Satya Nadella says Microsoft could lose “social permission” to burn electricity for AI unless it offers something useful — “That, to me, is ultimately the goal”

by

Generative AI uses a huge amount of power and water for cooling. Recent reports have shown just how much energy it takes to run these large, expensive data centers.

OpenAI CEO Sam Altman says that each response from ChatGPT uses a small amount of energy – 0.34 watt-hours. To put that in perspective, it’s roughly the same energy an oven uses in just over a second, or a high-efficiency lightbulb uses in a few minutes.

Interestingly, recent research shows that earlier GPT models, like OpenAI’s GPT-3, use significantly more water than initially estimated. Even the newer GPT-4 can use as much water as three water bottles just to create 100 words of text.

Microsoft and Google together use so much electricity that their combined power consumption is higher than that of over 100 countries. Microsoft CEO Satya Nadella recently stated at the World Economic Forum in Davos that the cost of energy will be a major factor in determining which nation leads the way in artificial intelligence development.

If these digital tokens don’t demonstrably improve things like health, education, and efficiency in both the public and private sectors – for businesses of all sizes – we’ll quickly lose public support for using valuable resources like energy to create them. Ultimately, that’s what we should be aiming for: real-world benefits.

Microsoft CEO, Satya Nadella

The executive explained that economic growth worldwide will depend heavily on the amount of energy used by artificial intelligence. He also described “tokens” as the new essential global resource – these are essentially credits that people purchase to use AI chatbots and other AI models.

As I understand it, Nadella was emphasizing that the core purpose of any economy, and each business within it, is to convert resources – what he called ‘tokens’ – into actual economic growth. He also pointed out that having a lower-cost product is always a significant advantage.

There’s currently a worldwide shortage of DRAM memory, causing prices to rise sharply. Major AI research companies have been building large data centers globally, significantly increasing the need for these components. When demand exceeds supply, prices naturally go up, and that’s what we’re seeing happen now.

Microsoft’s CEO believes AI companies need to create a widespread system for both energy and digital transactions. He also stressed the importance of businesses incorporating AI into their daily operations, describing it as a tool that boosts human capabilities and provides access to a vast pool of knowledge.

Read More

2026-01-21 15:41