A large building full of AI chips that trains models and answers questions, and needs a huge amount of electricity and cooling.
AI Drafted with AI help · Reviewed by Keshav Aggarwal · Updated
In simple words
Imagine a giant kitchen that cooks answers for millions of people at once. It needs thousands of stoves, a huge power line and lots of cooling so nothing overheats.
An AI data centre is that kitchen for AI: rows of computer chips that train models and answer questions.
The technical side
An AI data centre houses large numbers of GPUs or other AI chips, with the power supply, networking and cooling they need for training and running (inference) large models. Its demand is measured in megawatts, so access to power and grid capacity shapes where sites are built.
On 4 November 2026, Bitcoin miner Riot Platforms announced a 15-year, $28bn lease of 910MW at its Corsicana, Texas, site to an AI tenant, reusing power infrastructure first built for mining. OpenAI launched a Grid Protectors Program so future sites can cut power use at peak times, and voters in at least 29 US jurisdictions backed limits on new data centres. More on our AI news feed.
AI Data Centre kya hai?
Ek bada building jismein hazaaron AI chips din-raat chalte hain. ChatGPT jaise tools ka har jawab yahin banta hai. Isme bahut bijli aur cooling lagti hai, isliye log iski jagah aur bijli ke kharch par sawaal uthate hain.
Related terms
- GPU (Graphics Processing Unit)AIA computer chip that does thousands of small calculations at once. First made for video games, it now trains and runs most AI models.
- InferenceAIThe moment an AI model is used to answer: it takes your input and works out the output. Each answer you get is one inference.
- LLM (Large Language Model)AIAn AI model trained on huge amounts of text so it can read and write language, like the models behind ChatGPT, Claude and Gemini.
- Crypto MiningCryptoUsing powerful computers to secure a proof-of-work blockchain like Bitcoin and earn newly created coins and fees in return.
