How Yotta Is Building Its AI Infra Playbook To Serve Global Enterprises

For Yotta Data Services, it’s a race against time to build one of India’s largest pools of AI computing power, and the neocloud infrastructure giant is chasing global demand to fill it, rather than just tap the domestic research and training workloads. 

Over the next four months, the company expects to scale its AI cloud to more than 40,000 NVIDIA Blackwell GPUs and to around 85,000 GPUs by the end of this financial year. 

Yotta cofounder, managing director and CEO Sunil Gupta told Inc42 nearly all of this new GPU deployment is being driven by international customers and enterprises that are willing to sign billion-dollar contracts if delivery can be guaranteed within four to five months. 

That marks a sharp shift from Yotta’s earlier inward focus. Until February, nearly 70% of Yotta’s GPUs were dedicated to India’s AI ambitions, supporting startups and institutions such as Sarvam AI, IIT Bombay and IIT Madras. 

Once those training workloads concluded, Yotta began reallocating capacity overseas, where demand is stronger and pricing is roughly 50% higher — flipping its customer mix to 75% international and 25% domestic, with Gupta expecting 90:10 by year-end. 

Funding the build-out is a capital raise already well under way. Yotta has raised $150 Mn from non-institutional investors at a valuation of ₹37,000 Cr (close to $3.9 Bn), with cofounder and CEO Sunil Gupta telling Inc42 that most of the round closed by March and some transactions ran through May. 

There is no promoter offer-for-sale, with Yotta preparing to file its DRHP in the coming months. The company is targeting about $1.5 Bn in all across the pre-IPO rounds and the IPO itself. The CEO is not bothered by the chatter around the valuation inflation before the IPO. 

“As Yotta continues signing larger customer contracts, valuation continues to increase,” Gupta told us. “Our valuation today is already around 30% higher than it was in March.”

The company competes with players such as Neysa, E2E Networks, NxtGen, CtrlS and ESDS, although several of these companies derive a substantial portion of their revenue from traditional cloud or data centre services rather than AI infrastructure alone. 

Among these, ESDS has also filed its DRHP as it prepares to go public this year, with an ₹700 Cr IPO. One could say these companies represent a different side of the AI boom, and perhaps one that will become more visible to public markets investors in the near future.  

GPU Demand Comes On, Fundraising Strategies Shift

Founded in 2019 by entrepreneur Sunil Gupta and Darshan Hiranandani as part of real estate giant Hiranandani Group, Yotta Data Services began as a hyperscale data centre and cloud infrastructure company.

Yotta’s big moment came around 2023-24 when it procured thousands of NVIDIA’s GPUs, the specialised chips foundational to current high-performing AI systems. This gave it a leg up over other domestic cloud infrastructure and compute providers. 

Enterprises are often mandated to use domestic cloud providers when working on government contracts and tenders, so Yotta earned an early lead when it comes to GPU compute. 

With the launch of the IndiaAI Mission by the Indian government, Yotta also became one of the earliest contributors to its common compute facility. Today, it is one of the largest compute providers to the MeitY-led programme, which has a stated figure of 38,000 GPUs overall.

The company expects its cumulative investment in AI infrastructure to exceed $7 Bn by the end of FY27, driven by surging global demand for GPU capacity, of which about $4 Bn has already been committed to buy AI infrastructure.

To finance its planned $7 Bn AI infrastructure expansion spree, Yotta is moving beyond traditional debt and equity by adopting a GPU financing model pioneered in mature markets. 

Under the structure, partners such as Gorilla, a US-listed technology company, fund and own the GPUs, while Yotta installs and operates them in its data centres and signs customer contracts.In April, the companies deepened their partnership, adding another $2.8 Bn to an agreement that was initially valued at $500 Mn.

Revenue is shared, with roughly 75% going to the funding partner and 25% to Yotta, reflecting the higher capital invested in GPUs. After four to five years, Yotta has the option to buy back the GPUs at market value and continue monetising them through its AI inference, or ‘token factory’, business.

Yotta’s Three-Tiered Business 

Yotta began as a colocation and integrated data centre services provider before expanding into sovereign cloud offerings for enterprises and government agencies. Today, Yotta has a three tiered business model which includes colocation, sovereign cloud and managed services, and neocloud offerings.

Yotta derives around 20% of its business from colocation, 25-30% from sovereign cloud and managed services, and about 50% from GPUs, Gupta said. 

While all three businesses are growing, AI infrastructure increasingly dominates the company’s revenue mix. “We have secured one of the largest colocation contracts in the market, so that business will continue to grow. Our sovereign cloud business is also seeing strong momentum because of our government partnerships through NIC,” he said. 

However, the GPU business is expanding much faster, backed by global customer contracts already in hand. Gupta said that over the next two years, colocation’s share of the business is expected to decline to around 15%, with the remaining 85% coming from sovereign cloud and AI infrastructure, where “the GPU business will be the largest contributor.”

That said, Yotta’s sovereign cloud is deriving much interest from several other countries, said Gupta. 

“Outside India, governments across South Asia, the Commonwealth of Independent States (CIS), the broader Global South and some parts of the West are approaching us with a common request of whether the sovereign offering can be replicated in their countries,” he said. 

He added that Yotta is in discussions with several countries and expects to announce some international partnerships in the coming months. Sovereignty in AI requires domestic AI infrastructure and compute providers like Yotta, but global partnerships are equally critical to keep pace with the international AI ecosystem, and attract foreign investments alongside domestic funds. 

A Race For The Public Markets

Building on the momentum of 2025, India’s IPO pipeline remains robust in the first half of this year. As many as 28 startups have filed their DRHPs with SEBI, while more than 24 others are in various stages of preparing for a public listing.

Yet, despite the AI boom, AI and AI infrastructure companies remain largely absent from the IPO wave. So far, companies such as E2E Networks and Fractal Analytics are among the few AI-focused firms to have entered the public markets, highlighting the gap between the sector’s private market momentum and its public market reality and representation.

Even among VCs in the private market, there is some caution at play, because early AI bets are failing before they even have enough time to build. 

As per Inc42’s Indian Investor Ranking & Sentiment Survey Q1 2026, about 92% of the 70 institutional investors surveyed are taking a rather selective approach to their AI bets. 

A turning of this tide could begin from the public market, once the neocloud and AI infra companies become retail investor targets. Indeed, the holes in the India AI narrative stand in contrast to global markets, particularly the US, where the AI infrastructure boom has already produced a new class of publicly traded companies, and even the first ever trillionaire. 

In the US, giants such as CoreWeave have listed, while companies including Nebius Group and Applied Digital have repositioned themselves around AI compute, attracting significant investor interest. Anyone who is even remotely in the compute infrastructure space has seen their stock fly. 

Investors globally have become increasingly willing to back capital-intensive AI infrastructure businesses, viewing access to compute as a strategic asset rather than just another cloud service. Will this reality also dawn in India with companies like Yotta, Neysa, ESDS, E2E, NxtGen and others emerging? 

The emergence of credible listed players in the AI infrastructure, compute and data centre space might not just be a big checkbox that global investors can tick, but it’s a very real boost to the domestic and sovereign AI ambitions, since infrastructure is the heart of the AI stack.

[Edited by Nikhil Subramaniam]

The post How Yotta Is Building Its AI Infra Playbook To Serve Global Enterprises appeared first on Inc42 Media.