At the Global Fintech Fest 2026, industry leaders argued that India’s AI strategy must move beyond data storage to full control over the technology stack. This shift requires massive capital for compute infrastructure and localized AI models to reduce dependence on foreign architectures.
At the Global Fintech Fest 2026 in Mumbai, top executives from India’s digital infrastructure and telecom sectors called for a major reorientation of the nation’s sovereign AI strategy. The consensus among industry leaders is that current policies, which focus primarily on data residency—keeping data within Indian borders—are insufficient to ensure true technological autonomy. Instead, experts are advocating for comprehensive control over the entire AI technology stack, from hardware and compute infrastructure to foundational models.
The Infrastructure and Capital Gap
Sunil Gupta, CEO of Yotta Data Services, emphasized that the most significant barrier to India’s AI ambitions is the massive capital spending required for computing power. Building a sovereign AI ecosystem is not just about writing code; it requires heavy investment in data centers, energy supply, and high-performance GPU clusters. According to industry views, India’s domestic capital markets must increase their appetite for this type of infrastructure-heavy risk. Without significant investment in this 'brute computing power,' the country risks falling behind and ceding competitive advantages to international technology giants.
Defining True Sovereignty
Rahul Vatts, Group Chief Regulatory Officer at Bharti Airtel, outlined a multi-dimensional framework for what true cloud and AI sovereignty should look like. He argued that keeping data within Indian borders is merely a starting point. A robust sovereign AI strategy must be built on five pillars: data residency, digital infrastructure control, operational independence, jurisdictional clarity, and technological sovereignty. The industry view is that India currently faces a notable gap in technological sovereignty, meaning the ability to develop and own the underlying AI architecture rather than relying on foreign-designed systems.
Localizing AI for Efficiency
Beyond infrastructure, there is an urgent need to address the efficiency of AI models used within Indian enterprises. Rishi Bal, CEO of BharatGen, highlighted that relying on Western-centric AI models creates two problems: higher costs and linguistic inefficiency. Because these models are often trained on English-dominant data, they require more computational resources to process Indian languages, leading to higher inference costs for local businesses.
To bridge this gap, initiatives like the government-backed 'Param 2'—a 17-billion parameter multilingual foundation model—aim to provide culturally relevant and cost-effective AI solutions. For investors, this shift toward locally developed models could prove vital. It offers a way for businesses to maintain security and compliance while reducing the overhead costs associated with foreign-designed AI platforms.
Investor Monitorables
The pivot toward sovereign AI creates both opportunities and risks. The primary challenge remains the high cost of execution and the ongoing need for capital investment in data centers and specialized hardware. Investors tracking this sector should watch for further developments in domestic compute infrastructure, the adoption rate of locally trained foundation models by large enterprises, and whether government policy begins to favor incentives for hardware and semiconductor-related infrastructure over simple software-based applications.