bets on nuclear energy to solve its AI power problems.
The Future Investors tweet media
English
6
50
317
28.9K
Nuclear Hazelnut 👷🏻‍♀️
Nuclear Hazelnut 👷🏻‍♀️@NuclearHazelnut·
Did you know? A single AI response requires about 10–100 watt-seconds of energy. Now, imagine millions of requests per hour globally. 💡 In 1 hr, an AI like @ChatGPTapp could easily process 1 million requests, consuming an estimated 27 kWh—equivalent to running a refrigerator for 15 days. Multiply that by every data center worldwide, and you get a massive energy demand.⚡ Nuclear energy provides a continuous, reliable source of clean power, producing up to ~1 GW per reactor—enough to power AI systems 24/7 while maintaining net-zero emissions. 🌍 The future of AI is bright, but it needs sustainable, round-the-clock power. Nuclear Energy is the answer.
Nuclear Hazelnut 👷🏻‍♀️ tweet media
English
21
71
438
24.8K
Eric
Eric@EricFlaningam·
Among the many bottlenecks for AI data centers, energy might be the most important and the most difficult to address. IF estimates of data center energy consumption turn out to be true (or even in the vicinity of truth), our current energy infrastructure will not be able to support those demands. Unfortunately, there’s no easy way to increase energy capacity quickly. I think it will take many solutions to solve this problem, just a few of them are: - Off-Grid Power Generation - Nuclear - Long-Duration Energy Storage (Batteries) - Building Where the Power’s At - Grid Optimization Tech I shared some thoughts on this in the newsletter this morning (link below).
Eric tweet media
English
26
183
819
169.1K
Tsarathustra
Tsarathustra@tsarnick·
Sam Altman says abundant intelligence and abundant energy will lead to solving every problem in physics and eventually we'll be talking not about using nuclear fusion or solar to power AI but a Dyson Sphere
English
58
120
690
86.9K
Tsarathustra
Tsarathustra@tsarnick·
Alexandr Wang says 20 gigawatts of energy will be needed in the next 2-3 years to power new AI data centers - 5 Chicagos worth of energy
English
60
115
487
543K
Man of Letters.
Man of Letters.@Letter_to_Jack·
This is the complete breakdown of the Power Demand and Energy Consumption of nearly all domestic appliances that we use in Nigeria based on your requests. Go through this carefully. Source: Ikeja Electric.
Man of Letters. tweet media
English
23
316
479
213.8K