Before artificial intelligence can transform society, technology will first have to learn to live within its means.
Right now, generative AI has an “insatiable demand” for electricity to power the tens of thousands of computing clusters needed to run large language models like OpenAI’s GPT-4, the director warned of marketing. Friend Badani Since chip design Arm Holdings company.
If generative AI is ever to be able to run on every mobile device, from a laptop and tablet to a smartphone, it will need to be able to scale without overloading the power grid at the same time.
“We won’t be able to continue progress in AI without addressing the question of power,” Badani said at Fortune’s Brainstorm AI conference in London on Monday. “ChatGPT requires 15 times more energy than a traditional web search.”
Not only are more companies using generative AI, but the tech industry is in a race to develop new, more powerful tools that will drive an increase in computing demand – and energy consumption with it , unless nothing can be done.
The latest advancement from OpenAI, the company behind ChatGPT, is Sora. It can create super realistic or stylized clips of video footage up to 60 seconds long based solely on user text prompts.
GenAI’s wonder comes at a high cost
“It takes 100,000 AI chips operating at full computing capacity and full power consumption to train Sora,” Badani said. “It’s a huge amount.”
Data centers, where most AI models are trained, currently account for 2% of global electricity consumption, according to Badani. But as generative AI is expected to become more widespread, she predicts it could end up devouring a quarter of all energy in the United States by 2030.
The solution to this conundrum is to develop semiconductor chips optimized to operate with minimal energy.
This is where Arm comes in: it’s RISK processor designs currently run on 99% of all smartphonesas opposed to the rival x86 architecture developed by Intel. The latter was a standard for desktop computersbut proved too inefficient for operating battery-powered portable devices like smartphones and tablets.
Arm takes the same design philosophy to AI.
“If you think about AI, it has a cost,” Badani said, “and that cost unfortunately is energy.”
