Stay Ahead of the Curve

Latest AI news, expert analysis, bold opinions, and key trends — delivered to your inbox.

G42 and Cerebras to Deploy 8 Exaflops of Sovereign AI Compute in India

4 min read G42 has partnered with U.S. chipmaker Cerebras to deploy 8 exaflops of AI computing power in India through a locally hosted supercomputer system. February 20, 2026 14:52 G42 and Cerebras to Deploy 8 Exaflops of Sovereign AI Compute in India

India’s AI infrastructure race just got a sovereign upgrade.

G42 has partnered with U.S. chipmaker Cerebras to deploy 8 exaflops of compute power in India through a new supercomputer system. The announcement was made on the sidelines of the India AI Impact Summit in New Delhi.

The system will be hosted locally and comply with India’s data residency and security requirements — a key point as nations increasingly push for sovereign AI infrastructure.

The compute resources will support:

  • Educational institutions

  • Government agencies

  • Small and medium-sized enterprises

In short: this isn’t just cloud access. It’s national-scale AI infrastructure.


Sovereign AI Takes Center Stage

Manu Jain, CEO of G42 India, described the initiative as critical for “national competitiveness,” emphasizing that local researchers and enterprises will be able to build AI systems while maintaining full data sovereignty.

The project also includes:

  • Mohamed bin Zayed University of Artificial Intelligence (MBZUAI)

  • Centre for Development of Advanced Computing (C-DAC)

Notably, MBZUAI and G42 previously released Nanda 87B, a Hindi-English large language model built on Meta’s Llama 3.1 70B — designed to better understand casual speech in both languages.

Cerebras’ chief strategy officer Andy Hock said the deployment would significantly boost India’s ability to train and run large-scale AI models tailored to local needs.


The Bigger Infrastructure Wave

This announcement comes amid massive AI infrastructure pledges in India:

  • Adani Group committed $100 billion to build up to 5 GW of data center capacity by 2035.

  • Reliance Industries pledged $110 billion over seven years for gigawatt-scale data centers.

The signal is clear: India doesn’t want to just consume AI — it wants to host, train, and govern it domestically.


Why It Matters

8 exaflops is not a small experiment. It’s infrastructure at national scale.

As AI becomes central to economic and geopolitical power, countries are racing to secure compute sovereignty. For G42 and Cerebras, India represents both a massive talent pool and a fast-growing AI market.

For India, it’s about control.

In the next decade, AI power won’t just be measured in model size — it will be measured in who owns the compute.

And India is making sure it has its own.

User Comments (0)

Add Comment
We'll never share your email with anyone else.

img