Amanda McDermott Posted January 28 Share Posted January 28 Hi all, We're thinking about AI use and environmental impacts within our business; does anyone know how to quantify the carbon footprint of the AI visualizer per use? We would also be interested to know if it uses Vectorworks' server to generate from, or if it is making use of a larger data centre? Our use of the AI visualizer to date has been limited, and of limited success for the outputs we would wish to generate - particularly if (as we suspect) it has a significant carbon footprint, we might just give up on it. 2 Quote Link to comment
Vectorworks, Inc Employee Dave Donley Posted January 28 Vectorworks, Inc Employee Share Posted January 28 Hello @Amanda McDermott the cloud is someone else's computer. In this case the AI images are generated on our servers in Maryland, and/or Amazon's servers if ours are full. It is the same as if you were to do a rendering with your local PC. The servers have pretty good GPUs in them but aren't more power hungry than high end consumer versions. These servers are used for cloud rendering or AI Viz jobs, when needed. The image generation takes 30 seconds or less per image. HTH 2 Quote Link to comment
Amanda McDermott Posted February 5 Author Share Posted February 5 Thanks @Dave Donley, I'm not quite sure if it answers the question... are you saying that the carbon and water consumption footprint of using the AI visualiser is no greater than doing a 'normal' offline rendering using Vectorworks? Quote Link to comment
Vectorworks, Inc Employee Dave Donley Posted February 5 Vectorworks, Inc Employee Share Posted February 5 Yes, it is a shared computer and when you have an image to generate it does that, just like a local PC would. Our PCs are turned on but idle if they don't have jobs, and they are started and stopped depending on the load. If the load is greater than what our servers can handle we use Amazon's servers. Quote Link to comment
line-weight Posted February 5 Share Posted February 5 I'm interested in this too. My understanding is that the carbon footprint is pretty hard to pin down because it depends on the energy source for the relevant servers and data centres. And a lot of the ones being built at present are using "dirty" sources like gas power plants. And then there is the question of whether you only take into account the energy used by your specific query or a portion of the energy used to generate the model it uses. A reasonable round-up here https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/ Surprisingly it seems that image generation may be less bad than text generation. 1 Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.