Archive · 15 Nov 2025
Data Centers: The Unseen Engine Driving Generative AI and LLMs
Imagine Generative AI and chatbots, which can create amazing things like stories or computer code, are like super-smart robots.
These robots need a special, huge building called a data center to work.
Data centers are like the robots' homes and power stations.
They store all the information the robots need to learn and have super-fast computers to help them think.
Training these robots, called LLMs, needs tons of electricity and super-quick connections between computers.
This means data centers have to be bigger and use more power, like a whole city's worth!
They also need special ways to keep cool because all those computers get very hot.
Companies are building more data centers and making them more energy-efficient, even using renewable power.
This is all to make sure our smart AI robots can keep learning and helping us with new and exciting things.
Data centers are the essential infrastructure enabling the capabilities of Generative AI and Large Language Models (LLMs).
Training large AI models requires significant computational power from thousands of GPUs or custom accelerators, demanding sophisticated data center environments with high-speed interconnects.
AI is projected to drive a 165% increase in data center power demand by 2030, necessitating substantial infrastructure expansion and modernization.
Data centers must provide robust storage and processing for vast datasets, with efficient networking crucial for low-latency communication between compute nodes.
The intense energy consumption of AI models is driving innovation in cooling technologies and the adoption of renewable energy sources within data centers.
- Who
- Data center operators, Generative AI developers, LLM developers, Goldman Sachs, Nasscom, CtrlS Datacenters CEO Sridhar Pinnapureddy
- What
- Data centers are identified as the critical, yet often unseen, infrastructure enabling the development and deployment of Generative AI and Large Language Models (LLMs).
- When
- Currently and projected towards 2030, with historical context of the past seven years for India's data center growth.
- Where
- Global AI infrastructure, India (data center capacity expansion and green-certified capacity growth).
- Why
- The immense computational power, data storage, and high-speed networking required for training and running LLMs necessitate significant advancements and expansion in data center capabilities, including power, cooling, and connectivity.
- Projected Data Center Power Demand Increase by 2030
- 165%
- India's Data Center Capacity Expansion (Past 7 Years)
- More than fourfold
- Networking Investment in GenAI Training
- Largest single infrastructure expenditure
- Green-Certified Data Center Capacity Growth in India
- 15% CAGR
