Why AI Needs Massive Data Centers

Added: 2026-05-17

[00:00:00]running your code. So, you send your prompt, it travels to somebody's servers over the internet, which by the way means that there has to be an extent of security involved. Then it gets processed on specialized hardware in a data center, and then the responses come back. So, specialized hardware is a keyword here, and data center as well.

[00:00:20]So, a data center is a collection of specialized hardware. So, that's why these companies are investing billions and billions on these data centers and on the energy production infrastructure that is needed to run the data center, and particularly the air conditioning to keep them cool. And when you do that, um you never touch the weights in that you can't tailor the model you're running um with you run exactly what they are providing you with, which is usually really advanced stuff. But, it's not something you can tweak too much. And you and I discussed in previous lives how your chosen large language model will get used to how you work. So, there will be a bit of personalization, but you are still not changing the the substance, the essence of the model.

[00:01:09]And um as a result of all these things also, you can't really download a copy of those models for reasons of intellectual property, but also because it would never fit on your machine. So, because those models are so big, they are much more capable of what you can run locally. But, yeah, even if you had the space, your hardware wouldn't be able to do that.