Environmental impact

Ask Me Anything Q&A Speeroamaquestionrepo

Topic summary

A user raises concerns about the environmental impact of AI usage, specifically regarding energy consumption and water usage for cooling data centers. They ask whether Shopify and Speero have strategies to mitigate environmental damage, details about data center ownership, power sources, and geographic locations (including wildfire risk).

Key questions raised:

  • Is AI necessary for all users?
  • Can it be disabled when not needed?

Response highlights:
A community member (not directly affiliated with infrastructure decisions) suggests:

  • Using AI intentionally rather than as an “always on” default
  • Reducing unnecessary prompts and model usage
  • Minimizing data footprints by limiting logging and storage
  • Considering local LLM options like Ollama (though 24/7 operation may negate environmental benefits)

Status: The discussion remains open with no official response from Shopify or Speero regarding their specific environmental policies or data center practices.

Summarized with AI on October 30. AI used: claude-sonnet-4-5-20250929.
1

What is Shopify, and Speero, doing to lessen the environmental impact of the use of AI? These computations take huge amounts of energy to produce and water to cool the data centers. Has there been any focus in the work to mitigate or eliminate this environmental damage? Do you own these data centers or do they belong to another company? What kind of power grid do they work on? Where are the data centers located geographically? Is there any danger of these data centers being in a wildfire zone?

Is this use of AI necessary for all users? Is there a way to disable it if not?

1 Like
2

To clarify, I’m not directly involved with Shopify or Speero’s infrastructure decisions, but I can speak to the broader concerns you’re highlighting.

AI shouldn’t be a default “always on”. Use it intentionally and know what you want to get out of it.

Be intentional with usage: Not every task needs AI.

Reducing unnecessary prompts, training, or model use can cut down energy waste.

Minimise data footprint: Large-scale logging, unnecessary image generation, or storing outputs indefinitely all add up.

Being conscious of what data is collected and how long it’s kept makes a difference.

You can still get the benefits of LLMs, but run them locally with options like https://ollama.com/ (however if you’re running it 24/7 any gains may be negligible)

Paul

Visit our website: https://speero.com/

Follow me on LinkedIn: https://www.linkedin.com/in/paulrandall/