OpenAI is no longer just about AI models. The ChatGPT creator is moving to own the entire AI stack and is targeting countries worldwide to build data centres through a new initiative called OpenAI for Countries.

While the startup explores opportunities in India, the country is pushing ahead with its own plans to build a foundation model under the IndiaAI Mission and develop a sovereign AI stack.

OpenAI is reportedly looking to set up a data centre in India. The company also announced that organisations using its APIs, ChatGPT Enterprise, and ChatGPT Edu will be able to store customer content within India. 

This data residency option extends to other Asian countries, including Japan, Singapore, and South Korea, allowing local data storage while using OpenAI products.

According to a Bloomberg report, OpenAI is targeting the Asia-Pacific region to build new data centres. Company officials are likely to tour the region to explore opportunities around AI infrastructure and software usage. 

Notably, OpenAI opened its first office in Korea recently. 

In a recent congressional testimony, OpenAI CEO Sam Altman said that he wants the entire world to be on the US stack and use US Chips and services like ChatGPT, hoping that it would push for the adoption of US applications. Altman believes the AI stack should be built as a unified system rather than separated across countries or vendors.

Most recently, OpenAI, along with a group of global firms, including G42, Oracle, NVIDIA, SoftBank Group, and Cisco, announced the launch of Stargate UAE, a 1-gigawatt AI compute cluster that will operate within the newly established UAE–US AI Campus in Abu Dhabi. 

 In addition, ChatGPT Plus will soon be available for free to users in the United Arab Emirates (UAE).

Can Stargate Challenge IndiaAI Mission?

IndiaAI mission features over 18,000 GPUs through public-private partnerships with companies such as Jio Platforms, NxtGen Data Centre, Locuz Enterprise, E2E Networks, CtrlS DataCenters, CMS Computers, Orient Technologies, Tata Communications, Vensysco, and Yotta Data Services. 

According to IT minister Ashwini Vaishnaw, eligible users will get access to this computing power at up to 40% lower cost. The total budget for the initiative is ₹10,372 crore. A recent report said the government plans to add around 15,000 more GPUs to the common compute cluster, bringing the total to 29,000. 

Seven shortlisted companies cleared the technical round of the IndiaAI Mission’s second GPU tender, and their commercial bids were recently opened. The companies are Netmagic (NTT Global), Cyfuture India, Sify Digital Services, Vensysco Technologies, Locuz Enterprise Solutions, Yotta Data Services, and Ishan Infotech.

Meanwhile, Yotta has partnered with Microsoft, integrating Microsoft Azure AI services into Yotta’s Shakti Cloud platform to support the IndiaAI Mission. Google has also shown interest in working with the Indian government on its IndiaAI Mission and exploring ways to collaborate with the electronics and IT ministry.

AIM reached out to several local cloud providers in India to understand the difference they found between the IndiaAI Mission and OpenAI’s Stargate, and which one they would choose.

Yotta CEO Sunil Gupta told AIM that the cost to build a hyperscale data centre in India is significantly lower, around $5 million, compared to $10 million in many developed markets. “Affordable land, skilled manpower, and policy incentives from state governments make India one of the most cost-effective destinations globally for building and operating data centres,” he said. 

He shared that Yotta has deployed over 8,000 NVIDIA H100 GPUs, with B200 and GB200 Blackwell chips on the way, and under MeitY’s AI Mission, Yotta provides high-end GPUs like the NVIDIA H100 at an effective rate of sub ₹100/hour, with the government subsidising the cost from a base of ₹130/hour.

Gupta explained that the IndiaAI Mission takes a fundamentally sovereign-first, public-private collaborative approach, which distinguishes it from global initiatives like OpenAI for Countries.  Moreover, Gupta said that if OpenAI ever plans to build Stargate in India, the company is open to working with them.

On the other hand, NxtGen Cloud CEO A S Rajgopal told AIM that it would not partner with OpenAI, emphasising its focus on leveraging open-source AI models as the foundation and fine-tuning them to meet specific customer use cases. “We are focused on building and serving the enterprise inference market,” the company said. 

He explained that it goes against its core principle of offering sovereign cloud and AI services for India. Rajgopal believes the future of AI lies in open source, and the company is not interested in aligning with closed, proprietary agendas. “We may not really add any value to OpenAI. We don’t have the money or talent, and we do not want to end up as an insignificant partner,” the company added.

“The IndiaAI Mission represents a fundamentally different approach compared to initiatives like OpenAI for Countries. While OpenAI’s model aims to provide nations access to its proprietary models and infrastructure, which is still largely hosted and governed under US jurisdictions, the IndiaAI Mission is focused on enabling sovereign, self-reliant AI infrastructure within India’s borders,” said Kesava Reddy, chief revenue officer, E2E Cloud. 

“Unlike OpenAI’s model, IndiaAI is championing open-weight models, publicly funded datasets, and deployment on sovereign cloud infrastructure that aligns with Indian laws, languages, and societal needs.”

However, he added that E2E is open to partnering with OpenAI if it ever builds Stargate in India. 

Ambani to the Rescue 

If there’s one player in India that is capable of helping OpenAI build data centres, it’s Reliance Jio. Operating Stargate-like facilities requires a massive investment, and Reliance has both the scale and resources.

To give an idea, OpenAI’s first data centre in Abilene, Texas, has around 2,200 people on site, according to Chase Lochmiller, founder of Crusoe. In a recent interview with Bloomberg, he said the 1,200-acre campus will eventually house eight buildings that can hold up to 400,000 chips, making it one of the world’s largest known compute clusters. “The whole thing is supposed to be done by mid-2026,” he said.

Oracle will spend around $40 billion on NVIDIA’s higher-performance chips to power OpenAI’s new US data centre. 

Meanwhile, Reliance is developing a large data centre in Jamnagar and has reportedly held talks with OpenAI and Meta about offering data hosting services. With a planned capacity of 3 gigawatts, the facility could support the local deployment of AI models, addressing key concerns around data sovereignty.

By partnering with Reliance, OpenAI plans to offer ChatGPT services to Indian customers at a lower cost, similar to what it did in the UAE. However, if the deal goes through, it could undermine the IndiaAI Mission’s efforts to develop a world-class, homegrown model.

For instance, SarvamAI recently faced backlash from the community for building on top of the Mistral model. If OpenAI’s models become widely accessible to Indian consumers, local alternatives may struggle to gain traction.

In a previous interaction with AIM, IndiaAI Mission CEO Abhishek Singh said that while models from Meta and OpenAI are being embraced, India is also strategically investing in building its own foundational models trained on Indian datasets. This addresses the limitations of existing models, which may lack a nuanced understanding of the Indian context.

During a recent panel discussion, IT minister Vaishnaw stated that one of the primary reasons India needs to develop its LLM is that AI models may not remain open source indefinitely. 

Vaishnaw also took a jibe at OpenAI, pointing out how it transitioned its models to closed source after GPT-2, suggesting that open-source model makers today might take a similar approach.  “We should also change [OpenAI’s] name,” he joked.

The clash between IndiaAI Mission and OpenAI’s Stargate reflects a deeper contest over control, not just of technology, but of national direction in the AI era. Accepting OpenAI’s infrastructure might solve short-term gaps, but it could permanently fracture India’s control over its AI destiny.

The post OpenAI’s Stargate Is a Threat to IndiaAI Mission appeared first on Analytics India Magazine.