The Uncomfortable Truth About AI's Carbon Footprint (And What We're Doing About It)

Let's start with a truth that most AI companies prefer to keep in the footnotes: AI is not a cloud. It's a factory.

Every time you ask a large language model (LLM) to write an email, summarize a document, or generate an idea, a chain reaction of computation kicks off in a massive, power-hungry data center somewhere in the world. These data centers are filled with thousands of specialized processors that consume enormous amounts of electricity and, in many cases, vast quantities of water for cooling.

A single query can be the energy equivalent of charging your phone. Training a new flagship model can have a carbon footprint comparable to hundreds of transatlantic flights.

For a company like Earthly Insight, whose entire mission is to support ecological restoration, this presents a profound and uncomfortable paradox. How can we build a business on a technology that contributes to the very problem we're trying to solve?

This isn't a question we shy away from. It's the question that shapes our entire product philosophy.

Acknowledging the Cost: Beyond Carbon Tunnel Vision

The environmental cost of AI isn't just about CO₂ from electricity. It includes:

• Water Consumption: Data centers use millions of gallons of water for evaporative cooling, often in water-stressed regions. A recent study estimated that a typical conversation with a chatbot could "cost" a 500ml bottle of water.
• Hardware & E-Waste: The rapid advancement of AI requires a constant cycle of manufacturing and discarding specialized hardware, which has its own complex supply chain and environmental footprint.
• The Rebound Effect: As AI becomes more efficient, we tend to use it more, potentially increasing total energy consumption even as the cost per query drops.

We can't solve a problem we're unwilling to measure or acknowledge. So, our first commitment is to be honest about this reality. We are not building a "green" or "eco-friendly" AI. That doesn't exist today.

Instead, we are building a more responsible AI, one that intentionally works to do less harm and contribute more good.

Our Strategy: Doing Less Harm by Design

Most AI platforms are in an arms race for bigger, more complex features. We’re in a race for smarter, more efficient impact. Here’s how we're tackling our own footprint:

1. We Are Text-First, by Design.The most energy-intensive AI tasks are often visual. Generating a single image can require as much energy as hundreds or even thousands of text queries. That's why Earthly Insight is—and will remain—a text-based platform. We have intentionally chosen not to build features like image generation. It's a product decision rooted in our commitment to minimizing our compute, energy, and water usage.

2. We Integrate, We Don't Train (For Now).Training a foundational LLM from scratch is the most resource-intensive part of the AI lifecycle. We don't do it. Instead, we use a multi-model approach, integrating leading third-party models (like those from OpenAI, Anthropic, Google, and others). This allows us to select the most efficient model for the job and leverage the billions of dollars they have invested in optimizing performance. Our long-term goal is to secure funding to build our own smaller, highly efficient LLM tuned specifically for our users' needs, which would dramatically lower our energy use per query.

3. We Believe in Financial Transparency and Direct Action.We can't claim our platform "offsets" its own footprint. That math is murky at best. Instead, we embrace a simpler, more honest model: we use our revenue to fund the people already doing the work.

We donate 33% of our subscription revenue directly to our partner, the Global Rewilding Alliance. This isn't a carbon credit scheme; it's a direct investment in restoring the planet's natural systems. Rewilding is the most holistic, high-impact approach to fighting climate change and biodiversity loss, and our model ensures that as our platform grows, so does our ability to fund this critical work.

A Step in the Right Direction

Building an AI company while caring deeply about the planet is a constant exercise in navigating complexity and embracing imperfection.

We can't claim to have solved the environmental impact of AI. No one has. But we can promise to be transparent about the trade-offs, to design our product with intention, and to use our business as a vehicle for tangible, measurable restoration.

Using Earthly Insight is not an act of environmental purity. It's a vote for a different kind of tech company, one that acknowledges its impact and takes real, structural steps to fund a healthier planet.

It's a step in the right direction. And we're glad to be taking it with you.

‍