The Indian data centre industry is experiencing rapid growth, with its current capacity expected to double in the next three years. This surge is driven by the increasing digitisation of industries, the rise of AI applications, and the expanding demand for cloud services. However, as AI data centres expand, so do their power needs. Meeting these requirements with traditional energy sources is becoming increasingly challenging, necessitating a shift to more efficient and sustainable power solutions.
This is where small modular reactors (SMRs) emerge as a potential game-changer. SMRs, with their compact size and ability to produce consistent, high-output power, could provide a reliable energy source for AI data centres, solving the dual challenge of increasing energy demand and environmental sustainability.
The growing energy demand of AI data centres
AI data centres are specialised facilities designed to support the intensive computational requirements of AI applications. These centres house high-performance GPUs and advanced cooling systems to handle the substantial heat generated by AI workloads. With the rapid growth of AI, machine learning, and big data, AI data centres provide the robust infrastructure needed for processing massive datasets and training complex models. Their ability to deliver dense, consistent power and cooling solutions makes them essential for advancing technology and innovation in various industries. This surge in AI applications is driving the increasing energy consumption of these centres, making it more challenging to meet power demands using traditional energy sources. As per a report by the International Energy Agency, data centres worldwide already consume approximately 1-1.5% of global electricity, a figure set to rise as AI continues to advance.
According to ICRA, the country’s data centre capacity is anticipated to increase from 950 MW in FY24 to over 2,000 MW by FY27, essentially equating the power needs of data centres to those of an entire city. This growing demand for continuous power is driven by emerging technologies such as AI, machine learning, cryptocurrency, and blockchain, which utilise high-performance GPUs. These advanced processors, despite their compact size, generate substantial heat and require sophisticated cooling systems, thereby further increasing power consumption.
Traditional energy grids are struggling to keep pace with these demands, often leading to higher operational costs and sustainability challenges. Additionally, renewable energy sources like solar or wind, while beneficial for the environment, often face reliability issues due to their dependency on weather conditions. AI data centres, with their 24/7 operational needs, require a more stable, high-density power source.
“The reality is that, as AI technology evolves, power consumption remains a significant challenge that needs to be solved. Traditional power sources like fossil fuels, wind, and solar have their limitations, and standardisation is still lacking, making it difficult to scale data centres with these options,” said Siddharth Sharma, Head of IT Operations, Digi Yatra.
Why SMRs?
“While we’re moving towards green energy, it’s not yet possible to meet 100% of our needs through traditional renewable sources. That’s why nuclear energy, especially in a modular form, is gaining traction. The energy demand is increasing exponentially, and nuclear energy seems to be the most practical solution to address this growing gap,” shared Anil Pawar, Chief AI Officer, Yotta Data Services. Therefore, “the most promising alternative seems to be nuclear power,” mentioned Sharma.
SMRs offer several advantages that make them an attractive solution for powering AI data centres:
– High energy density: SMRs produce a significant amount of energy in a compact form. This allows them to generate a substantial amount of energy, which is perfect for data centres that need a reliable and consistent power supply.
– Reliability and stability: Unlike renewable energy, SMRs provide continuous power without fluctuations, ensuring that AI data centres run smoothly around the clock.
– Low carbon emissions: SMRs are a clean energy source, generating minimal greenhouse gases, helping AI data centres reduce their carbon footprint.
– Scalability: SMRs can be scaled to meet the specific needs of a data centre, from smaller facilities to massive, enterprise-level operations.
– Shorter deployment time: Compared to traditional large-scale nuclear plants, SMRs can be deployed faster, which is crucial for meeting urgent energy needs in rapidly expanding sectors like AI.
– Enhanced safety features: Modern SMRs are designed with advanced safety mechanisms, reducing risks of accidents or meltdowns.
– Cost-effective: Over time, the operational and energy costs of SMRs could be more economical than the volatile costs of fossil fuels or the intermittent supply of renewables.
Global initiatives
“Many countries, including the U.S. and China, have been exploring SMRs. The U.S. launched an initiative called NuScale, aiming for small nuclear plants under 100 megawatts,” shared Sharma. Constellation Energy has signed an exclusive deal with Microsoft to restart one of the units at the Three Mile Island nuclear plant in Pennsylvania, providing 835 megawatts of energy to Microsoft’s data centres. The UK, Russia and Sweden have also begun research and development in SMR technology, recognising its potential to transform energy landscapes and power critical infrastructure, while China and Canada plan to go live with their projects by 2026 and 2028 respectively.
Is India ready?
India has already established itself as a nuclear power, with a long history of nuclear energy development. The country is also committed to expanding its clean energy capacity to meet its sustainability goals. In the context of data centres, India has the potential to leverage SMRs to address the growing power demands, especially as the government pushes for a more robust digital economy.
Feasibility studies have been conducted on the use of nuclear power for industrial purposes, but there is still much work to be done to translate these ideas into reality for data centres. The Indian government’s initiatives around clean energy, combined with its goal to increase data centre capacity, present an opportunity to explore nuclear as a dedicated power source for these critical facilities.
The Central government is exploring the feasibility of providing dedicated, uninterrupted power to large data centre parks through direct lines from power producers or by setting up small nuclear plants nearby. This initiative aims to support the growing energy demands of India’s digital infrastructure while ensuring sustainability and reducing carbon emissions. The government is also committed to making India a global data centre hub, with significant investments in energy-efficient, sustainable, and green data centres.
State governments, such as Karnataka, are keen on adopting sustainable practices to ensure that the growing tech industry remains eco-friendly. “I need to ensure we follow sustainable practices, particularly in energy and water. We’re exploring options like setting up microgrids to feed directly into the grid,” said Priyank Kharge, Minister for IT & BT and Rural Development & Panchayat Raj, Govt. of Karnataka. This sentiment echoes a broader push across the country for innovative energy solutions, including nuclear energy, to meet the increasing demands of AI-driven data centres.
Challenges and risks
Despite the potential, adopting SMRs for AI data centres is not without challenges. Regulatory hurdles remain one of the biggest obstacles, as nuclear energy projects often face stringent approval processes. Public scepticism surrounding nuclear safety is another concern, especially in densely populated areas where the idea of a nuclear reactor might face resistance.
“Setting up nuclear reactors for data centres will require new infrastructure, policies, and a specialised skill set. There are potential side effects, and we need to address all of these factors carefully,” mentioned Pawar.
The initial capital investment required to build SMRs is also high, which could deter stakeholders looking for short-term returns. Additionally, the risks associated with nuclear energy, such as waste management and long-term environmental impacts, will need to be addressed before SMRs can gain widespread acceptance.
However, the future of AI data centres is closely tied to the ability to secure reliable, sustainable, and cost-effective energy sources. SMRs present a promising solution that could revolutionise the way we power these data centres, offering high energy density, low emissions, and unmatched reliability. While challenges remain, particularly in terms of regulatory frameworks and public perception, further research and investment in SMR technology could pave the way for a cleaner, more efficient energy future. With India’s data centre capacity set to grow exponentially, exploring SMRs as a power source could position the country as a global leader in both AI and nuclear innovation.
