meta_pixel
Tapesearch Logo
Log in
On Point with Meghna Chakrabarti

How will we power the AI boom?

On Point with Meghna Chakrabarti

WBUR

On Point, News, Daily, Talk Show, Npr

4.33.9K Ratings

🗓️ 13 February 2025

⏱️ 50 minutes

🧾️ Download transcript

Summary

Just one query to ChatGPT takes as much energy as powering a lightbulb for some 20 minutes. As AI balloons, so does its demand for electricity. That's driving tech companies to increase their demand for fossil fuels.

*** Thank you for listening. Help power On Point by making a donation here: wbur.org/giveonpoint

Transcript

Click on a timestamp to play from that location

0:00.0

Support for this podcast comes from All About Change, a podcast with firsthand stories from

0:05.9

activists who are changing the world, like Coda actor Troy Kotzer, V. Aaron Brockovich, and

0:12.0

disability activist Shane Burkaw. Listen to stories that give us all hope on All About Change,

0:18.5

wherever you get your podcasts.

0:26.9

Support for WBUR comes from MathWorks, creator of MATLAB and Simulink Software for technical computing and model-based design. MathWorks, accelerating the pace of discovery

0:32.4

in engineering and science. Learn more at MathWorks.com.

0:37.7

This is on point. I'm Megna Chakra-Bardi.

0:40.3

And today we're going to start off with a question for chat GPT.

0:45.0

Got a window open right here in front of me, chatgpt.com.

0:48.6

And this is live radio, so I'm not 100% sure this is going to work, but we're going to try it out.

0:52.9

So here's my query for chat GPT. How much electricity does chat? You guys are now finding out how

1:00.6

a slova typeer I am. Chat GPT use per query compared to an incandescent light bulb with a B.

1:12.4

How much electricity does chat GBT use per query compared to an incandescent light bulb?

1:16.9

Enter.

1:18.1

It is thinking.

1:18.9

Oh, here it goes.

1:19.8

The amount of electricity chat GPT uses per query is relatively low, but calculating it directly can be complex.

1:30.3

It depends on a number of factors. However, the bot says, let's make some general comparison runs on powerful cloud servers. When

1:37.3

a user sends a query, it goes through multiple layers of processing from the network and storage

1:42.3

to the machine learning model itself.

1:52.5

The extent of energy consumption for one query, okay, here's finally the answer, is likely in the range of a few watt seconds.

2:08.2

For context, one study suggested that a large model like GPT3 could consume 0.5 to 1.5 kilowatt hours for training. Okay. Wow. So this is not actually answering my question.

...

Please login to see the full transcript.

Disclaimer: The podcast and artwork embedded on this page are from WBUR, and are the property of its owner and not affiliated with or endorsed by Tapesearch.

Generated transcripts are the property of WBUR and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.

Copyright © Tapesearch 2026.