OpenAI Eyes India for Massive AI Data Center Under Stargate Expansion

New Delhi, Sept 2025 — OpenAI is preparing to build a gigawatt-scale data center in India, marking one of its biggest international infrastructure bets to date. The project is part of the company’s ongoing “Stargate” expansion, a global program designed to provide the enormous computing power needed for training and running frontier artificial intelligence (AI) models.

Reports suggest the facility could reach a capacity of 1 gigawatt (1GW) — placing it among the largest planned data centers in Asia. OpenAI is currently in talks with local partners to secure land, power, and regulatory approvals.

👉 For deeper insights into OpenAI’s latest advancements, check out our coverage of OpenAI GPT Realtime and its breakthrough in live AI interactions, as well as OpenAI Learning Accelerator India, which highlights how OpenAI is strengthening its research and education footprint in the country.

Why India?

India has emerged as one of the world’s fastest-growing AI markets, with millions of developers and businesses experimenting with generative AI. Locating a major data center here would bring several advantages:

• Lower Latency: Faster access to OpenAI services for Indian users and enterprises.
• Compliance: Support for India’s evolving data protection and localization requirements.
• Market Growth: Access to a large pool of enterprise clients in IT, banking, telecom, and retail.

Experts say the move also signals OpenAI’s recognition of India’s role as both a consumer market and a talent hub for AI.

What Is “Stargate”?

“Stargate” is the name given to OpenAI’s multi-gigawatt global infrastructure program. First reported earlier this year, the initiative involves building hyperscale data centers with the backing of partners such as Microsoft and Oracle.

In July, OpenAI and Oracle confirmed a deal to advance 4.5GW of Stargate capacity in the U.S.. India now looks set to be the next frontier.

Scale of the India Project

A 1GW facility is extraordinary in scale. To put it in perspective:

• Power Demand: Equivalent to the electricity used by a medium-sized city.
• Compute Capacity: Enough to host hundreds of thousands of GPUs and AI accelerators.
• Investment: Likely to run into billions of dollars spread over multiple years.

The project will not be a single building. Industry analysts expect a campus model with phased build-outs of 100–250MW each.

Possible Locations

While OpenAI has not confirmed the exact site, industry watchers point to four leading candidates:

• Maharashtra (Mumbai/Navi Mumbai) — close to financial hubs and undersea cables.
• Telangana (Hyderabad) — fast-emerging as India’s data center capital.
• Karnataka (Bengaluru) — India’s technology hub with deep engineering talent.
• Delhi NCR — access to government and enterprise clients.

Final decisions will hinge on power availability, renewable energy sourcing, land acquisition, and fiber connectivity.

Partners in Focus

In the U.S., OpenAI has leaned heavily on Microsoft Azure and Oracle Cloud for infrastructure. Industry insiders believe a similar partnership model will be replicated in India:

• Microsoft Azure: Already operates large cloud regions in India.
• Oracle: Actively expanding its AI data center footprint worldwide.
• Local Utilities and Colocation Providers: Needed to manage land, power, and operations.

No official partners have been announced yet, but talks are reportedly underway.

Power and Sustainability Challenge

The single biggest question for a 1GW facility is energy. Data centers of this scale require constant, reliable power — and OpenAI will be under pressure to secure it sustainably.

• Renewable PPAs: Long-term solar and wind power purchase agreements will likely be part of the deal.
• Energy Storage: Batteries and backup systems to stabilize supply.
• Cooling Innovation: Use of liquid and immersion cooling to reduce water consumption.

Analysts warn that without a strong green energy strategy, projects of this magnitude could face public and regulatory pushback.

Chip Supply and Global Race

AI infrastructure depends on specialized chips such as NVIDIA’s GPUs. With global demand surging, allocation remains tight.

OpenAI’s India project will need to align with global supply chains for accelerators, memory, and networking gear. This could lead to local assembly and vendor opportunities for Indian firms involved in data center construction, cooling, and hardware integration.

Regulatory Landscape

India is rolling out new rules on data protection and AI oversight. Locating infrastructure in India would allow OpenAI to:

• Store sensitive data in compliance with local laws.
• Provide enterprise customers with assured security and privacy.
• Participate in policy discussions around AI ethics and safety.

Industry observers say that OpenAI’s move could push other AI players to localize compute in India.

Economic Impact

The planned data center could create:

• Direct Jobs: Construction workers, engineers, and data center staff.
• Indirect Jobs: Supply chain vendors, logistics, and security.
• Skilling Opportunities: Training in AI operations, data engineering, and machine learning ops.

Analysts predict tens of thousands of jobs could be supported over the life of the project.

What to Watch

While OpenAI has not shared a public timeline, three milestones will be key:

1. Official Announcement & Office Opening in India.
2. Power Purchase Agreements (PPAs) with renewable suppliers.
3. Land Acquisition & Construction Filings at the state level.

Experts say the project will be built in phases over several years, with the first stage likely focused on inference capacity before full-scale training clusters come online.

Reactions

Industry experts welcomed the move:

• “A 1GW AI data center in India would mark the country’s arrival as a global AI hub,” said a senior analyst at a leading research firm.
• “This will not only bring jobs but also reshape India’s cloud and semiconductor ecosystem,” added a Bengaluru-based AI startup founder.

However, some raised concerns about energy use and sustainability. “Powering a gigawatt-scale campus responsibly will be India’s biggest challenge,” noted a data center consultant.

FAQs

Q. Has OpenAI confirmed the project?
Not officially. Multiple reports confirm OpenAI is in discussions, but details remain confidential.

Q. What does 1GW mean?
It’s enough power to support hundreds of thousands of AI chips and could equal the energy usage of a mid-sized city.

Q. Where will it be built?
Locations under consideration include Mumbai, Hyderabad, Bengaluru, and NCR. Final choice will depend on land, power, and connectivity.

Q. Who are the likely partners?
Microsoft, Oracle, and local utilities are strong candidates, but no names are official yet.

Q. When will it go live?
Construction timelines are not confirmed. Industry watchers expect phased rollout over several years.

Conclusion

OpenAI’s plan to establish a 1GW data center in India is one of the boldest bets on AI infrastructure in Asia. While details are still emerging, the move underscores India’s growing importance in the global AI race. If completed, the project could transform the country’s role in AI research, enterprise adoption, and digital infrastructure — while testing its capacity to deliver sustainable, gigawatt-scale power for the future of artificial intelligence.