They don't want you to know

00:05:28
https://www.youtube.com/watch?v=-lzQxbcrscc

Ringkasan

TLDRAs technology companies like Google and Microsoft rapidly expand their AI capabilities, the growth of data centers has surged. These centers require substantial amounts of fresh water and energy, raising environmental concerns. In 2023, Google's data centers alone utilized an immense amount of water equivalent to part of Lake Athabasca to keep servers cool. Microsoft has taken steps towards nuclear energy, agreeing to utilize facilities like the Three Mile Island reactor, to power their data centers. The growing demand calls for water that is clear of impurities, impacting local water supplies essential for daily human use. Communities near these centers, such as in Santiago, Chile, face challenges over resource allocation in already water-stressed areas. As AI drives advancements in fields like medicine and energy efficiency, it is crucial to balance its benefits against its environmental cost. Ensuring sustainability while minimizing ecological impact is key, pressing tech companies towards innovative water reuse and energy sourcing solutions.

Takeaways

  • 📈 Rapid growth of AI data centers is affecting energy and water resources.
  • 💧 Each use of AI technologies like ChatGPT consumes fresh water.
  • ⚡ Data centers require significant power, prompting some companies to use nuclear energy.
  • 🌍 The environmental impact of AI infrastructure includes water scarcity and high energy demands.
  • 🤖 AI offers potential breakthroughs in medicine and renewable energy efficiency.
  • 🌊 Some communities face water shortages due to local data center demands.
  • 🔍 Transparency and understanding of data center operations are crucial for sustainability.
  • 📊 Mid-sized data centers consume substantial water daily, equivalent to 100,000 homes.
  • 🔄 Solutions for water reuse and renewable energy sourcing are necessary.
  • 🧠 Companies race to control AI, impacting resource allocation globally.

Garis waktu

  • 00:00:00 - 00:05:28

    As AI progresses, companies like Amazon plan significant investments in data centers, consuming vast amounts of water and energy, often in water-stressed regions. Google's data centers in drought-prone areas show the environmental stress caused by such expansions. The lack of regulation on water usage by data centers is problematic, but transparency is improving. Nuclear power deals, like Microsoft's, illustrate attempts to sustainably power these centers, though the energy is solely for the centers’ operations. Balancing AI's potential in fields like medicine with its environmental toll is crucial, urging companies to adopt greener solutions.

Peta Pikiran

Video Tanya Jawab

  • How much water is used when asking ChatGPT questions?

    Asking ChatGPT 3 10 to 55 questions uses roughly 500mL of fresh water.

  • Why do data centers require so much fresh water?

    Data centers require fresh water to cool the systems, which need to be free of impurities to function effectively.

  • What energy sources are being used for AI data centers?

    AI data centers are increasingly using nuclear energy, as exemplified by Microsoft's investment in the Three Mile Island nuclear reactor.

  • How much water does a mid-size data center consume daily?

    A medium-sized data center uses about 300,000 gallons of water per day.

  • What environmental impact do AI data centers have?

    They consume significant amounts of water and energy, contributing to water scarcity and increased energy demand.

  • Why are local communities concerned about new data centers?

    Communities worry about data centers taking water from local freshwater sources, potentially leading to shortages.

  • What potential benefits does AI technology provide?

    AI can aid in designing new medication and understanding diseases, offering large strides in healthcare and efficiency improvements in renewable energy systems.

Lihat lebih banyak ringkasan video

Dapatkan akses instan ke ringkasan video YouTube gratis yang didukung oleh AI!
Teks
en
Gulir Otomatis:
  • 00:00:00
    as tech companies race to win the AI
  • 00:00:02
    Revolution the amount of data centers
  • 00:00:04
    that they are opening is growing
  • 00:00:06
    exponentially this requires increasing
  • 00:00:08
    amounts of fresh water and energy in
  • 00:00:10
    2023 alone Google used the equivalent of
  • 00:00:13
    1 18 of Lake athabaska to cool their
  • 00:00:17
    data centers and Microsoft recently
  • 00:00:19
    committed to restarting nuclear reactors
  • 00:00:21
    like Three Mile Island the site of the
  • 00:00:23
    1978 partial nuclear meltdown to create
  • 00:00:26
    enough energy for their growing amount
  • 00:00:27
    of AI intensive data centers now there
  • 00:00:30
    is a lot of talk right now about the
  • 00:00:31
    future potential of AI I'm working on a
  • 00:00:34
    video about Google mind and Alpha fold's
  • 00:00:36
    ability to create breakthrough
  • 00:00:37
    medications and there's a ton of
  • 00:00:39
    excitement and optimism in other spaces
  • 00:00:41
    too but one thing we know for sure about
  • 00:00:42
    AI right now is that Microsoft Tesla
  • 00:00:45
    meta all these companies they're
  • 00:00:46
    building massive amounts of data centers
  • 00:00:48
    in order to power their AI systems in
  • 00:00:51
    their race to monopolize artificial
  • 00:00:53
    intelligence and this will have a huge
  • 00:00:56
    impact on our futures AI technology
  • 00:00:58
    relies on data centers that require so
  • 00:01:00
    much energy that they get so hot that
  • 00:01:03
    they need water pumped through them to
  • 00:01:05
    be cooled in order to function this is a
  • 00:01:07
    Google data center those colorful pipes
  • 00:01:09
    those aren't wires that's kind of what I
  • 00:01:11
    thought initially those are pipes of
  • 00:01:12
    Cold clean water which flow through the
  • 00:01:15
    system keeping the technology from
  • 00:01:16
    overheating in 2023 alone Microsoft data
  • 00:01:19
    centers consumed 7.8 million cubic met
  • 00:01:22
    of freshwat up 34% from the year before
  • 00:01:26
    because of how many more data centers
  • 00:01:27
    they were running to power their new AI
  • 00:01:29
    Tech technology Pursuits the water that
  • 00:01:31
    cools and cleans the technological
  • 00:01:33
    systems needs to be free of impurities
  • 00:01:36
    bacteria debris and dirt this is the
  • 00:01:38
    reason why these data centers rely on
  • 00:01:40
    fresh water for cooling the water needs
  • 00:01:43
    to be extremely clean therefore the
  • 00:01:44
    water is taken from the same sources
  • 00:01:47
    that we use to drink cook and wash with
  • 00:01:49
    so how much water are you consuming
  • 00:01:51
    every time you use an AI technology like
  • 00:01:54
    chat GPT based on Microsoft's 2022
  • 00:01:57
    annual report on fresh water used for
  • 00:01:59
    cooling system asking chat GPT 3 10 to
  • 00:02:02
    55 questions or prompts uses roughly 500
  • 00:02:05
    mL of fresh water I think this is
  • 00:02:07
    important to know and remember that when
  • 00:02:09
    you're using AI technology like chat GPT
  • 00:02:11
    there is an environmental impact it
  • 00:02:14
    makes sense that we haven't really
  • 00:02:15
    thought about this that much companies
  • 00:02:16
    aren't talking about this as they expand
  • 00:02:18
    their AI initiatives Amazon is planning
  • 00:02:20
    to spend a hundred billion dollar in the
  • 00:02:23
    next 10 years to build data centers
  • 00:02:25
    around the world consuming massive
  • 00:02:26
    amounts of energy and water in fact
  • 00:02:29
    earlier this year they secured the
  • 00:02:30
    approval to build two new data centers
  • 00:02:32
    near Santiago Chile Santiago Chile I uh
  • 00:02:35
    have been there once that's why I know
  • 00:02:37
    how to say it so while is already a
  • 00:02:39
    water stressed City so there has been
  • 00:02:41
    Mass protests and concern as people are
  • 00:02:43
    wondering where this freshwat is going
  • 00:02:44
    to come from cuz it's likely going to
  • 00:02:46
    come from their personal freshwater
  • 00:02:47
    sources in the DS Oregon I've never been
  • 00:02:50
    there I don't know if it is pronounced
  • 00:02:51
    DS Google has three data centers and one
  • 00:02:54
    quarter of the city's water supply is
  • 00:02:56
    being consumed by these three data
  • 00:02:58
    centers and they're planning to build
  • 00:02:59
    two more there has been a lack of
  • 00:03:01
    Regulation and standards about knowing
  • 00:03:03
    how much water these data centers are
  • 00:03:05
    consuming cuz this is all quite new
  • 00:03:06
    technology but we are getting more
  • 00:03:08
    transparency than ever before based on
  • 00:03:10
    estimates from 2022 a midsize data
  • 00:03:13
    center uses about 300,000 gallons of
  • 00:03:15
    water a day equal to the water
  • 00:03:17
    consumption of about 100,000 homes
  • 00:03:20
    Microsoft recently announced it had
  • 00:03:22
    signed a 20-year deal to purchase energy
  • 00:03:24
    from the dormant 3M Island nuclear
  • 00:03:26
    reactor the one that had a meltdown in
  • 00:03:28
    the 1970s the reason was to create
  • 00:03:30
    nuclear energy to power their growing
  • 00:03:32
    need for data centers in their AI race
  • 00:03:35
    now I am all for this I'm all for
  • 00:03:37
    nuclear energy but this is only being
  • 00:03:40
    used to power data centers these nuclear
  • 00:03:42
    reactors aren't being used to create
  • 00:03:43
    energy for us this is all at of time
  • 00:03:45
    when we're trying to figure out how to
  • 00:03:47
    get energy that isn't going to be
  • 00:03:49
    burning fossil fuels but again these
  • 00:03:50
    nuclear reactors are just powering data
  • 00:03:52
    centers but as stated earlier I do think
  • 00:03:55
    that AI can be very important in
  • 00:03:56
    specific Fields Alpha fold and AI system
  • 00:03:59
    developed by Google deep mine won a
  • 00:04:02
    Nobel Prize in chemistry because among
  • 00:04:04
    other things it can predict and design
  • 00:04:06
    protein structures this AI technology
  • 00:04:08
    could see breakthroughs in our
  • 00:04:09
    understanding of disease create new
  • 00:04:11
    medicines and all of this we would not
  • 00:04:13
    be able to do without assistance of ai
  • 00:04:15
    ai in theory can help Renewable Energy
  • 00:04:17
    Systems become more efficient but at the
  • 00:04:19
    same time one of our biggest issues as
  • 00:04:21
    humanity is the climate crisis and water
  • 00:04:23
    scarcity so we need to balance the pros
  • 00:04:25
    and cons of AI right now as all of these
  • 00:04:27
    companies sort of Skyrocket creat
  • 00:04:29
    creting these data centers to make sure
  • 00:04:31
    they don't do more harm than good I
  • 00:04:33
    think this video is just about us
  • 00:04:35
    understanding I'm back in school right
  • 00:04:36
    now doing my masters and lots of people
  • 00:04:37
    are just using chat GPT all the time I
  • 00:04:40
    think it's good for us to think about
  • 00:04:41
    the water consumption that goes along
  • 00:04:43
    with us using chat GPT in our day-to-day
  • 00:04:46
    life I think we need to put pressure on
  • 00:04:47
    these companies to figure out ways to
  • 00:04:49
    maybe reuse this water or have the fresh
  • 00:04:51
    water end up in productive places like
  • 00:04:53
    on farms hopefully this video and public
  • 00:04:55
    understanding and increased transparency
  • 00:04:57
    around these data centers will help us
  • 00:04:59
    understand how this AI Tech actually
  • 00:05:01
    works if these data centers become
  • 00:05:03
    powered by renewable energy nuclear
  • 00:05:05
    energy who knows it could create
  • 00:05:06
    breakthroughs and Innovation this could
  • 00:05:08
    be a good thing but this video was just
  • 00:05:09
    about you and I understanding what's
  • 00:05:11
    going on right now with these data
  • 00:05:13
    centers thanks for watching us all these
  • 00:05:14
    years we've been doing this for 12 years
  • 00:05:16
    which is insanity for all you young ons
  • 00:05:18
    out there do not ask how old we are hit
  • 00:05:19
    that like And subscribe button I don't
  • 00:05:21
    even know if that's still what people do
  • 00:05:23
    and thanks for watching we will see you
  • 00:05:25
    soon for a new science video I see peace
Tags
  • AI
  • data centers
  • water usage
  • energy consumption
  • Google
  • Microsoft
  • nuclear energy
  • environmental impact
  • tech companies
  • resource allocation