India's efforts to shore up its artificial intelligence (AI) infrastructure will bolster, rather than cannibalise, established Southeast Asian tech hubs like Singapore and Malaysia, according to industry leaders from the subcontinent. Speaking at the Gitex AI Asia 2026 conference in Singapore, tech executives at the forefront of India’s AI infrastructure boom outlined how the country’s massive scale will serve as a testing ground for the broader Asian market.
Sunil Gupta, co-founder and chief executive of Indian datacentre giant Yotta Data Services, and Jay Chandan, chairman and CEO of Gorilla Technology, a UK-based supplier of AI-powered smart city applications, recently inked a landmark agreement to deploy thousands of graphics processing units (GPUs) across India. During a fireside chat, the two leaders discussed the details of the roll-out, driven by New Delhi’s push to build sovereign AI capabilities to protect national data and cater to domestic needs.
Sovereign AI and data sovereignty
With a population of 1.4 billion, including a billion smartphone users connected to the internet, India currently accounts for over half of the world’s digital payment transactions. This has led to an increased demand for processing and storing data within the country’s borders. And with the growing adoption of AI in recent years, users are now worried about what will happen to their data, said Gupta, particularly in terms of privacy and security concerns regarding how their information is used and managed by AI systems. “People want sovereign AI and sovereign models trained on sovereign data,” he said. “That’s a huge wave in India right now, supported fully by the government.”
Through the state-backed IndiaAI Mission, the Indian government is heavily subsidising computing costs, paying infrastructure providers to allocate GPUs to local model builders, researchers and academia. “We’re looking at about 5,000 GPU cards to be deployed for our AI workloads in the first six months,” said Chandan, adding that the partnership aims to eventually scale up to 36,000 GPUs. Under the agreement, Gorilla will provide the GPU infrastructure, while Yotta will operate the GPUs at its Navi Mumbai datacentre to deliver AI compute services, including GPU clusters, bare-metal GPUs, AI lab workstations and AI model endpoints to enterprises and government customers.
India's push for data sovereignty is also driven by its recently enacted Digital Personal Data Protection Act, which requires sensitive personal data to be stored and processed within the country. This legislation, combined with the IndiaAI Mission, creates a powerful incentive for global companies to locate AI infrastructure in India. The country already hosts major hyperscale datacentre investments from players like Amazon Web Services, Microsoft, and Google, and the addition of sovereign AI compute capacity further cements India as a data and AI processing hub.
Solving the ROI challenge
Gupta noted that while many enterprises have developed AI use cases across a range of industries, including finance, media, entertainment and manufacturing, not all have made it to production. “The CFO [chief financial officer] is not convinced,” he added. “If I invest in these GPUs, models, datasets and skill sets, will I get a return? We have to bring down the cost so they can cross that line into production. If they start experimenting and become successful, then they’ll scale up.”
By offering GPU infrastructure through an elastic, low-cost consumption model, Yotta and Gorilla aim to make enterprise AI commercially viable, generating returns on investments in three to five years, said Chandan. The model is similar to cloud computing's pay-as-you-go approach but tailored for AI workloads. It allows companies to test model training without upfront capital expenditure, which has been a major barrier for small and medium-sized enterprises in India and across Asia. The partnership also provides AI lab workstations and pre-configured model endpoints, reducing the time needed to deploy AI applications from months to weeks.
The challenge of return on investment is not unique to India. Globally, many organisations have struggled to move from AI experimentation to production due to high infrastructure costs and uncertain business outcomes. According to recent surveys, only about 20% of AI projects reach full-scale deployment. By lowering the cost of compute, the Yotta-Gorilla initiative hopes to raise that percentage significantly, creating a ripple effect across industries such as healthcare, supply chain logistics, and agriculture.
India’s role in the global AI ecosystem
However, the build-up of India’s AI infrastructure has raised questions about whether the subcontinent could syphon tech investments away from Southeast Asian digital hubs. “In all the meetings I’ve had, people ask, ‘Is India going to replace Singapore, Malaysia and Vietnam?’ That’s not going to happen,” said Chandan. “India is not here to replace anybody. India is here to help you build scale and velocity. It’s here to show you that you can build these large-scale models, and you can be successful with an efficient cost base.”
Gupta added that India’s sprawling datacentres are also helping to solve global supply chain challenges. He revealed that due to GPU shortages elsewhere, enterprises from Europe and the Middle East are increasingly looking to India to host their AI training and inference workloads. “Because India is geopolitically safe compared to many other areas, it has the potential to become a major hotspot for serving global AI demand,” he said.
The GPU shortage, driven by skyrocketing demand for AI, has created bottlenecks in regions like North America and Europe. Export controls on advanced chips have further tightened supply, making India an attractive alternative. The country's stable government, improving power infrastructure, and proactive policies like the IndiaAI Mission contribute to its appeal. Moreover, India’s vast pool of engineering talent ensures that GPU deployments can be effectively managed and optimised, reducing the total cost of ownership.
Looking at specific use cases, the Yotta-Gorilla infrastructure will support training of large language models (LLMs) for Indian languages, computer vision for manufacturing quality control, and predictive models for financial services. The Indian government also plans to use sovereign AI for public services such as healthcare diagnostics, education, and agricultural advisory. By hosting this compute capacity locally, the government ensures data privacy and reduces latency for applications that require real-time processing.
The partnership’s focus on Navi Mumbai is strategic. The datacentre campus is located in a region with robust power availability, proximity to undersea cable landing stations, and cooling advantages due to its coastal climate. Yotta’s facility already hosts tens of thousands of servers, and the addition of GPU clusters for AI will triple its compute capacity over the next two years. Chandan noted that the scalability of the operation allows for rapid expansion as demand grows, potentially making India a primary hub for AI compute in the Indo-Pacific region.
In parallel, Indian startups are leveraging the subsidised compute to build innovative AI solutions. For example, agricultural tech startups are developing crop disease prediction models, while fintech firms are building fraud detection systems. The IndiaAI Mission also funds research in ethical AI and bias mitigation, ensuring that the models trained on sovereign data adhere to Indian social and cultural norms.
From a global perspective, the IndiaAI Mission aligns with similar efforts in other nations, such as the European Union’s AI Act and Japan’s AI strategy. By creating its own sovereign AI infrastructure, India can set standards for data governance that other developing countries might adopt. This has long-term implications for how AI technologies are shaped, ensuring that diverse perspectives are represented in the training data and model outputs.
The GPU deployments by Yotta and Gorilla are just the beginning. Other Indian datacentre providers, such as CtrlS and Nxtra, have announced similar plans to expand AI-ready capacity. Nvidia, a key supplier of GPUs, has strengthened its partnerships with Indian companies, providing both hardware and software stacks. This ecosystem approach ensures that India not only hosts AI workloads but also develops local expertise in model development, fine-tuning, and deployment.
As India continues to invest in sovereign AI, the rest of Asia is watching closely. Singapore, with its strong research ecosystem, can complement India’s scale by providing advanced R&D and financing. Malaysia and Vietnam, both experiencing datacentre booms, can serve as regional redundancy sites. Far from competing, these hubs can form a distributed network of AI infrastructure, with India acting as the heavy-lift engine for training large models, and Southeast Asian hubs handling inference and edge computing.
In summary, India's push for sovereign AI is not an isolationist move but a pragmatic strategy to meet domestic needs while opening new opportunities for regional collaboration. The combination of policy support, private investment, and technical expertise positions India as a catalyst for the entire Asian AI ecosystem. As Gupta stated, “The wave is huge, and it’s going to lift all boats.” The challenge now is to sustain momentum and ensure that the benefits of sovereign AI reach every sector and citizen.
Source: ComputerWeekly.com News