The Dark Side of AI: We’re Heading Towards An Energy Crisis

Summary

The video script addresses the looming electricity crisis due to AI's soaring energy demands, highlighting the irony of tech giants advocating for energy conservation while contributing to the problem. It points out the rapid growth of AI energy needs, far outpacing electricity production. The script proposes Bitcoin mining as a practical solution, explaining how it can support renewable energy projects, provide a secondary revenue stream for energy providers, and act as a flexible partner during energy crises. Examples from around the world illustrate how Bitcoin mining is already positively impacting renewable energy financing and consumption.

Takeaways

• 🔋 The looming shortage of electricity could impede the operation of AI chips, suggesting a potential energy crisis driven by AI's growing energy demands.
• 💡 AI is creating an energy crisis that could significantly increase electricity bills, highlighting a dark side of AI development that is not widely recognized.
• 🏢 Tech giants like Tesla, Google, and Microsoft are contributing to the AI energy crisis despite advocating for energy conservation and sustainability.
• 🌐 The rapid growth in AI's energy consumption is outpacing the increase in electricity production capacity, indicating a potential future shortfall.
• 💰 The high costs and energy demands of data centers are making them a unique asset class in real estate, with significant implications for energy supply.
• 🚀 Tesla's gigafactory in Austin, Texas, exemplifies the scale of energy consumption for AI, with massive data centers requiring substantial power to train AI models.
• 🔌 The energy-intensive process of AI model training is compared to the energy usage of constructing skyscrapers, emphasizing the scale of the problem.
• 🌿 Bitcoin mining is presented as a solution to the AI energy crisis, as it can help finance renewable energy projects and provide a steady buyer for energy.
• ⚡ Bitcoin miners can act as flexible energy consumers, adjusting their consumption during peak and off-peak hours, which benefits both the miners and energy providers.
• 🌍 Global examples of Bitcoin mining aiding renewable energy projects, such as in the United Arab Emirates, Iceland, and Kenya, demonstrate the practicality of this solution.

Q & A

What is the potential next shortage that could impact technology, according to the video?

The potential next shortage discussed in the video is electricity, as there might not be enough electricity to run all the chips required for AI and technology advancements.

What is described as the 'dark side' of AI in the video?

The 'dark side' of AI refers to the energy crisis that AI is creating, which could lead to a significant increase in electricity bills and affect everyday energy usage.

Why are data centers considered a unique asset class in real estate according to the video?

Data centers are considered a unique asset class in real estate because each building can cost four to five billion dollars and require massive amounts of power, which are unprecedented in the real estate business.

What are the two main questions raised by the video regarding the development of data centers?

The two main questions raised are: where will we get the energy from to power these data centers, and who will pay for it.

How does the video suggest that large tech companies contribute to the AI energy crisis?

The video suggests that large tech companies, while advocating for energy conservation and net carbon zero, are major contributors to the AI energy crisis due to their massive energy consumption for AI and data centers.

What is the significance of the Tesla gigafactory in Austin, Texas, as mentioned in the video?

The Tesla gigafactory in Austin, Texas is significant because it is the largest car manufacturing plant in the country and is also set to become one of the largest data centers in the world for training AI models.

What does the video claim about the growth rates of electricity production capacity and AI energy demands?

According to the video, electricity production capacity is growing by about 2% a year through 2035, while AI energy demands are growing at a much faster rate of 23% a year.

How does Bitcoin mining provide a solution to the AI energy crisis as per the video?

Bitcoin mining is presented as a solution to the AI energy crisis because it helps finance renewable energy projects, provides a secondary revenue source for energy providers, and can shut off operations during severe weather or energy shortages, making it a flexible and beneficial partner.

What are the three key ways in which Bitcoin mining helps bootstrap the AI revolution according to the video?

Bitcoin mining helps bootstrap the AI revolution by making financing renewable projects more feasible, providing a secondary revenue source for utilities, and being a better partner during energy crises by being able to shut off operations when needed.

What is the role of battery storage in the context of the AI energy crisis as discussed in the video?

The video questions the feasibility and cleanliness of battery storage projects as a solution to the AI energy crisis, pointing out the challenges in scaling up battery storage to meet the projected demands and the environmental impact of battery production and disposal.

Outlines

00:00

🔋 AI's Looming Energy Crisis

The video script discusses the impending energy crisis due to the rapid advancement of AI and its insatiable appetite for electricity. It posits that the next major shortage will be in electricity, as the demand for power to run AI chips outstrips supply. The script suggests that before hitting a capital investment threshold, energy constraints will become a limiting factor. It highlights the dark side of AI, where billionaires are aware of the energy crisis that AI is creating, which could lead to a surge in electricity bills. The video aims to expose this reality and propose solutions to prevent a dystopian future. It questions the sustainability of the data center business, given the massive capital expenditures by tech giants and the unique challenges of power supply and cost associated with these facilities.

05:00

🌐 The Digital-Physical Transformation Gap

Paragraph 2 delves into the disparity between the growth of digital transformation, fueled by AI, and the physical infrastructure's ability to keep pace. It cites studies showing that while electricity production capacity is increasing by about 2% annually, AI's energy demands are growing at a staggering 23% per year. The script discusses the energy-intensive process of AI model training, which is likened to the energy consumption of constructing a skyscraper. It also touches on the continuous energy usage of inference models that run in the background. The paragraph emphasizes the irony of tech giants advocating for energy conservation while simultaneously contributing to the AI energy crisis. It contrasts the growth in carbon emissions of companies like Google and Microsoft with their sustainability reports and goals, suggesting a disconnect between their actions and their environmental commitments.

10:02

⚡ Bitcoin Mining as a Renewable Energy Solution

The final paragraph presents Bitcoin mining as a practical solution to the AI energy crisis. It outlines three key ways in which Bitcoin mining can support renewable energy projects: by making financing more feasible, providing a secondary revenue stream for energy providers during off-peak hours, and being a more flexible partner during energy crises compared to AI data centers. The script discusses real-world examples of Bitcoin mining being used to monetize stranded energy and fund renewable projects globally, such as in the United Arab Emirates, Iceland, and rural Kenya. It challenges the feasibility of large-scale battery storage as an alternative solution, questioning both the environmental impact and the cost of such projects. The video concludes by advocating for Bitcoin mining as a tangible and already implemented method to support the growth of AI in a sustainable manner.

Keywords

💡Electricity Shortage

Electricity shortage refers to the inability to meet the demand for electrical power, which is a central concern in the video. It is discussed in the context of the increasing demand for power due to the growth of AI and data centers. The script mentions that the next crisis could be a lack of electricity to run all the chips, illustrating the potential impact on technology and infrastructure reliant on a stable power supply.

💡AI

Artificial Intelligence (AI) is the simulation of human intelligence in machines to perform tasks like learning and problem-solving. In the video, AI is highlighted as a technology driving an energy crisis due to its high computational demands. The script discusses the 'dark side of AI', suggesting that its energy consumption could lead to higher electricity bills and potential shortages.

💡Data Centers

Data centers are facilities used to house servers, storage systems, and networking equipment for the purpose of storing and managing data. The video script mentions data centers in relation to their high energy consumption and the valuation of this business sector. It points out that these centers are unique in real estate due to their high power demands, which can range from 100 to 500 megawatts.

💡Energy Constraints

Energy constraints refer to limitations in the availability or capacity to supply energy. The video discusses how energy constraints could become a limiting factor in the development of new data centers and AI technologies. It suggests that before reaching a point where further investment in AI is not worth it, we might encounter energy limitations that hinder progress.

💡Sustainability

Sustainability in the context of the video pertains to the ability to maintain processes or states in a way that avoids depletion of natural resources and minimizes negative environmental impact. The script critiques tech giants for not practicing what they preach about sustainability, as their energy consumption for AI and data centers has increased significantly.

💡Carbon Emissions

Carbon emissions are the release of carbon dioxide (CO2) or other carbon compounds into the atmosphere. The video uses carbon emissions as a measure of environmental impact and sustainability. It points out that companies like Google and Microsoft, despite their sustainability reports, have seen an increase in their carbon emissions, which is at odds with their public commitments to environmental responsibility.

💡Bitcoin Mining

Bitcoin mining is the process of validating transactions and adding them to the public ledger of the Bitcoin blockchain. In the video, it is presented as a solution to the AI energy crisis because it can utilize excess energy from renewable sources, thus helping to finance and bootstrap renewable energy projects. The script provides examples of how Bitcoin mining is already being used to monetize stranded energy and reduce emissions.

💡Renewable Energy

Renewable energy refers to power sources that are naturally replenished and can be used sustainably, such as wind, solar, and hydroelectric power. The video suggests that renewable energy projects are key to addressing the energy demands of AI and data centers. It highlights how Bitcoin mining can support these projects by providing a market for excess energy and helping to finance their development.

💡Hydropower

Hydropower, or water power, is a form of renewable energy that uses the potential energy of water to generate electricity. The script mentions hydropower in the context of Iceland, where Bitcoin miners are utilizing the country's abundant waterfall energy to mine Bitcoin, turning it into a profitable revenue stream and supporting the local economy.

💡Energy Consumption

Energy consumption refers to the use of electrical or fuel energy by devices or systems. The video discusses the high energy consumption of AI data centers and how it is outpacing the growth of electricity production capacity. It provides statistics to illustrate the rapid increase in AI energy demands and compares it to the slower growth of overall energy production.

💡Inference

In the context of AI, inference is the process of deriving conclusions or making decisions based on a trained model. The video explains that inference, which involves continually running and updating AI models, is energy-intensive. It contrasts inference with the training of AI models, noting that both processes contribute significantly to the overall energy consumption of AI technologies.

Highlights

Next year may see a shortage in electricity due to the increasing demand for powering AI chips.

AI is creating an energy crisis that could lead to a significant increase in electricity bills.

The data center business might be facing a bubble due to high valuations and unprecedented capital spending.

Data centers are built to meet the needs of big tech companies, creating a demand for massive amounts of power.

The energy required for AI is so high that it could affect everyday energy use for appliances and lights.

Elon Musk's quote about not being able to find enough electricity to power AI chips is highlighted.

Tesla's gigafactory in Austin, Texas, is set to become one of the largest data centers in the world.

Tesla, in partnership with Nvidia and Dell, is working on Colossus, the largest AI training model.

Tech giants are contributing to the AI energy crisis despite advocating for energy conservation.

Google's and Microsoft's sustainability reports show an increase in carbon emissions despite their eco-friendly claims.

The global spending on cloud infrastructure is outpacing the spending on new power plants by electric utilities.

AI energy demands are growing at a rate that far exceeds the growth of electricity production capacity.

Training AI models is as energy-intensive as building skyscrapers.

Bitcoin mining is presented as a solution to bootstrap AI and renewable energy projects.

Bitcoin mining can provide a secondary revenue source for energy providers during off-peak hours.

Bitcoin miners can shut off operations during severe weather crises, making them better partners than AI data centers.

Examples of Bitcoin mining helping to fund renewable projects in the UAE, Iceland, and rural Kenya are given.

Bitcoin mining is positioned as a practical solution for funding AI projects and building infrastructure.

Transcripts

then the next shortage will be

electricity so I think next year you'll

see the electricity that they just can't

find enough electricity to run all the

chips there is a capital question of

like okay at what point does it stop

being worth it to put the capital in but

I actually think before we hit that

you're going to run into energy

constraints there's a dark side of AI

that the billionaires know that the

average American doesn't AI is creating

an energy crisis that's going to cause

your electricity bill to to explode

today we'll uncover this dark reality

and propose some solutions for how we

avoid this dark Future Let's Usher in

the AI revolution in a sensible

[Music]

way is there a bubble in the data center

business given some of the valuations I

mean the question is and we we all watch

the reports you know for the big tech

companies how much what's their Capital

spend going to be in the following years

but this asset class is is very unique

for Real Estate the each one of these

buildings could cost four or5 billion do

you're talking about power numbers that

we've never en encountered in the real

estate business 100 megawatt 200 up 500

megawatts those buildings it's so what's

amazing about them power as difficult it

is to get becomes the uh constraint to

new development so you don't have over

supply of these data centers and for the

most part they're built least to one of

the the big hyperscale tech companies

the two obvious questions that should

come from this are where are we going to

get the energy from and who's going to

pay for it these sheer amount of

resources that are needed to provide

compute for another supercomputer are

just going to rob us of everyday energy

applications like our appliances or our

lights or the ability to charge a cell

phone don't take my word for it the

billionaires already know that we don't

have enough energy to fuel this AI

Revolution we do need way more energy in

the world than I think we thought we

needed before my my whole model of the

world is that the two important

currencies of the future are compute SL

intelligence and energy um you know the

ideas that we want and the ability to

make stuff happen and uh the ability to

like run the compute and I think we

still don't appreciate the energy needs

of this technology let's go back to this

quote that we started the video with

from Elon Musk they just won't be able

to find enough electricity to power

these chips well Tesla is one of the

main contributors to this AI energy

crisis that we are barreling into right

now the Tesla gigafactory in Austin

Texas is already the largest car

manufacturing plant in the country and

soon to be one of the largest data

centers in the world to train models for

their full self-driving capabilities in

their Tesla cars in a tweet earlier this

summer he announced that they're

building for 130 megaw of capacity and

soon to be tripling that number in just

18 months this past weekend he announced

Colossus the largest training model in

AI data center in the world he announced

these plans in conjunction with their

Partners Nvidia and Dell and they will

have over 100,000 of the h100 chips

which is nvidia's chip in the most

powerful AI chip in the world the crazy

part about this is they said they did it

in 122 days this is by far the fastest

that anyone's been able to stand up this

amount of energy and compute for an AI

data center project to dat and they're

not the only ones we haven't even gone

into other projects from meta or Google

or any of the other large Tech Giants

the irony is these large Tech Giants

were the main ones talking about how we

need to conserve energy how we need to

go net carbon zero and yet they're the

main culprits in this AI energy crisis

that we're heading head first into I dug

into Google's 2024 sustainability report

and what's interesting about the

sustainability report is there's an AI

insight button which produces more

emissions and makes us less sustainable

but when I asked it the question of has

Google increased its carbon emissions

here's what it had to say Google has

been cost planing us and saying that

they're doing good for the environment

when actually their carbon emissions

have increased 47% from 2019 to 2023 and

you know what Microsoft sustainability

report wasn't much better Microsoft has

increased their carbon emissions from

2020 to 2023 by 30% what's funny is that

when when you look at this chart their

carbon emissions are growing 10% a year

and yet they expect to go negative

carbon emissions by 2030 I'm not good at

math but that math doesn't math to me

the problem is pretty simple Mark Mills

the author of The Cloud Revolution and

AI energy expert said it in a pretty

simple manner that we can easily

understand in dollar terms uh we

spending more money building

infrastructure for the cloud globally

Capital then all the electric utilities

in the world are spending combined

to build new power plants in a nutshell

it's pretty simple digital

transformation is outpacing physical

transformation according to an eel

energy study the amount of electricity

production capacity is growing by about

2% a year through 2035 but the AI energy

demands are growing at 23% a year and if

you think that number is crazy Boston

Consulting groups numbers aren't that

far off they're projecting 50 to 20%

growth a year from now to 2030 the

capacity for AI data center energy usage

is expected to Triple to 72% of total

energy consumption why is that well

there's two types of training there's

training a model which has to be updated

daily and retrained every time there's

new data and there's new information and

this takes about as much energy as

building a skyscraper this training is

electricity intensive taking about 1300

kwatt hours per day to train one of

these models then there's inference

which means means that the model that's

continually being updated and trained is

continually running in the background at

all times that's why a chat gbt search

takes 3 to 30 times as much energy as a

basic Google query and if any of you

have ever used any of these sort of

models in the past you know that it

doesn't nail the answer on the first

time contributing to a larger number of

queries than a general Google search

which has more results on a single page

and the crazy thing about all of this

this is just the AI data Center's impact

on our energy usage we're not even

considering the fact that we want to

have more electric vehicles and we want

to onshore the manufacturing that's gone

overseas to China over the past several

decades so what's the solution to our AI

energy crisis I've Heard lots of things

from our folks that have been talking

about the Green Revolution for several

decades but there's actually a real

practical and already implemented

solution happening all over the world

that is bootstrapping AI generation and

renewable energy projects globally

that's Bitcoin mining Bitcoin mining

helps to bootstrap the AI revolution in

three key ways number one it helps to

make financing these projects more

feasible you see when you're standing up

a large power project or a renewable

energy project typically that energy

source is far away from City centers and

people that actually buy and use that

energy and so the further that energy is

away from these people or the buyers of

that energy the more infrastructure you

have to put in place

Transformers lines CA

all of these things to transport that

energy from the source to the buyer then

you also have the fact that energy

dissipates the more distance that it

travels and so the main issue when

financing these projects is you don't

have a ready and early buyer for this

energy until you've sunk a ton of money

into the project insert Bitcoin mining

they bring the mining machines right to

the source of the energy and they have

an Always On Demand buyer of this

electricity and so you're able to

actually fund and get profits from

projects much sooner than you otherwise

would win-win for everybody number two

Bitcoin miners provide utilities and

energy providers and producers a

secondary Revenue Source let's take a

look at this chart throughout a typical

day peak hours for energy and the costs

of energy are most expensive between

8:00 a.m. and 1: p.m. with the most

costly part of the day being between 4:

p.m. and 10 p.m. when we all get off of

work or if you're in Texas or California

you come home and you turn on your air

conditioning and so the cost of this

energy is unsustainable and not really

good for Bitcoin miners but when energy

usage is at its lowest or off peak hours

from 12:00 p.m. or 12: at night to 8:00

in the morning well then you have a

great time for Bitcoin miners to be able

to come online consume that energy at

more profitable rates for the utility

and energy producers this seems like a

win win win for everyone opportunity

number three Bitcoin miners are just

Flatout better Partners than our AI

overlords when there's a severe weather

crisis or there's a need for people that

actually need electricity to have it

back Bitcoin miners are able to shut off

and turn off at a moment's notice let's

compare this to someone like Tesla who

maybe they have a large AI training

computer that's running in their Austin

Texas gigafactory that's used to program

and provide updates to all of their cars

running their full service driving

technology well what if they need to

make a critical update to all of the

cars that are on the road driving

themselves while you're asleep do you

think that because there's bad weather

outside or there's inclement conditions

that they're just going to turn that off

for the sake of being a good partner I

don't think so Bitcoin miners have

already been doing this and have

demonstrated this capability in places

like Texas and Tennessee for years and

before you think this is just some crazy

Pie in the Sky idea this is already

being done at several places around the

world in the United Arab Emirates

Bitcoin miners are soaking up stranded

solar farm energy and desalination plant

energy that would otherwise be wasted to

profitably fund more renewable projects

in the world what about Iceland where

they have a large amount of potential

Hydro power that's not being monetized

while Bitcoin miners are going near

waterfalls and actually consuming this

energy through the mechanical output of

this Hydro or this water and turning it

into revenue streams that are profiting

Iceland what about a village in rural

Kenya where a company like gridless with

Eric hman is taking Bitcoin miners

straight to a volcano and Mining this

stranded energy in order to build the

infrastructure to power 5,000 homes in

rural Kenya that otherwise wouldn't have

been able to finance infrastructure

projects to transmit that energy out to

those rural places a lot of people in

the Bitcoin industy already know about

cruso energy who's able to slash

emissions by 63% and prevent the spewing

of methane gas into the atmosphere by

literally capturing that flared gas

putting it into a generator and Mining

Bitcoin with it win win win for everyone

there are Bitcoin miners like ter wolf

who use 93% of zero carbon energy

sources in order to fund their Bitcoin

operations through nuclear and through

Hydro there's several other publicly

traded miners like core scientific or

Marathon who've already been capturing

this trend the solution is simple

Bitcoin actually makes renewable

projects profitable the next push that

you're going to see versus real things

that have been implemented in the wild

is projects like what the economist is

pushing ter

battery storage for electricity grids

it's taking us 20 or 30 years to get to

200 megaw of battery storage for grids

but somehow we're going to get to a

gigawatt by 2030 and a terawatt of

battery storage by 2050 I'm not a math

guy but that math doesn't math to me two

obvious questions should arise again is

the actual clean battery storage

projects that are being built are they

actually clean how are you going to

dispose and create these batteries in

the first place coal and natural gas and

the next question question that's

obvious is who's going to pay for it the

Ia will say that we need more batteries

nvidia's CEO Jensen Wong will say that

we need the AI chips to actually be

implemented so that AI can help us

determine whether or not we should have

electricity on it'll help us dynamically

balance the grid that sounds convenient

for someone whose sole job is to sell

more chips Bitcoin mining seems to

provide a real tangible practical and

already implemented solution to how do

we bootstrap these AI projects how do we

profitably build out the infrastructure

to fund this AI Revolution I'm all ears

for feasible Solutions and today I

haven't heard any if Bitcoin mining

isn't the solution to funding our AI

Revolution then what is I'd love to hear

your thoughts in the comments this is

part two of our Bitcoin in AI series I'm

going to be diving deep for the next few

weeks on the implication of Bitcoin as

it relates to an AI future so tune in

next Tuesday on what we're covering next

Bitcoin and AI payments and as always

I'm all about building real Community

Real expertise real connections real

people engaging with you on your Bitcoin

Journey I've had the pleasure for doing

this for hundreds and thousands of

people now and I want to help you do

this through Columbus head to join

columbus.com if you want to find a

Bitcoin Community online to help you

navigate your journey and with that

we're signing off today hope that I was

able to explain Bitcoin simply for you

guys this is Dante cook happy stacking