The Growth Trajectory: From 1 GW to 3 GW by 2030

In the first quarter of 2024, India’s total data center capacity crossed the 1 GW mark, a milestone that signifies the country's growing role in the global digital ecosystem. According to industry estimates, this capacity is expected to reach 3 GW in the next five years. A significant driver of this growth is the increasing demand for AI and GPU processing, which requires massive computational power.

AI workloads involve processing vast amounts of data, making them more resource-intensive than traditional IT tasks. This shift is pushing data center operators to invest in GPU-based infrastructure, which offers the high-performance computing needed to support AI applications like machine learning, natural language processing, and advanced data analytics.

Sunil Gupta, Managing Director and CEO of Yotta Data Services, notes that AI and GPU-based workloads will occupy 40-50% of India’s data center capacity by 2030. “We expect the Big Three – Microsoft, Amazon Web Services (AWS), and Google – to own 30% of this total capacity,” says Gupta. These global cloud players are not only expanding their footprints in India but also building their own captive data centers to cater to the rising demand for AI and cloud services.

The Big Three: Building the Future of Captive Data Centers

Microsoft, AWS, and Google are at the forefront of India’s data center growth, with aggressive plans to establish their own data centers across the country. These tech giants are making substantial investments to secure land and build facilities that will power the next generation of digital services.

For example, Microsoft has acquired two parcels of land in Pune's Hinjawadi area and another in the nearby Pimpri-Chinchwad area. The company is also expanding in Hyderabad, another key location for data center development. Similarly, AWS and Google are building new capacity in Mumbai, a critical hub for India’s digital infrastructure. Google is reportedly negotiating the acquisition of 22.5 acres of land in Navi Mumbai, which would house its first captive data center in India, while AWS is exploring land parcels near Mumbai for further expansion.

These developments are driven by the need to meet growing demand for cloud services and AI workloads, which are set to dominate the future of digital transformation. Captive data centers will allow these tech giants to offer more reliable, scalable, and cost-effective solutions to their global and Indian customers.

Why India? The Competitive Edge in AI Data Centers

India's emergence as a potential offshore AI data center hub is no accident. The country offers several competitive advantages that make it an attractive destination for building AI-led data centers, especially when compared to other locations in the Asia-Pacific (APAC) region.

• Cost-Efficient Infrastructure: Building a data center in India is significantly cheaper than in neighboring countries. Cushman & Wakefield estimates that the cost to build a 1 MW data center in India ranges from $6-8 million, compared to $10 million in Indonesia and Australia, and $8 million in Thailand.
• Lower Power Costs: India's access to affordable power, particularly renewable energy, further enhances its appeal. As data centers consume vast amounts of electricity, especially for GPU-based AI workloads, access to low-cost, sustainable power sources is critical for long-term viability.
• Cheaper Labor: India's labor costs are lower than many other APAC countries, making it more affordable to build and operate large-scale data centers. This is especially important for companies like Microsoft, AWS, and Google, which are planning massive investments in infrastructure.

The Cost of AI: High Investment, High Reward

While India’s cost advantages are clear, AI-driven data centers come with their own set of challenges, primarily related to the capital expenditure required to build and maintain GPU-based infrastructure. Data centers that process typical IT workloads cost around $5 million per MW (excluding land costs). However, data centers built for AI and GPU workloads can require significantly higher investment. For instance, a 24 MW data center dedicated exclusively to GPU workloads could cost approximately $120 million, with additional expenses for high-end GPUs.

A facility with 16,000 high-end GPUs would increase the total cost to $720 million, reflecting the expensive nature of GPU infrastructure. Despite these high costs, the long-term potential for AI-driven data centers in India remains immense, particularly as demand for AI services continues to soar across industries such as healthcare, finance, and autonomous systems.

Yotta Data Services: Leading the AI Data Center Race

Yotta Data Services, a pioneer in India’s data center market, is setting the pace with its focus on AI and GPU processing. As the first data center provider in the country to offer Nvidia GPUs to clients, Yotta is already making strides toward establishing India as a major player in AI infrastructure. The company has placed an order for 16,000 Nvidia GPUs, 4,000 of which have already been delivered. The remaining GPUs are expected to arrive soon, with plans to acquire an additional 32,000 in the near future.

This strategic move positions Yotta as a leader in the AI data center space, with the capacity to meet the growing demand for AI workloads. Moreover, the company’s investments align with broader efforts by the Indian government to promote AI development. The Ministry of Electronics and Information Technology recently signed an agreement with Nvidia to acquire 10,000 GPUs, which will be allocated to startups, further fueling AI innovation in the country.

The Future: India as an Offshore AI Hub

India's data center capacity is expanding at an unprecedented pace, driven by a combination of favorable economic factors and the growing demand for AI-driven solutions. While the journey to becoming a global AI data center hub will take time, the country’s competitive advantages—lower construction costs, affordable power, and cheaper labor—make it well-positioned to capture a significant share of the global market.

As the Big Three cloud giants—Microsoft, AWS, and Google—continue to build their own data centers in India, the country’s digital infrastructure will only strengthen, enabling it to support a wide range of AI applications and cloud services. By 2030, AI and GPU-based workloads could account for up to half of India’s total data center capacity, marking a major shift in the way data is processed and managed in the country.

The data center revolution is underway, and India is at its forefront. With the right investments and strategic initiatives, the country has the potential to become a global leader in AI-driven digital transformation, paving the way for a more connected and intelligent future.

Key Highlights:

• 40-50% of India's data center capacity will be dedicated to AI/GPU workloads by 2030.
• Total capacity is expected to treble to 3 GW by 2030.
• Microsoft, AWS, and Google are building captive data centers, aiming to control 30% of the market.
• India's cost-efficient infrastructure makes it a competitive hub for AI data centers.
• Yotta Data Services is leading the way with its investment in Nvidia GPUs for AI processing.