Reliance Buys Advance NVIDIA AI Chips for World's Largest Data Centre - Report

According to a Bloomberg report, Mukesh Ambani's Reliance Industries is planning to build what could become the world's largest data centre in Jamnagar, Gujarat. The facility is expected to have a capacity of 3 GW, significantly surpassing the current largest data centre, Microsoft's 600-megawatt site in Virginia. The project could cost between $20 billion and $30 billion.

Reliance Industries Limited is procuring advanced AI semiconductors from Nvidia to support this ambitious project. These high-performance chips are essential for complex computations required by AI-driven tools.

The data centre will be powered primarily by renewable energy sources, including solar, wind, and hydrogen power. This initiative aligns with Reliance's strategy to boost India's AI capabilities and make AI more affordable and accessible.

This ambitious project aims to significantly boost India's artificial intelligence capabilities.

The facility is expected to have a total capacity of 3 GW, which is three times larger than the largest existing data centres.

As of now, the largest data centre in the world is the China Telecom-Inner Mongolia Information Park in Hohhot, Inner Mongolia, China. This massive facility spans 10.7 million square feet and consumes 150 megawatts of power. It includes not only data storage and processing infrastructure but also contact centers, warehouses, offices, and housing quarters.

Cost and Investment

Building such a massive facility could require an investment of $20 to $30 billion. Reliance Industries has about $26 billion in cash reserves, but funding such a project would still be a challenge.

Ambani aims to lower the cost of AI inferencing, making AI more affordable and accessible to businesses and startups in India.

"AI inferencing" is the process by which a trained machine learning model applies its learned patterns to new, unseen data to make predictions or decisions. For example, an AI model trained to recognize spam emails can infer whether a new email is spam or not based on the patterns it learned during training.

AI inferencing requires significant computational power, often involving GPUs or specialized hardware like TPUs. The cost of using these resources can be substantial, especially for large-scale models.

For example, companies using NVIDIA's AI inference platform can achieve significant cost savings through optimized performance and energy efficiency. Similarly, startups like Pipeshift offer modular inference engines that can reduce GPU usage by up to 75%, leading to substantial cost savings.

This initiative aligns with Reliance's strategy to disrupt the market, similar to how Reliance Jio revolutionized the telecom sector by offering affordable services. The cost of using these resources can be substantial, especially for large-scale models.

This announcement comes within few days after the announcement of The Stargate Project, a massive artificial intelligence initiative, announced by OpenAI, SoftBank, Oracle, and MGX. The project aims to invest $500 billion over the next four years to build AI infrastructure in the United States. Masayoshi Son, the founder of SoftBank, will chair the project, while Sam Altman from OpenAI will manage operations..