In February 2026, just over a week before India hosted the AI Impact Summit, New Delhi announced that it had signed digital public infrastructure (DPI) cooperation agreements with 23 countries. Amid bipolar geopolitical contests, India’s DPI diplomacy can be viewed as extending infrastructural soft power internationally.
India’s DPI, known as India Stack, comprises digital identity, digital payment, service delivery platforms and data governance and exchange. It supports several mobile apps for public service delivery, including DigiLocker for official documents and DigiYatra for air travel.
The success of these efforts has encouraged India to pursue ‘DPI diplomacy’ since its G20 presidency. This model has attracted interest in international development circles and across the Global South. As competition between the United States and China extends to critical minerals and the raw materials needed for digital and energy infrastructure, growing interest in DPI is timely. Countries seeking to build digital rails for public administration are increasingly concerned about digital sovereignty, and DPI seems to offer a suitable solution. Attention is also turning to developing AI capabilities using the building blocks that DPI provides.
The convergence of AI and DPI could reduce language barriers, improve public-service delivery efficiency and make services more accessible to citizens. The India AI Governance Guidelines outline a medium-term plan to integrate DPI and AI, with security built into systems designed for scalability and interoperability. The guidelines also envisage the development of multilingual AI applications such as Bhashini, the Indian government’s translation platform.
The story of India Stack began in the mid-2000s when the Indian government decided to implement a digital identity system to ensure that welfare and social-security benefits reached impoverished families. This developed into Aadhaar, a biometric database containing the fingerprints and iris scans of the vast majority of India’s population. This occurred as awareness grew of data’s economic value and its ability to enable surveillance. As the scale and potential of the digital data economy became more widely understood, Aadhaar’s importance became increasingly apparent.
The election of Prime Minister Narendra Modi’s government in 2014 marked another turning point. Despite criticising Aadhaar while in opposition, Modi became a strong supporter of the biometric identification system. During his tenure, Aadhaar expanded alongside the Unified Payments Interface and other digital products built on India Stack’s underlying infrastructure. During India’s 2023 G20 presidency, DPI became a central feature of its digital diplomacy, including through initiatives such as the Global DPI Repository.
Rebranding the much-maligned Aadhaar alongside the other more successful applications enabled by India Stack has extended India’s infrastructural soft power. Infrastructural power refers to a state’s ability to control and govern its territory. But India’s promotion of DPI on the international stage — and especially among Global South countries — extends India’s unique soft power by making an attractive policy offering rather than imposing a coercive outcome.
This marks the rise of India’s infrastructural soft power, where government technology developed for the country’s sovereign territory is being offered as a depoliticised software service available for international adoption. Infrastructural power is often tied to state power. Yet combined with the concept of soft power, it can also explain how states extend their influence beyond their territorial boundaries.
DPI is presented as a depoliticised technology stack that can be adopted in different political systems and institutional settings without any ideological requirements. But its use raises questions about personal data collection. Evidence of privacy breaches, surveillance and the sale of personal biometric data in India raise concerns about how DPI would be used in countries with different levels of state control.
India’s data governance framework emerged from a long consultation process on non-personal data in 2017 and the Personal Data Protection Bill in 2019 and culminated in the Digital Personal Data Protection Act 2023. Even as the regulation is slowly being implemented, concerns around personal data remain, as illustrated by the Kudankulam Nuclear Power Plant data leak. DPI cannot simply be exported without regard to local context, especially as interoperability between data systems and agentic AI is increasingly desired but almost impossible to regulate and govern.
The convergence of DPI and AI is opening up further governance challenges as agentic AI becomes more widespread. Applications built on a shared technology stack will need to balance openness and interoperability with clear rules on liability, data provenance and developer verifiability. This challenge will become more pressing as AI systems increasingly influence agent-to-agent interactions. As a new era emerges in which AI agents generate a growing share of web traffic and activity, there is an urgent need to define what makes DPI genuinely public.
Preeti Raghunath is Lecturer in Digital Media and Society at the University of Sheffield.