DesignLed uses AI, and that comes with an environmental cost. AI infrastructure consumes electricity and water and relies on resource-intensive data-centre hardware. We can't eliminate that impact, and current data doesn't allow us to measure our exact share with perfect accuracy.
We therefore take a two-part approach: reducing unnecessary AI compute wherever we can, and building environmental funding into the cost of using DesignLed, with a portion of revenue from platform transactions going towards measurable water and ecological restoration.
As our ability to measure DesignLed's environmental footprint improves, we'll continue to adjust what we contribute and publish how the money is being used.
We don't claim that these measures make AI impact-free. They don't. Our aim is to minimise the resources DesignLed uses and make a meaningful, measurable contribution towards repairing the environmental impact associated with running it.
1. What are we doing to reduce DesignLed's own impact?
-
DesignLed is being built from the outset with resource use in mind. As we develop the platform, we continually assess how each part of the process works and look for ways to achieve the same result with less computing power, less storage and fewer unnecessary AI calls.
That means AI isn't the default way of doing everything. Much of the design process happens without generative AI, using reusable 3D assets, structured product data and DesignLed's own design rules. AI is reserved for the parts of the experience where it genuinely adds value — such as creating a photorealistic view of a design.
We're also designing the platform to avoid unnecessary processing and duplication. Saved rooms, for example, store a lightweight record of a user's design rather than repeatedly storing copies of the underlying products, models and assets.
This isn't a one-off exercise. As DesignLed evolves, we'll continue to review how the platform works, measure what we can, and look for more efficient ways of delivering the same experience with fewer resources.
I particularly like that last paragraph because it establishes continuous assessment as part of how you build the company, rather than claiming you've already found the environmentally optimal technical solution.
2. Contributing to environmental restoration
-
Environmental funding is built into our financial model. A portion of the revenue generated through the platform will be set aside to support carefully selected projects working on water and ecological restoration.
We're starting small and deliberately: initially supporting one project in Ireland and one in the UK, so that we can understand where the money goes, how it is used and what measurable difference it makes before expanding our support.
We will prioritise projects where funding contributes to practical work on the ground and where the resulting environmental outcomes can be reported clearly.
This isn't an offset and we won't describe it as one. Supporting restoration elsewhere doesn't erase the resources DesignLed consumes. It's an additional commitment alongside reducing our own impact.
3. Measuring and reporting what we do
-
Measuring the environmental impact of an individual AI service is currently difficult. The information available about energy, water and infrastructure use doesn't allow us to calculate DesignLed's exact footprint with the precision we'd like.
We won't manufacture certainty where it doesn't exist.
Instead, we'll use the best information available to us, improve our measurement as better data becomes available, and publish what we can measure.
That will include how much DesignLed has contributed to environmental projects, where the money has gone and the outcomes reported by the organisations receiving it. As our ability to measure the platform's own resource use improves, we'll publish that information too.
If better evidence shows that we should be doing more, we'll adjust our contribution accordingly.