Today, the term “digital public infrastructure” is used around the world to refer to a range of digital projects across different contexts. Over the past five years, this fuzzy concept was promoted by US funding bodies and international organisations to refer to a development model based on India’s Aadhaar digital ID, UPI payments system and data exchanges. Now, having achieved global popularity, it is being taken up and reinterpreted in countries around the world. As such, digital public infrastructure is at a turning point.
In the initial period of digital public infrastructure discourse, before 2020, the term was used in an ad hoc manner. Its meaning was not yet formalised and it was often used interchangeably with other similar terms. The acronym “DPI” had not yet emerged. This period was characterised by openness and improvisation.
In this phase, people using the term were generally not aware of others’ use of the term; they were operating separately, yet the term clearly indexed a broad zeitgeist. This phase coincides with the beginning of critiques of Big Tech’s power and with conversations around what a “public interest” internet might look like. The range of projects it referred to include systems for deliberative democracy and the public sphere, projects to increase governmental efficiency, to promote a domestic software industry, to improve socio-economic development outcomes, to share open public data, to build hardware and connectivity and to ensure cybersecurity.
Meanwhile, the Indian experiment with Aadhaar, UPI and digital platforms had been taking place since 2009. Since 2015 these platforms had been referred to collectively in India as IndiaStack. From 2019, they were called digital public infrastructures. This rebranding was troubling, for IndiaStack is not meaningfully public. It was first proposed by a software industry think tank as a system for giving the private sector programmatic access to state promoted platforms via Application Programming Interfaces, or APIs.
In both its design and outcomes, using public resources to enable corporate profits while avoiding public oversight, IndiaStack should accurately be called “digital public-private infrastructure”. Most IndiaStack platforms are run by private companies and lack public accountability. For example, the National Payments Corporation of India, which runs UPI, is excluded from the Right to Information law. What’s more, for the past decade, journalists, activists and academics have found that the implementation of these systems often did more harm than good. Such critical civil society input has been largely ignored by the state. Yet, in 2023, when India hosted the G20, it engineered a remarkable global consensus on DPI, one of the event’s most significant outcomes. In 2024, a global commitment to DPI was ratified as part of the United Nations’ Global Digital Compact. This global consensus was not only the handiwork of Indian diplomats, it was supported by international organisations and large institutional funders in the US. The Gates and Rockefeller Foundations, for example, have committed hundreds of millions of dollars to promoting DPI around the world. This money is spent on pilot projects around the world as well as on creating a large-scale research agenda for DPI at institutions like Oxford University and University College London. This research has sought to create an evidence base for the positive impacts of DPI while minimising its drawbacks. One article, co-authored by a prominent progressive economist, redefines “public” to refer not to state ownership but to an abstract conception of “public value”. This research programme has been remarkably effective, creating a retrospective justification for the success and publicness of the Indian model.
The shift to a consolidated, IndiaStack-influenced model of DPI heralds a narrowing down of what had been a capacious term. In effect, the consolidated version of DPI reduces “digital” to software, stripping it of its material and social dimensions. If in the early days the term digital public infrastructure encompassed hardware, connectivity and social infrastructures, now reflecting IndiaStack it had been whittled down to refer only to a thin layer of APIs, which enable different digital systems and databases to communicate with each other. Today’s DPI excludes hardware and connectivity. The devices people use, the networks they access, the data centres that process data all fall beyond the remit of DPI. In other words, if an Aadhaar authentication fails because of poor mobile connectivity, or a bank system is inaccessible due to a power cut, it’s not a DPI problem. Excluded also are social infrastructures: the community structures, human intermediaries and local organisations through which we access the digital. In India, the excision of these factors has led to extremely uneven outcomes.
Marginalised Indians have suffered – even starved to death – on account of Aadhaar’s mandatory imposition into government welfare, while the rollout of the UPI payments interface without any attention to digital and financial literacy has facilitated the emergence of an enormous scam industry. Yet, India’s richest man Mukesh Ambani credits access to the Aadhaar API for enabling Reliance Jio to onboard of over a million customers a day, making it the dominant player in Indian telecom today. Unfortunately, it is the very excisions that have caused DPI to fail in implementation that make it an attractive, low-cost vision of technological “disruption” palatable to a range of powerful actors.
DPI’s appeal outside India is a response to the distinctive global conjuncture we find ourselves in today. When Covid-19 pandemic and lockdowns led to a shutdown of face-to-face engagement for those who could afford to move online, the digital claimed a new salience, and there was a broad desire for infrastructures that could be digital and public. Meanwhile, the escalation of a new tech cold war between the US and China also led to a desire for an alternative way of building and operating digital technology. It is in this conjuncture that something called digital public infrastructure – which, according to the new evidence base constructed by international funders, had been successfully implemented by India – offered a compelling promise of a non-aligned digital future.
Today, the term DPI has captured the global imagination. The amount of institutional and financial capital that has been invested into it has ensured that it is here to stay, at least until this hype cycle fizzles out and everyone turns to the next big thing. But as it spreads everywhere from Europe to Latin America, DPI’s meaning is being stretched thin. There may be a global consensus around the term DPI, but there is new range in what it means in practice on the ground. At this moment, some of what was excised from DPI is reappearing.
In a Europe threatened by US Big Tech and Trumpian foreign policy, digital public infrastructure refers to a new “EuroStack” that seeks to orient massive investments into infrastructure. Notably, a recognition of US Big Tech’s control over global hardware is central to the EuroStack project. Meanwhile, the Open Future Foundation sees “public digital infrastructure” as the basis for a new public sphere, responding to the whittling away of sites for democratic debate and deliberation by Big Tech. The opposition to US Big Tech is a factor in Brazil’s embrace of the term. In Brazil, as in India, digital public infrastructure is also imbued with developmental goals, with delivering government welfare more effectively, and the term is synonymous with digital ID and payments systems. The difference is that Brazil’s ID and payments systems are built by a state with robust public oversight and with civil society participation. These new meanings that the term has picked up around the world are a positive development but they are not enough.
Today, no one actor has a monopoly over digital public infrastructure. Although it claims credit for it, the Indian government does not own DPI, and neither do the international organisations that have promoted it in backrooms around the world. We mustn’t delude ourselves: the powers that consolidated it are strong. But the term’s localisation presents a sliver of opportunity for those who desire a more just digital future.
“Digital public infrastructure” has become an effective way of naming a broad aspiration to break free from US Big Tech. It will inform the digital development trajectories of many countries around the world, especially in the global south. Even in India, where Aadhaar and UPI are fait accompli, it may not be too late to reimagine how DPI in sectors like health, education and agriculture are implemented.
In these cases, we can’t limit our ambitions to replacing US Big Tech with state surveillance and domestic crony capitalists. Infrastructures that are meaningfully digital and public must be grounded in the needs and aspirations of communities. They must account for the varying material and social conditions that mediate access to the digital in different places. In short, their development must be led by the people. If digital public infrastructures are to be worth the name, they must move beyond the Indian model.
This article is based on a recent paper published by the Open Future Foundation.
Mila T. Samdub researches the infrastructures and aesthetics of digital governance in India and beyond. He is a fellow with the Open Future Foundation and the CyberBRICS project at FGV Rio, and a visiting fellow at the Information Society Project at Yale Law School.
