Inside India’s foundational LLM mission: How Yotta is bridging the compute gap

Yotta has also invested in STPI (Software Technology Parks of India) and NIC (National Informatics Centre) to build sovereign cloud infrastructure for government applications.

As India advances its efforts to establish sovereign AI capabilities, homegrown data center and cloud services provider Yotta is emerging as a key player in the country's quest to develop an indigenous Large Language Model (LLM).

What makes Yotta a significant player in the overall scheme of things is that it will provide 50 percent of the 18,693 advanced GPUs earmarked under the mission. Currently, Yotta has pitched close to 8,200 Nvidia H-100 GPUs and over 1,000 H-200 GPUs for the IndiaAI mission.

In a recent conversation, Yotta highlighted two important aspects of the process that are in the works right now. First: vetting, approval, and the subsequent permission to access the compute resources on the government’s IndiaAI Mission portal. Secondly, integrating the India AI portal through an Application Programming Interface (API) linkage to the Cloud for provisioning of Graphics Processing Units (GPU).

Explaining the details of the two aspects, Sunil Gupta, co-founder, Chief Executive Officer, and Managing Director, Yotta Data Services, told Moneycontrol that entities can select their required compute resources, submit supporting documentation, and receive instant access once their request is cleared by the government through a single unified interface.

“The moment government does that and approves, it will also select the name of the service provider who will give it, depending on the logic which government has defined,” Gupta told Moneycontrol on the sidelines of Nasscom Technology & Leadership Summit, 2025.

The second key part of the mission is Yotta’s Shakti Cloud, which is currently being integrated into the India AI portal through an automated API linkage.

“Our software team is working with IndiaAI Mission’s portal team to create an API linkage between IndiaAI portal and Shakti Cloud… the end-user account will be created in Shakti Cloud, and the service which end-user is asking for will get delivered on my (Shakti) cloud,” Gupta explained.

Yotta is also empanelled under the government’s India AI Mission, which will ease the burden of expensive compute by providing a 40 percent subsidy on GPU Cloud services to eligible entities.

Through the Shakti Cloud and the India AI portal, Indian AI startups, research institutions, and enterprises can access subsidised GPU compute power.

To be sure, the Ministry of Electronics and Information Technology (MeitY) has launched a continuous empanelment process for AI cloud service providers, allowing firms to apply on an ongoing basis to supply GPUs and related AI compute services under the IndiaAI Mission.

The empanelment process will ensure competitive pricing; as new firms can challenge existing lowest-price (L1) bidders by matching or lowering the rates. Apart from Yotta, existing L1 bidders include Jio Platforms, NxtGen Data Center, Locuz Enterprises, E2E Networks, CtrlS DataCenters, CMS Computers, Orient Technologies, Tata Communications, Vensysco, and Data Services.

The last date for the first round of applications under this continuous empanelment process is April 30, 2025, with evaluations set to begin in May 2025.

Also, read: Govt focuses on foundational AI, allocates Rs 1,500 cr, rules out tweaks to existing models

Gupta has been working for decades on data center businesses and co-founded Yotta in 2019 with the backing of real estate billionaire Niranjan Hiranandani.

A Sovereign AI Future

While there’s a debate on whether India should build its foundational model or stick to building applications on top of it and let the big boys of Silicon Valley build LLMs, Gupta believes the push for India’s sovereign AI infrastructure is not just about technology—it’s also about national security and business.

“We may have lost the data centre race, we are trying to come up. We may have lost the cloud race, we are trying to come up. So if at this stage itself, we can build a huge AI compute infrastructure in India, we can then create services on top of that,” said Gupta, who is also known as the “Data Centre Man of India”.

After developing use cases for Indians, the country can also aspire to export to the rest of the world. “Then exporter is not just again of our bodies and our skill sets, the way we were doing in software, we will be exporting India-created AI products.”

Moreover, Yotta has invested in STPI (Software Technology Parks of India) and NIC (National Informatics Centre) to build sovereign cloud infrastructure for government applications. By investing in GPU-powered Data Centres, Gupta believes he is ensuring that Indian ministries and departments can host critical AI workloads on India-owned infrastructure, reducing foreign cloud dependency.

“We may not be the as sophisticated cloud as Amazon. Obviously, they have done ages of R&D before they launched their cloud. But even if I am launching 50 services compared to their 300 services, at least I am running a great service and I am running it fully in the controls of Indians,” Gupta added.

Why Yotta’s Cloud Matters

For years, India has been a leader in software exports, but its data centre infrastructure has lagged.

While the US has nearly 17,000 megawatts of Data Centre capacity, India, despite its four times larger population, has just 1,400 megawatts. Moreover, 20 percent of the world’s data is either produced or consumed by Indians, yet only 3 percent of global data is hosted in India.

This lack of local cloud infrastructure has forced Indian companies to rely heavily on foreign cloud providers, creating concerns around data sovereignty, security, and long-term digital independence.

Yotta is building a sovereign cloud ecosystem with a massive Data Centre campus in Mumbai and a similar setup in Greater Noida.

“We build Data Centres but we never excelled in the race for building our own cloud infrastructure. We all ended up reselling AWS, Azure, and Google Cloud services. All the tech now is with them and we are just leasing out spaces to them,” he said. “God forbid, a situation comes of you being in a corner and getting isolated, if you are not sovereign in terms of your technology infrastructure, people can control you.”

Nevertheless, it would be important to see how the bid to pursue India's AI infrastructure pans out with multiple domestic players eyeing the same pie even as the competition heats up.