RackBank plans to build an AI data center in Raipur, Chhattisgarh, with an initial investment of 1,000 crore rupee (US$118.3 million).

The facility will cover 13.5 acres and house up to 100,000 GPUs. An additional 2.70 hectares have been set aside in the Chhattisgarh SEZ for future expansion.

The total investment may reach 3,000 crore rupee (US$354.89 million) over the next five years.

The data center will be built in four phases, starting with an 80 MW capacity that can scale up to 160 MW by the fourth phase.

The project is expected to create 200 jobs initially, with plans to increase that number to 500.

🔗 Source: The Economic Times

🧠 Food for thought

1️⃣ India’s rapid AI data center buildout reflects broader strategic shift in the market

RackBank’s entry into Raipur represents a departure from India’s traditional data center hubs like Mumbai, Hyderabad, and Chennai, highlighting a new phase in the country’s digital infrastructure expansion.

The data center market in India is projected to grow from $6.48 billion in 2024 to $10.70 billion by 2030 at a CAGR of 8.72%, with capacity expected to increase by 66% (adding 604 MW) by 2026 alone 12.

This rapid growth is being fueled by significant investments across the country, including NTT DATA’s ₹10,500 crore AI cluster in Hyderabad and Reliance Group’s planned 3 GW facility in Gujarat, which is set to become one of the world’s largest data centers 34.

RackBank’s focus on housing 100,000 GPUs reflects the specialized hardware requirements driving the latest generation of data centers, as AI applications demand substantially different infrastructure than traditional computing workloads.

The shift toward tier-2 cities like Raipur demonstrates how data center operators are balancing factors like land availability, power costs, and latency requirements to optimize their investments beyond saturated primary markets.

2️⃣ Cooling technology emerges as critical competitive differentiator in AI-focused facilities

RackBank’s emphasis on proprietary liquid cooling technology highlights an industry-wide shift toward specialized thermal management for AI workloads, which require significantly higher power densities.

Power consumption for AI data centers is projected to increase by 50% by 2027, creating immense thermal challenges that cannot be addressed with traditional air cooling methods 5.