AI could consume a quarter of US electricity by 2030: ARM warning
Here's your revised text in Russian:
ARM chief warns that artificial intelligence could consume a quarter of all US electricity by 2030, and calls for greater efficiency and regulation of the industry .
AI is a growing consumer
The threat of rising energy consumption
Rene Haas, the head of ARM, said that by 2030 artificial intelligence could consume a quarter of all electricity generated in the United States. Currently, AI data centers consume 4% of electricity consumption. However, without significant increases in energy efficiency and/or increased government regulation, the current AI development trend will not be possible, Haas believes.
High energy consumption of ChatGPT
According to January estimates from the International Energy Agency, processing a single ChatGPT request requires 2.9 Wh of electricity, which is ten times higher than a regular Google search. Thus, with a full transition to AI, Google would have to spend at least 11 TWh of energy annually instead of the current 1 TWh.
OpenAI developments and energy costs
Haas's remarks came on the eve of the expected cooperation agreement between the United States and Japan in the field of artificial intelligence, as well as recent OpenAI developments such as Sora. According to Factorial Funds, the current version of Sora requires using at least one NVIDIA H100 GPU for an hour to generate a five-minute video. Energy consumption of one 700-watt H100 processor reaches about 3740 kWh per year.
US Energy Production and Consumption
The US Energy Information Administration reported that the United States produced 4,240 TWh of electricity in 2022, of which only 22% came from renewable sources. At the same time, consumption amounted to 3900 TWh.
Energy growth and regulation
Achieving sustainable development requires taking into account the likely growth in demand for electricity from other industries, as well as the extent to which renewable and non-renewable resources are being used. Taking into account the almost doubling of the cost of electricity since 1990, stricter government regulation of this area seems justified.
Glossary
- ARM is a British company that develops processor architectures for mobile devices and embedded systems.
- AI (artificial intelligence) is a system capable of imitating human thinking and behavior.
- ChatGPT is an artificial intelligence chatbot created by OpenAI.
- OpenAI is an American company engaged in research in the field of artificial intelligence.
- NVIDIA H100 is a graphics processor used in artificial intelligence computing.
Links
- International Energy Agency Report
- Factorial Funds data on Sora energy consumption from OpenAI
- Statistics on changes in the cost of electricity in the United States
- Source: Tom's Hardware
Questions answered
What is an advertising platform for brands and agencies?
What features does this platform provide?
How does the platform help manage advertising campaigns?
What are the benefits of using this platform?
How is data security and privacy ensured on the platform?
Hashtags
Save a link to this article
Discussion of the topic – AI could consume a quarter of US electricity by 2030: ARM warning
ARM CEO Rene Haas said that by 2030, artificial intelligence could consume up to 25% of all US electricity. This is due to the growing popularity of AI and the increasing amount of data required to train it.
Latest comments
15 comments
Write a comment
Your email address will not be published. Required fields are checked *
Сергей
Damn, these numbers are just scary! 🤯 AI power consumption looks unsustainable in the long term. We must look for more efficient solutions so as not to burden the planet.
Франческа
I agree, the situation is alarming. But we shouldn’t forget about the incredible opportunities that AI opens up for humanity. 💻 It is important to find a balance between progress and sustainable development.
Ганс
What if we focused our efforts on developing renewable energy sources? 🌞 Then AI consumption will not be so critical for the environment.
Ольга
Great idea, Hans! Investments in green energy could be the key to solving the problem. In addition, it will create new jobs and boost the economy. 💹
Джованни
Let's not forget that AI helps us in various fields - medicine, science, education. 🧠 Perhaps technological progress will reduce energy costs in the future.
Януш
By the way, have you heard about the OceanMind company? 🌊 They are developing underwater data centers that use cold sea water for cooling. An innovative approach to reducing energy consumption!
Грумпи
Again these fashionable things... 🙄 It would be better to do something useful, and not waste resources on entertainment for gamers and talkers.
Анна
Don't be such a pessimist, Grumpy! 😄 AI really opens up a lot of possibilities. Take a look at ChatGPT - it helps people learn, create and work.
Пётр
I've heard that Midjourney and Stable Diffusion use relatively little power to generate images on demand. 🖼️ Maybe we should focus on such energy-efficient AI applications?
Елена
By investing in AI, we are investing in the future of humanity. 🌐 Of course, environmental consequences must be taken into account, but progress is inevitable. Let's work towards the sustainable development of these technologies!
Жан-Люк
Did you know that the Japan Science and Technology Agency is developing a computer with very low power consumption? 💻⚡ Maybe cooperation with the US will help speed up these developments.
Софи
It seems to me that part of the solution may lie in the use of cloud computing. 🌩️ If you combine resources, you can achieve greater efficiency and energy savings.
Мартин
Don't forget that thanks to AI we can optimize many processes and reduce emissions. 🏭 For example, smart cities and transport will help reduce the carbon footprint.
Наталия
Great discussion! 👏 It is clear that AI is a multifaceted topic that requires an integrated approach. Let's join forces and find a reasonable balance between innovation and ecology.
Виктор
I completely agree! 👍 Only by collaborating and taking into account different points of view can we build a truly sustainable future for AI and all humanity.