\n
![](\"https:\/\/fortune.com\/img-assets\/wp-content\/uploads\/2024\/04\/53656347289_79e4b94620_o-e1713261654314.jpg?w=2048\")
<\/p>\n [ad_1]
\n<\/p>\n
Before artificial intelligence can transform society, the technology will first have to learn how to live within its means.<\/p>\n
Right now generative AI have an \u201cinsatiable demand\u201d for electricity to power the tens of thousands of compute clusters needed to operate large language models like OpenAI\u2019s GPT-4, warned chief marketing officer\u00a0Ami Badani<\/a>\u00a0from\u00a0chip design<\/a> company Arm Holdings.\u00a0<\/p>\n If generative AI is ever going to be able to run on every mobile device from a laptop and tablet to a smartphone, it will have to be able to scale without overwhelming the electricity grid at the same time.<\/p>\n \u201cWe won\u2019t be able to continue the advancements of AI without addressing power,\u201d Badani told Fortune\u2019s Brainstorm AI conference in London on Monday. \u201cChatGPT requires 15 times more energy than a traditional web search.\u201d\u00a0<\/p>\n Not only are more businesses using generative AI, but the tech industry is in a race to develop new and more powerful tools that will mean compute demand is only going to grow\u2014and power consumption with it, unless something can be done.\u00a0<\/p>\n The latest breakthrough from OpenAI, the company behind ChatGPT, is\u00a0Sora<\/a>. It can create super realistic or stylized clips of video footage up to 60 seconds in length purely based on user text prompts.\u00a0<\/p>\n \u201cIt takes a 100,000 AI chips working at full compute capacity and full power consumption in order to train Sora,\u201d Badani said. \u201cThat\u2019s a huge amount.\u201d\u00a0<\/p>\n Data centers, where most AI models are trained, currently account for 2% of global electricity consumption, according to Badani. But with generative AI expected to go mainstream, she predicts it could end up devouring a quarter of all power in the United States in 2030.<\/p>\n The solution to this conundrum is to develop semiconductor chips that are optimized to run on a minimum of energy.<\/p>\n That\u2019s where Arm comes in: its\u00a0RISC<\/a>\u00a0processor designs currently run on\u00a099% of all smartphones<\/a>, as opposed to the rival x86 architecture developed by Intel<\/a>. The latter has been a\u00a0standard for desktop PCs<\/a>, but proved too inefficient to run battery-powered handheld devices like smartphones and tablets.\u00a0<\/p>\n Arm is adopting that same design philosophy for AI.<\/p>\n \u201cIf you think about AI, it comes with a cost,\u201d Badani said, \u201cand that cost is unfortunately power.\u201d\u00a0\u00a0<\/p>\n<\/div>\nThe marvel of GenAI comes at a steep cost<\/h2>\n
Subscribe to the Eye on AI newsletter to stay abreast of how AI is shaping the future of business. Sign up<\/a> for free.<\/div>\n<\/div>\n