Indian data centre companies are looking to attract neocloud firms such as Coreweave, Lambda Labs, Crusoe and Nebius, which offer high-end computing for generative artificial intelligence training at nearly one-third the price of hyperscalers like Amazon Web Services, Google Cloud and Microsoft Azure.
Backed by large investors, neoclouds, or ‘babyscalers’—boutique operators offering specialised infrastructure and flexible GPU rentals—are fast snagging business from Microsoft Azure, AWS and Google Cloud. A study by digital infrastructure consultancy Uptime Institute showed that the average hourly cost of an Nvidia DGX H100 GPU system from a hyperscaler was $98, while neocloud players offer it at about $34—a 66% saving.

These attractive prices, along with cheap data centre space and power in India, make a powerful case for neocloud companies to invest here, industry experts said. Sharad Agarwal, CEO of Chennai-based data centre services provider Sify Infinit Spaces, said neocloud companies are watching the Indian market keenly, especially because of the world’s lowest rack rentals and power costs here. Many neocloud operators evolved from Bitcoin mining, gaming and high-frequency trading backgrounds. As AI chips grew in demand post ChatGPT’s release in 2022, these firms quickly amassed thousands of chips to offer on rent.
Notably, Wall Street financiers have created a lucrative debt market for the neocloud firms which use the funds to acquire thousands of Nvidia chips and further rent out on hourly/monthly usage basis.
Most neocloud firms are cash rich. Large private equity firms like Blackstone, Carlyle and BlackRock have pumped in $11 billion in this new cohort of companies. Coreweave made an impressive $23-billion debut on Nasdaq this March.
According to semiconductor research firm Semianalysis, besides the four neocloud leaders, there are 131 emerging players in the segment globally who offer GPU rentals at discounted prices.

India’s offer
Sify Infinit Spaces has introduced a pay-per-use pricing model, offering power and rack space to neocloud companies deploying GPU clusters in India. “Our pay-per-use pricing model is a disruptor to the data centre industry,” Agarwal said. “Huge entry cost is a barrier for emerging cloud infra firms to expand in foreign markets, and hence this pricing structure could derisk companies’ fixed cost.”
Hiranandani Group’s Yotta Data Services – which by itself is India’s sovereign neocloud -- said it has listed 8000 Nvidia H100 GPUs on Nvidia’s marketplace alongside Lambda Labs, Coreweave and Nebius to offer their GPUs-as-a-service. “This shift from hyperscalers to AI-specialised clouds is not just a competitive realignment—it’s critical for democratising access to high-performance compute and accelerating innovation,” said Sunil Gupta, cofounder, CEO and managing director of Yotta Data Services, which has deployed its own sovereign neocloud with IndiaAI Mission, NIC, and STPI.
Essar Group’s Blackbox said it is tapping North America and Europe expansion of neocloud firms to provide cabling, design, and construction services, and participate in their “multi-billion gigawatt-scale expansion.”
Blackbox is also pursuing global rollouts of gigawatt-scale campuses of neocloud operators for a range of services. “These include structured cabling, data centre design, and turnkey construction services for both greenfield and retrofit projects,” said Sanjeev Verma, president and CEO of Black Box.
Although current discussions are focused on North America and Europe markets, expansion in India is not a question of if, but when, Verma added. CtrlS Datacenters is also in early talks with neocloud firms “to deploy GPU clusters in India, driven by increasing AI model training and inference workloads, especially in sectors like BFSI, healthcare and manufacturing,” said Ranjit Metrani, president, managed services, at CtrlS Datacenters.
Shift in cloud business
Experts say it is the right time for India’s infrastructure players to strike long-term partnerships with neocloud firms as the industry is growing rapidly and India is evolving to be one of the largest markets for AI use. “Neocloud companies are gaining market share globally at an incredibly fast pace comparable to the early growth of hyperscalers,” said Jitesh Karlekar, director-research at real estate research and advisory firm JLL.
“This is a fundamental shift in the cloud business model… Neocloud firms are architected for the AI-native era. They offer more specialised infrastructure, lower latency, flexible pricing, and are often better optimised for high-performance computing workloads.” Therefore, it is crucial for India’s infrastructure players to catch on this momentum early,” he said.
Backed by large investors, neoclouds, or ‘babyscalers’—boutique operators offering specialised infrastructure and flexible GPU rentals—are fast snagging business from Microsoft Azure, AWS and Google Cloud. A study by digital infrastructure consultancy Uptime Institute showed that the average hourly cost of an Nvidia DGX H100 GPU system from a hyperscaler was $98, while neocloud players offer it at about $34—a 66% saving.

These attractive prices, along with cheap data centre space and power in India, make a powerful case for neocloud companies to invest here, industry experts said. Sharad Agarwal, CEO of Chennai-based data centre services provider Sify Infinit Spaces, said neocloud companies are watching the Indian market keenly, especially because of the world’s lowest rack rentals and power costs here. Many neocloud operators evolved from Bitcoin mining, gaming and high-frequency trading backgrounds. As AI chips grew in demand post ChatGPT’s release in 2022, these firms quickly amassed thousands of chips to offer on rent.
Notably, Wall Street financiers have created a lucrative debt market for the neocloud firms which use the funds to acquire thousands of Nvidia chips and further rent out on hourly/monthly usage basis.
Most neocloud firms are cash rich. Large private equity firms like Blackstone, Carlyle and BlackRock have pumped in $11 billion in this new cohort of companies. Coreweave made an impressive $23-billion debut on Nasdaq this March.
According to semiconductor research firm Semianalysis, besides the four neocloud leaders, there are 131 emerging players in the segment globally who offer GPU rentals at discounted prices.

India’s offer
Sify Infinit Spaces has introduced a pay-per-use pricing model, offering power and rack space to neocloud companies deploying GPU clusters in India. “Our pay-per-use pricing model is a disruptor to the data centre industry,” Agarwal said. “Huge entry cost is a barrier for emerging cloud infra firms to expand in foreign markets, and hence this pricing structure could derisk companies’ fixed cost.”
Hiranandani Group’s Yotta Data Services – which by itself is India’s sovereign neocloud -- said it has listed 8000 Nvidia H100 GPUs on Nvidia’s marketplace alongside Lambda Labs, Coreweave and Nebius to offer their GPUs-as-a-service. “This shift from hyperscalers to AI-specialised clouds is not just a competitive realignment—it’s critical for democratising access to high-performance compute and accelerating innovation,” said Sunil Gupta, cofounder, CEO and managing director of Yotta Data Services, which has deployed its own sovereign neocloud with IndiaAI Mission, NIC, and STPI.
Essar Group’s Blackbox said it is tapping North America and Europe expansion of neocloud firms to provide cabling, design, and construction services, and participate in their “multi-billion gigawatt-scale expansion.”
Blackbox is also pursuing global rollouts of gigawatt-scale campuses of neocloud operators for a range of services. “These include structured cabling, data centre design, and turnkey construction services for both greenfield and retrofit projects,” said Sanjeev Verma, president and CEO of Black Box.
Although current discussions are focused on North America and Europe markets, expansion in India is not a question of if, but when, Verma added. CtrlS Datacenters is also in early talks with neocloud firms “to deploy GPU clusters in India, driven by increasing AI model training and inference workloads, especially in sectors like BFSI, healthcare and manufacturing,” said Ranjit Metrani, president, managed services, at CtrlS Datacenters.
Shift in cloud business
Experts say it is the right time for India’s infrastructure players to strike long-term partnerships with neocloud firms as the industry is growing rapidly and India is evolving to be one of the largest markets for AI use. “Neocloud companies are gaining market share globally at an incredibly fast pace comparable to the early growth of hyperscalers,” said Jitesh Karlekar, director-research at real estate research and advisory firm JLL.
“This is a fundamental shift in the cloud business model… Neocloud firms are architected for the AI-native era. They offer more specialised infrastructure, lower latency, flexible pricing, and are often better optimised for high-performance computing workloads.” Therefore, it is crucial for India’s infrastructure players to catch on this momentum early,” he said.