Create a story

Sustainability

AI uses energy. Here's how we think about it.

Published October 3, 2026 · Last reviewed October 3, 2026

Lil Lanterns was founded by someone trained in sustainability management, so we couldn't build with AI without asking what it costs the planet. This page shares what the research says, the choices we've made because of it, and what we don't know yet.

In short

  • Writing a text story uses a small amount of energy. Making AI images uses far more, so we don't.
  • Our AI provider doesn't publish exact energy figures yet, and we say so plainly.
  • We'll update this page as better data becomes available.

What the research says

  • Data centres are a small but fast-growing share of electricity use. They used about 415 terawatt-hours in 2024, around 1.5% of the world's electricity. The International Energy Agency projects their use will more than double to about 945 terawatt-hours by 2030, with AI as the biggest driver.[1]
  • A single text request uses a small amount of energy. Google reported that a typical text prompt to its Gemini assistant uses about 0.24 watt-hours, about the same as watching TV for under nine seconds.[2] Epoch AI, an independent research group, estimated about 0.3 watt-hours for a typical ChatGPT question.[3] Longer responses use more.
  • Making images takes far more energy than writing text. In a peer-reviewed study that measured many AI models, generating images used over 60 times more energy on average than generating text. The least efficient image model used about as much energy per image as half a smartphone charge.[4]

Choices we've made

  • Stories, not AI-generated pictures. Lil Lanterns writes text stories and doesn't use AI to generate illustrations. Given the energy research above, and our respect for human artists, when we add illustrations they'll come from the independent illustrators we partner with, not from an image generator.
  • No AI narration service. “Read aloud” uses the voice already built into your browser or device. We don't run a separate AI voice service to narrate stories.

What we don't know yet

We want to be straightforward about the gaps. Our stories are written by Claude, an AI model from Anthropic. Anthropic doesn't publish energy or emissions figures for individual requests, so we can't tell you exactly how much energy one Lil Lanterns story uses. Our stories are also longer than a typical chat reply, so they likely use more than the per-prompt figures above.

More about our hosting

Our website is hosted by Vercel. Vercel says it has used Amazon Web Services (AWS) Lambda, the service that runs code like our story generator, since its founding,[7] and its computing regions correspond to AWS regions.[6] Amazon reports that the electricity used in 25 AWS data center regions was matched with renewable energy in 2025.[5] “Matched” means Amazon bought renewable energy equal to what those regions used over the year. It doesn't mean every hour ran on clean power.

What we're committing to

  • Keep Lil Lanterns text-first, and keep AI image generation out of it.
  • Favor AI and hosting providers that publish clear energy and emissions data.
  • Update this page as better data becomes available, and tell you when we do.

Questions or better data we should know about? Email us at hello@lillanterns.com.

Sources

  1. International Energy Agency (2025), Energy and AI: Executive summary
  2. Google Cloud (Aug 21, 2025), Measuring the environmental impact of AI inference
  3. Epoch AI (Feb 7, 2025), How much energy does ChatGPT use?
  4. Luccioni, Jernite & Strubell (2024), Power Hungry Processing: Watts Driving the Cost of AI Deployment? ACM FAccT
  5. Amazon Sustainability, AWS Cloud
  6. Vercel Docs, Global network and regions
  7. Vercel (Apr 5, 2023), Powering a serverless Web: Vercel joins AWS Marketplace

Published October 3, 2026. Last reviewed October 3, 2026.