- OpenAI’s CEO, Sam Altman, is actively working to establish a global AI alliance, engaging with governments and industry leaders to address concerns beyond chip production, including energy and data centre capacity.
- Recent meetings in the UAE and upcoming discussions in Washington underscore Altman’s heightened efforts to tackle the challenges of AI deployment outpacing critical infrastructure supply.
- Concerns about AI’s high energy consumption have been raised by industry leaders like Altman, with data showing significant electricity usage by AI models like ChatGPT and potential future challenges in power supply and infrastructure development.
OpenAI‘s CEO Sam Altman has shown signs he is dedicated to establishing a global AI alliance between governments and industry leaders, with topics extending beyond chip production to encompass energy, data centre capacity, and other aspects.
Also read: Sora won’t replace humans, and here’s why
Also read: Artists vs AI: Who will win the copyright fight of the century?
Altman wants support for large-scale AI infrastructure projects
Altman met investors and government officials in the UAE this week to discuss how the private sector can collaborate with governments to support expensive large-scale AI infrastructure projects. Furthermore, Altman engaged with officials from Western countries and is scheduled to hold meetings in Washington later this week, indicating his increased efforts and concerns about the supply of chips and other critical infrastructure unable to keep up with the rapid deployment of AI.
OpenAI believes AI systems require substantial energy, posing one of the biggest infrastructure challenges in the tech industry. Prior discussions by Altman emphasised the need for an energy breakthrough to support AI development, with solar power and nuclear fusion among the options.
Additionally, recent conversations between a co-founder and CEO of AI startup OpenPipe, Kyle Corbitt, and a Microsoft engineer responsible for the GPT-6 training cluster project revealed challenges in deploying infiniband-level links between GPUs across regions, reflecting Altman’s urgency in addressing AI’s energy issues.
ChatGPT consumes over 500,000 kilowatt-hours of electricity daily
Data shows that ChatGPT consumes over 500,000 kilowatt-hours of electricity daily, processing around 200 million user requests, equivalent to more than 17,000 times the daily electricity consumption of U.S. households. Meanwhile, if the search giant Google were to call AIGC in each user search, its annual electricity consumption would increase to about 29 billion kilowatt-hours, surpassing the annual electricity consumption of countries like Kenya and Guatemala.
As AI development progresses, several prominent figures in the industry have raised concerns about its high energy consumption. For instance, Arm CEO Rene Haas estimated in an interview this week that by 2030, AI data centre power consumption could account for 20%-25% of U.S. electricity demand, compared to the current proportion of 4% or less.
Elon Musk predicted a shortage of electricity
Elon Musk also predicted a shortage of electricity, stating that AI’s energy consumption is astonishing and could potentially limit AI development in the coming years. Bill Gates emphasised the critical role of electricity in determining the profitability of data centers and highlighted the staggering electricity consumption of AI.
These concerns have manifested in reality, with the construction of new data centres outpacing the construction of new power plants, leading to a supply-demand gap. According to CBRE Group, Inc., a commercial real estate services company, the construction time for data centres has been extended by 2 to 6 years due to delayed power supply.
