How AI’s thirst for energy is shaking up the IT sector’s sourcing strategy

Transform 2024 Back This July! On 400 business leaders will gather In san Francis Since July 9-11 has dive In THE advancement of GénAI strategies And engaging In which makes you think discussions In THE community. Find out how You can attend here.

A new white paper released last week by THE Electric Power Research Institute (EPRI) quantify THE exponential growth potential of AI power requirements. THE 35 pages report title, " Feed Intelligence: Analyzing Artificial Intelligence And Data Center Energy Consumption," projects total data center power consumption by WE. data centers alone could more that double has 166% by 2030.

According to has EPRI, THE request East be led largely by generative AI, which can require exponentially more power by request that traditional research. Notably, It is not including pictures And other rich content: "HAS 2.9 watt hours by ChatGPT request, AI queries are estimated has require 10x THE electricity of traditional Google queries, which to use about 0.3 watt hours each; And emerging, computationally intensive abilities such as picture, audio, And video generation to have No previous."

EPRI Energy To use by Model Picture Credit: EPRI Report

THE report studies five different to use case : Google research, ChatGPT, BLOOM And Powered by AI Google research. Among these, ChatGPT was THE less energy consuming of THE AI-based requests. However, THE researchers anticipated THE the integration of Google AI abilities In Google Research, noting he could be on 3x upper that ChatGPT: "If Google integrated similar AI In It is researches, THE electricity by research could increase has between 6.9 to 8.9 What."

Transform 2024 Registration East Open

Join business leaders In san Francis Since July 9 has 11 For A exclusive AI event. Connect with peers, explore THE opportunities And challenges of Generative AI, And learn how has to integrate AI apps In your industry. Register Now

A emerging provide constraint

EPRI developed four distinct forecasts For potential electricity use In WE. data centers between 2023 And 2030, base on miscellaneous annual growth scenarios: weak (3.7%), moderate (5%), high (ten%), And upper (15%). Below THE upper growth scenario, data center electricity use could increase has 403.9 TWh/year by 2030, A 166% increase Since 2023 levels. Even THE weak growth scenario projects A 29% increase has 196.3 TWh/year.

EPRI WE. Data Center Energy Consumption Projection 2023-20230 Picture Credit: EPRI Report

THE unequal geographical distribution of This growth creates localized challenges. Fifteen States accounted for For 80% of THE national data center load In 2023, with Virginia alone including 25%. Projection to show that of Virginie data center share of total electricity consumption could reach 46% by 2030 below THE upper growth scenario. Other States as Oregon, Iowa, Nebraska, North Dakota And Nevada are Also projected has to have data centers to understand 20% Or more of total electricity request.

Different types of data centers are contributing has This growth. Business data centers, possesses And exploited by individual companies For their own to use, account For 20-30% of THE total load. Collocation centers, Or companies to rent out sharing space And Infrastructure, And hyperscale centers built by cloud giants as Amazon, Google And Microsoft, together account For 60-70% of load. Hyperscale centers In particular are has THE Before...

How AI’s thirst for energy is shaking up the IT sector’s sourcing strategy

Transform 2024 Back This July! On 400 business leaders will gather In san Francis Since July 9-11 has dive In THE advancement of GénAI strategies And engaging In which makes you think discussions In THE community. Find out how You can attend here.

A new white paper released last week by THE Electric Power Research Institute (EPRI) quantify THE exponential growth potential of AI power requirements. THE 35 pages report title, " Feed Intelligence: Analyzing Artificial Intelligence And Data Center Energy Consumption," projects total data center power consumption by WE. data centers alone could more that double has 166% by 2030.

According to has EPRI, THE request East be led largely by generative AI, which can require exponentially more power by request that traditional research. Notably, It is not including pictures And other rich content: "HAS 2.9 watt hours by ChatGPT request, AI queries are estimated has require 10x THE electricity of traditional Google queries, which to use about 0.3 watt hours each; And emerging, computationally intensive abilities such as picture, audio, And video generation to have No previous."

EPRI Energy To use by Model Picture Credit: EPRI Report

THE report studies five different to use case : Google research, ChatGPT, BLOOM And Powered by AI Google research. Among these, ChatGPT was THE less energy consuming of THE AI-based requests. However, THE researchers anticipated THE the integration of Google AI abilities In Google Research, noting he could be on 3x upper that ChatGPT: "If Google integrated similar AI In It is researches, THE electricity by research could increase has between 6.9 to 8.9 What."

Transform 2024 Registration East Open

Join business leaders In san Francis Since July 9 has 11 For A exclusive AI event. Connect with peers, explore THE opportunities And challenges of Generative AI, And learn how has to integrate AI apps In your industry. Register Now

A emerging provide constraint

EPRI developed four distinct forecasts For potential electricity use In WE. data centers between 2023 And 2030, base on miscellaneous annual growth scenarios: weak (3.7%), moderate (5%), high (ten%), And upper (15%). Below THE upper growth scenario, data center electricity use could increase has 403.9 TWh/year by 2030, A 166% increase Since 2023 levels. Even THE weak growth scenario projects A 29% increase has 196.3 TWh/year.

EPRI WE. Data Center Energy Consumption Projection 2023-20230 Picture Credit: EPRI Report

THE unequal geographical distribution of This growth creates localized challenges. Fifteen States accounted for For 80% of THE national data center load In 2023, with Virginia alone including 25%. Projection to show that of Virginie data center share of total electricity consumption could reach 46% by 2030 below THE upper growth scenario. Other States as Oregon, Iowa, Nebraska, North Dakota And Nevada are Also projected has to have data centers to understand 20% Or more of total electricity request.

Different types of data centers are contributing has This growth. Business data centers, possesses And exploited by individual companies For their own to use, account For 20-30% of THE total load. Collocation centers, Or companies to rent out sharing space And Infrastructure, And hyperscale centers built by cloud giants as Amazon, Google And Microsoft, together account For 60-70% of load. Hyperscale centers In particular are has THE Before...

What's Your Reaction?

like

dislike

love

funny

angry

sad

wow