Breaking

Sam Altman Compares One California Almond’s Water Use With Thousands of ChatGPT Queries

September 2, 2026 at 14:53 EDT

Foundation Models

Sam Altman has compared the water used to grow a single California almond with thousands of ChatGPT queries, putting a familiar food item at the center of the AI industry’s resource debate. The OpenAI CEO made the comparison in the launch episode of Alex Heath’s Sources podcast, released September 1, 2026.

September 2, 2026 · OpenAI

One almond can equal thousands of ChatGPT queries

Sam Altman used California almond farming to frame a debate over the water cost of AI prompts. The size of the comparison depends on how each footprint is counted.

The scale in one glance

0.32 mL

Estimated water per ChatGPT query

3.56–12 L

Published California almond estimates

~11,000–37,500

Equivalent queries on those figures

The almond estimate has its own range

Equivalent ChatGPT queries, using 0.32 mL per query

Each block represents 1,000 equivalent queries; the one-query baseline is too small to render at this scale.

Why the comparison is useful—and incomplete

Useful scale check

A per-query estimate is tiny in isolation, and an everyday agricultural comparison makes that scale easier to understand.

Not a complete footprint

Indirect electricity water, hardware, training, local scarcity, model choice, and cooling design can change the result.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.