Story perspectives

India Allocates INR 10,000 Crore for AI Compute, Lagging US

7/7/2026

Summary

India Lacks Compute

India’s AI ambitions risk stalling because the country lacks sovereign compute infrastructure.
The United States dominates AI compute, with NVIDIA supplying about 80-90 % of advanced accelerators.
Training frontier models can require tens of thousands of GPUs, with a 100,000-GPU cluster costing billions at $25,000–40,000 per H100 GPU.
IndiaAI Mission has earmarked over INR 10,000 crore for a national compute facility, a modest sum compared with global multi-billion-dollar clusters.

Summary

Supermicro Launches Edge-AI

Supermicro launched edge-AI hardware in fanless DIN-rail, 1U rackmount and mini-tower formats, each up to 180 TOPS.
The devices use Intel Core Ultra CPUs that combine CPU, GPU and NPU on one chip, while Intel Arc Pro B-Series GPUs such as the B70 with 32 GB VRAM enable larger generative models.
Partnerships with StorMagic and Odine lower costs and meet data-sovereignty rules for the DIN-rail unit, which runs 0 °C-45 °C for vision.

History

Historical Context

In 2006 AWS launched EC2, giving the U.S. its first large-scale commercial cloud and capturing about 33% of the global infrastructure market. That model spurred national programs such as India’s AI Mission, which in 2024 earmarked INR 10,000 crore for a sovereign compute centre and partnered with firms like Supermicro and Intel to build edge-AI racks.

1 of 3

India Lacks Compute

India’s AI ambitions risk stalling because the country lacks sovereign compute infrastructure.
The United States dominates AI compute, with NVIDIA supplying about 80-90 % of advanced accelerators.
Training frontier models can require tens of thousands of GPUs, with a 100,000-GPU cluster costing billions at $25,000–40,000 per H100 GPU.
IndiaAI Mission has earmarked over INR 10,000 crore for a national compute facility, a modest sum compared with global multi-billion-dollar clusters.

1 / 3