FondationFAM Numérique 5ème cycleNumérique 5ème cycle
DonateDons
Chapitre 5Chapter 5 Part 2 · Digital strategies in India and the Global South Partie 2 · Stratégies numériques en Inde et dans le Sud Global

India Stack, another third way

Translated from the French, which remains the reference text. This translation is under review.

"The Indian experience shows a 'third way' of developing digital infrastructure, different from the more laissez-faire US approach or the more centralised, interventionist Chinese approach. In partnership with the private sector, the Indian government provided critical digital infrastructure to kick-start the digital economy (focused on micro-payments), ensured level competition, and provided tools to the poor and the informal sector to participate in the economy."

— Cambridge Judge Business School, DPI Playbook

5.1The context: systemic financial exclusion and the challenge of scale (Scale)

To understand the urgency of creating a digital public infrastructure, one has to go back to the socio-economic reality of India at the end of the 2000s. The country was then mired in an informal economy, dominated by cash, with very limited physical infrastructure (banks, roads, telecommunications). The central problem was not only poverty but administrative invisibility: hundreds of millions of Indians held no verifiable official identity document, which in effect excluded them from the banking system and fostered massive corruption in the distribution of the state's social benefits (leakages).

The International Monetary Fund (IMF) sets out the finding of this original exclusion:

"Verifiable identity, or an ID certifying that 'I am who I am,' is a key element of any economy and its level of financial inclusion. In 2008, only 1 in 8 Indians had a verifiable identity. [...] It is globally documented that, absent tangible collateral, a large majority of the adult population cannot borrow from the financial system. [...] The foundational layer of the India Stack focuses on addressing the absence of verifiable identity, which is the most important constraint to inclusion in a large, diverse and informal economy like India."

— IMF, Working Paper WP/23/78

Nandan Nilekani, a founding figure of the technology giant Infosys and the future architect of India's digitisation, understood very early that the problem could not be solved by following the pace of physical development of the Western countries. Digital technology had to be used to reach what he calls Population Scale (the scale of the entire population):

"I think that if we want to move the needle in India; if we really want to make an impact; if we want more development; if we want to actually change the lives of people, we have to think at scale because just having a small solution which works in a little pilot is not going to change anybody. [...] In India, unlike markets like Europe, is a common single market for services and a fragmented market for goods. This meant that if you roll out something which changes banks, the whole country can roll out the ID... So, I realised that digital services at national scale is where one can have the most impact."

— Nandan Nilekani, YourStory interview

5.2The genesis: Aadhaar, the JAM Trinity and the invention of "India Stack"

The history of India Stack truly begins with a strong political will in 2009. The Indian government decided to tackle the problem of identity by creating the UIDAI (Unique IDentification Authority of India), a state agency operating with the agility of a start-up, and appointed Nandan Nilekani to head it. Its mission: to give more than a billion residents a unique biometric identity (fingerprints, iris) named Aadhaar (which means "foundation" in Hindi).

The success of Aadhaar led the government to couple this digital identity with a vast programme of free bank account opening (Jan Dhan), against the background of the explosion of mobile telephony. This was the birth of the JAM Trinity, which brings together Jan Dhan, the government programme for the mass opening of free bank accounts (Pradhan Mantri Jan Dhan Yojana, launched in 2014), Aadhaar, the unique biometric digital identifier assigned to every Indian resident, and Mobile, the mass spread of mobile telephony, which gives everyone a channel of access.

"India’s digital public infrastructure did not emerge overnight. It was seeded through a deliberate convergence of identity, banking and connectivity. This convergence took shape as the JAM trinity. Jan Dhan bank accounts, Aadhaar enrolment and widespread mobile phone penetration created the base layer for India’s digital transformation. Together, they connected individuals to the state in a direct and verifiable manner."

— Government of India, Press Information Bureau

It is worth noting that "India Stack" was not, at the outset, an official government programme. The concept of a unified technology stack was forged by the iSPIRT foundation (Indian Software Product Industry Round Table), an influential volunteer think tank made up of IT experts (among them Dr Pramod Varma, chief architect of Aadhaar).

"The term 'India Stack' was coined by iSPIRT, a technology industry think tank co-founded by Sharad Sharma in 2013. iSPIRT positioned the branding to describe separate government systems as a unified, exportable model of 'digital public infrastructure' (DPI). [...] Each component (Aadhaar under UIDAI, UPI under NPCI, DigiLocker under MeitY) is separately governed; the unified branding suggests cohesion and openness."

— Institutional documentation / Wikipedia

5.3The vision: Digital Public Goods and architectural minimalism

The founding vision is to treat the digital primitives (identity, payment, data) not as exclusive commercial products but as Digital Public Infrastructure (DPI), accessible to all through open APIs (application programming interfaces). The aim is to make every Indian transaction presence-less, paper-less, and cash-less.

"India Stack is: a vision for the world / a vision for India; a vehicle for financial and social inclusion; national plumbing for the Internet Age; a collection of open APIs empowering people and businesses; inspired by the principles of open-source software and interoperability. India Stack is the moniker for a set of open APIs and digital public goods that aim to unlock the economic primitives of identity, data, and payments at population scale."

— Official manifesto, IndiaStack.org

The architecture is deliberately minimalist. Dr Pramod Varma theorises that a public infrastructure should provide only a base building block on which others can build.

"How do public infrastructure projects differ from commercial products? Public infrastructure projects, unlike commercial products, act as foundational layers. They are utilities meant for widespread use and integration, not standalone solutions. This minimalistic approach ensures they remain flexible and adaptable for various applications across sectors. [...] When we designed Aadhaar, UPI and other platforms, we aimed to create building blocks that others could use to innovate. [...] By asking for only the most essential information, we reduced the barriers to entry and made the system more accessible."

— Dr Pramod Varma, Infosys Finacle interview

Schema 5.1 · Digital public infrastructure read as a public service

In the interview he gave us for this investigation, Henri Verdier recounted an exchange with Pramod Varma, chief architect of Aadhaar and of India Stack. The latter told him, in substance : Europeans built their prosperity on large infrastructures, roads, railways, water, schools, on which the economy and society were able to develop ; then, when the digital came, the West stopped thinking of it as an infrastructure the public sector could build. Identity and payments are, in his view, today's roads, and that is what India Stack builds. Henri Verdier draws from it the reading the schema presents : what India builds falls under public service, in the sense French law gives the term.

Public-service principle Question put to the architecture Equality of access Does a person get help when authentication fails ? Continuity Does a change of provider interrupt the service ? Mutability Does a new version keep an access for the least equipped ? the state guarantees the infrastructure ; others may operate it
Henri Verdier proposes to read digital public infrastructure through the French conception of public service, rather than through the commons alone. The three principles administrative law attaches to it — equality of access, continuity, mutability — become concrete questions for the architect. It is an interpretive analogy he puts forward, not a documented historical filiation.After the interview with Henri Verdier (27 August 2026) and the editorial dossier of 11 September 2026.See also : Schema 5.3, Users at the centre of India Stack.

5.4The strategy: the hourglass architecture and the doctrine of "Public rails, private innovation"

To deploy its vision at the scale of a billion people, India had to devise an unprecedented systemic engineering. Pramod Varma and the researchers of the Harvard Kennedy School (its Project on Digital Era Government) describe this approach as a shift "from Platforms to Protocols" (From Platforms to Protocols).

"The problem with platforms is [...] they own consumers' data and use it for their own benefit. In contrast, protocols give control and decision-making to the consumers rather than keeping it centralized."

— Harvard Kennedy School, From Platforms to Protocols

This approach rests on a formula that Nandan Nilekani sums up as follows: “public rails, private innovation”.

By refusing either to privatise the base infrastructure (the American model) or to monopolise the end applications (the traditional state model), India demonstrated that a public standardisation of the "digital plumbing" drastically lowered the barriers to entry and multiplied commercial innovation. Rahul Sanghi, in his essay The Internet Country (Tigerfeathers), sums up this stance perfectly:

"India has approached digitisation as a public good. [...] Instead of building state-owned monoliths, the Indian government laid down a set of open, interoperable protocols. The idea was simple but profound: the state provides the public rails, and private enterprise builds the trains that run on them."

— Rahul Sanghi, The Internet Country

Rahul Sanghi's rails and trains point to a model well known to network architects, inherited from the design of the Internet (TCP/IP): the Hourglass Architecture (Hourglass Model).

The neck of the hourglass, defined by the state, is extremely narrow: it consists of only a few minimalist and agnostic APIs (identity, payment, consent). Below this neck lie the many institutional databases and registries. Above it, the private market is invited to build a multiplicity of competing applications.

Schema 5.2 · The hourglass of public infrastructure

Above : the private market a multitude of competing applications, built by private innovators and operators payments · credit · wallets · agents The waist : the minimal common a protocol, rules of the game identity · payments · consent Below : what already exists the institutional databases and registers to which the protocol connects banks · administrations · public registers The three layers of the hourglass, after Pramod Varma Innovation Governance Technology What the investigation retains the waist is set by the state and kept as narrow as possible : minimise the surface of the common
Pramod Varma presents three layers : technology, governance, innovation, and sums up his rule as 'government minimum, market maximum'. They match the hourglass this chapter has just described : a bottleneck, its waist, the protocol and its rules, between two bells — above, the multitude of innovators and operators on which adoption and scale rest ; below, the existing registers and institutions the protocol connects to. Governance, which he singles out as a layer of its own, sits in the waist with the protocol.After Pramod Varma, Inria, 16 September 2026, slide 20 (© NFH 2026).See also : Schema 7.1, Openness comes from underneath · Schema F9.2, Public, private and citizen · Stratification, first of the five operators (§ 7.2) · Pattern no. 4, Minimising the surface of the common.

5.5The building blocks of India Stack

Schema 5.3 · Users at the centre of India Stack

the person, the business, the agent Who I am India Stack : identity, legal entities next : DeDi directories, identified agents What I do India Stack : payments, signature next : Beckn networks, Finternet What I have India Stack : credentials, data access next : tokenised assets, OpenCred Our reading infrastructure is read from the one it serves rule of the report 'Judged by those one serves'
This schema, proposed by Pramod Varma, presents the functions of India Stack, extended with other projects of his foundation, from a user-centred approach. The chapter that follows presents the India Stack blocks by technical family — identity, payments, data.After Pramod Varma, Inria, 16 September 2026, slides 8 and 13 (© NFH 2026).See also : Schema 5.1, Digital public infrastructure read as a public service.

Aadhaar: the identity and documents building block, "presence-less / paper-less"

Before 2009, proving one's identity in India meant travelling in person with paper documents that were often forged or non-existent. The aim of this building block is to allow universal authentication from any connected device.

The technological core

The foundation is Aadhaar (a unique 12-digit number coupled with biometrics: fingerprints and iris). But the real technological revolution is e-KYC (Electronic Know Your Customer). This API allows an institution to verify a customer's identity instantly by querying the state's database (the UIDAI answering "Yes/No"). To this are added e-Sign (the legally valid electronic signature) and DigiLocker (a cloud wallet into which the state and the universities "push" official documents directly, in digital form).

The ecosystem of applications and the users

Opening the e-KYC API to private-sector organisations set off an extraordinary growth shock. The cost of acquiring a bank customer, which involved paperwork and physical checks, fell from about 23 dollars to 0.10 dollar, at last making the financial inclusion of the poorest possible.

On the side of the government ecosystem, the government linked bank accounts to Aadhaar to create the Direct Benefit Transfer (DBT) system: agricultural subsidies and social benefits are paid directly, eliminating corrupt intermediaries and "ghost beneficiaries". On the side of telecoms, the revolution came from Reliance Jio: in 2016, the operator used Aadhaar e-KYC for the instant activation of SIM cards in the street. The customer places a finger on a reader, and the line is activated. This process allowed Jio to acquire a record 100 million customers in only 170 days, bringing 4G within reach of the Indian countryside.

"Nobody has ever acquired 100 million consumers in six months, and it was entirely because of the paperless, presence-less onboarding made possible by India Stack."

— Nandan Nilekani, YourStory interview

The payments building block: UPI and the "Cash-less" ecosystem

Designed by the NPCI (National Payments Corporation of India), this layer set out to democratise money transfer in an informal economy dominated by cash. It is today the most widely used real-time payment system in the world.

The technological core: absolute interoperability

UPI (Unified Payments Interface) connects every bank account in the country. Its major innovation lies in separating the money from the user interface. Your money stays in your bank, but you can use any application to initiate a transfer from one account to another, instantly, around the clock, free of charge (Zero MDR), through a simple identifier (e.g. name@sbi) or a universal QR code (BharatQR).

The ecosystem of applications and the users

UPI has generated a frenzy of commercial innovation. India alone accounts for nearly half (46 % in 2022, 49 % in 2023) of all real-time digital transactions in the world, that is more than 15 billion transactions a month at the end of 2024.

The state provided "the pipes" but left the battle for the consumer interface to the technology giants, the private "Super Apps" (the TPAPs). The market is a duopoly dominated by PhonePe (owned by the American company Walmart) and Google Pay, followed by the Indian Paytm. Competition is conducted exclusively on the quality of service. Such is the success that Google officially recommended to the US central bank (the Fed) that it draw on UPI for its own system (FedNow).

The ecosystem has reached down to street commerce, to the tea sellers (chaiwalas) and vegetable sellers who refused costly bank-card terminals. To overcome illiteracy, street noise and the fear of fraud, the industry invented the Soundbox: a small, very low-cost connected loudspeaker, set on the stall, which announces aloud (in Hindi, Tamil...) "Payment of 20 rupees received" as soon as the customer scans the QR code.

Schema 5.4 · Rails and interfaces : two levels at which to measure power

Interfaces : a multitude of services service 1 service 2 service 3 service 4 service 5 service 6 … The rails : identity, a standard and open payment format guaranteed by the state — no service can unplug another For the person change service obtain redress keep acting without losing access to essentials at the level of interfaces, concentration remains possible : distribution of applications, customer relationship, knowledge of transactions control of the rails and plurality of services are assessed separately
Henri Verdier reads the UPI market-share cap as a sovereignty instrument : no operator, however present in usage, can unplug the others, holds the data, and each can itself be unplugged. The investigation draws a distinction from it : control of the rails, which the state guarantees, and plurality of services, which rests on a multitude of operators, are assessed separately. Concentration of the application layer remains an open question, which the 30 % cap set by NPCI — not 35 %, the figure he recalled — precisely seeks to address.After the interview with Henri Verdier (27 August 2026), the rules published by NPCI and the editorial dossier of 11 September 2026.See also : Schema 5.2, The hourglass of public infrastructure · Schema F9.2, Public, private and citizen.

The data building block: DEPA and the "Consent-based" ecosystem

This is the most sophisticated and most recent layer of India Stack. The founding idea, championed by Nandan Nilekani, is that "Indians will be data-rich before they are economically rich". The aim is to use the data generated by Aadhaar and UPI (purchase history, bills) as "collateral" to prove creditworthiness, rather than requiring real estate.

The technological core: consent and the "blind pipes"

This layer rests on DEPA (Data Empowerment and Protection Architecture). The central innovation is the Consent Artefact (a standardised and revocable digital document) that governs an end-to-end encrypted data flow. The central bank (RBI) created a new category of company: the Account Aggregators (AA), conceived at the outset to ease access to credit, by allowing a customer to share their banking data with a lender. These start-ups act as "blind postmen" (Data blind). They collect the customer's consent on their smartphone, then instruct the customer's bank to send the encrypted data to a lender. The AA can technically neither read, nor store, nor resell the data.

The institutional ecosystem and the use cases

This ecosystem is federated by the Sahamati alliance (a self-regulatory body recognised by the state). It involves hundreds of banks holding data (FIPs) and requesting Fintechs (FIUs).

The first use case is flow-based micro-credit (cash-flow based lending). In the past, a small street trader could not obtain a conventional bank loan. Today, through the Account Aggregator network, the trader consents in one click to share the cryptographic proof of their small daily cash receipts on UPI. The lender's algorithm analyses this data and grants them a short-term micro-credit within minutes on their phone. The second is health: the state replicated this architecture in the medical field through the Ayushman Bharat Digital Mission (ABDM). A patient uses an application to transfer their X-rays or prescriptions instantly and securely (with their consent) from hospital A to hospital B, creating the largest interoperable health network in the world.

According to the Sahamati report of the end of 2024, this network has already enabled the sharing of more than 120 million consents and has facilitated the disbursement of nearly 10 million loans, a large share of them directed to very small enterprises (MSMEs).

The combinatorial power of "the Stack" (The Stack)

The magic of India Stack lies in the combinatorial innovation made possible by stacking them (Stacking).

A loan process that used to take 3 weeks, and required a journey in person, land collateral and 50 pages of photocopies, now unfolds in four steps. Identity, first: Aadhaar identifies me formally in 2 seconds. Data, next: DEPA proves my creditworthiness digitally to the banker by analysing my purchase history. The document: e-Sign has me sign the loan contract digitally. Payment, finally: UPI instantly transfers the money to my bank account. The whole takes 5 minutes, from any village, on an entry-level smartphone.

Artificial intelligence and DEPA Training: "Digital Non-Alignment"

The irruption of AI has posed a major geopolitical challenge: the American giants train their models by scraping the global public internet. India's real wealth, however, lies in high-quality "siloed" data (medical records, banking histories, regional languages) protected by professional confidentiality or by privacy.

So that India does not become a digital colony of Silicon Valley, the iSPIRT think tank has extended India Stack's data protection architecture (the Data Empowerment and Protection Architecture) towards DEPA Training and DEPA Private Inferencing.

The vision: from data protection to the production of a sovereign AI

Indian doctrine holds that the country must use its public infrastructure to create its own AI models, suited to its citizens (such as the Bhashini project for instant translation of Indian dialects).

"India doesn't wish to be just a trade colony of China or technology colony of the US. [...] What makes AI different is that it needs a whole-of-nation approach. [...] To maintain competitiveness, India must embrace AI not merely as a tool but as the very foundation of its economic strategy going forward. It must ensure that it is not just a consumer of AI but a critical creator of AI. In fact, it must aim to emerge as one of the 3 AI superpowers in the world."

— iSPIRT on strategic sovereignty, AI manifesto

The diagnosis is clear: to win the race, India has to unlock its institutional vaults without selling privacy short.

"Let's look at health. India has far more proprietary datasets than the US. It is just frozen in the current setup. Unfreezing this will give us a play in AI. This is exactly what DPI for AI is doing – in a privacy-preserving manner."

— iSPIRT

The technical architecture: the "Confidential Clean Rooms"

How can an AI be trained on banking or medical data without transferring it to OpenAI or Google? India Stack proposes a complete inversion of the paradigm: the data no longer travels to the AI's servers; the AI algorithm comes and runs inside the vaults.

To that end, DEPA-Training deploys the concept of hardware Confidential Computing.

"The era of scraping public data to train AI is ending. The next frontier demands access to high-quality datasets locked within institutional silos. DEPA Training unlocks AI model training on protected datasets — with privacy, sovereignty, and scalability built in from the start. Confidential Clean Rooms — hardware-level isolation where models come to data, not data to models. [...] Instead of transferring data to the consumer's environment, data is released into a Confidential Clean Room (CCR), a hardware-based secure execution environment where computation happens under strict policy enforcement. [...] Importantly, the clean room environment ensures that the Data Consumer (or for that matter the cloud provider) cannot observe or tamper with the data or the computation."

— Samyog / iSPIRT technical documentation

This architecture rests on smart contracts that trace the use of the data end to end, guaranteeing full legal compliance (India's new DPDP Act).

Governance: the "Techno-Legal" model and the Samyog entity

In the Confidential Clean Rooms, encryption is absolute. But who holds the encryption keys? The Indian ecosystem refuses to entrust those keys to the state (a surveillance risk) or to a private company (a monetisation risk). It has therefore incubated a new, neutral and not-for-profit Self-Regulatory Organisation (SRO), called Samyog. This is what India calls "Techno-Legal" governance (coding the law into the architecture).

"If the confidential clean room is where trust is enforced, key management is what ensures that this trust cannot be bypassed. In DEPA Private Inferencing, keys for encrypting data as it is transferred... are generated and governed within a dedicated, transparent key management service (KMS) operated by Samyog, a neutral not-for-profit self regulated organization (SRO) set up for this purpose. [...] This creates a separation of control, ensuring that no single participant in the ecosystem can unilaterally access sensitive data."

— DEPA documentation, on governance by Samyog

Adoption, pilots and prospective impact

Unlike ONDC, which is already in general use, DEPA-Training is currently at the stage of advanced pilots (test beds / sandboxes in 2024-2025). Samyog and iSPIRT are developing use cases ("Data Collaboratives") to solve problems at the scale of a billion inhabitants.

The Samyog documentation names three fields of application. Healthcare: "AI-driven diagnostics, drug discovery, and public health — powered by securely pooled clinical data across hospital networks", with a pilot run with IIT Madras to detect brain tumours through medical imaging pooled across competing hospitals. Finance: "Credit access for 400M underserved individuals and MSMEs, enabled by privacy-preserving collaboration across financial institutions", that is, allowing AI to assess the credit risk of the poorest by cross-referencing telecommunications and tax databases, without revealing their raw identity. Education, finally: "Personalised learning at scale for 250M students, with portable skill credentials for a future-ready workforce".

5.6The governance of India Stack

Unlike the United States (where the infrastructure is left to private monopolies such as Google or Apple) or China (where the state centralises everything), India opted for a "Third Way" resting on a well-conceived public-private partnership.

An initiative coming from civil society

Let us note first of all that the initiative for India Stack came from civil society: from a community of entrepreneurs, many of whom had made their careers in IT abroad, keen to put their skills and their ideas at the service of India and its people.

The stance of the state: facilitator and guarantor (Enabler)

The Indian state acknowledged that it had neither the agility nor the talent to code systems at the scale of a billion people. It therefore positioned itself as the legislative guarantor, the provider of seed capital, and the operator of the sovereign databases (biometrics), while delegating the software architecture to the private sector.

"This approach to data governance neither favors excessive state intervention nor is it exclusively laissez-faire. Such a combination of private and public features encourages better regulation and innovation. [...] This approach has enabled the public sector to move away from responsibility for end-to-end delivery."

— IMF, report Data by People, for People

The role of Nandan Nilekani: the "Grand Architect"

Co-founder of the IT giant Infosys, Nandan Nilekani is the keystone of the ecosystem. In 2009 he was appointed by the government (with the rank of minister) to lead the creation of Aadhaar. He brought the culture of India's Silicon Valley (Bangalore) into the heart of the New Delhi bureaucracy. Today, although he no longer holds a ministerial mandate, he is often seen as the unofficial Non-Executive Chairman of digital India.

"Nandan Nilekani... considers IndiaStack as 'India’s single most important innovation to formalise India’s domestic economy'. [...] He operates at the intersection of state power and private capital."

— I. Shah and A. Sukumar, The Stack and the State

The role of iSPIRT and the limits of the techno-volunteer model

Despite its Open Source aura, the governance of India Stack is regularly criticised for its technocratic opacity. The very concept of "India Stack" was not invented by the government but by iSPIRT (Indian Software Product Industry Round Table), an influential think tank of IT experts founded in 2013. Working with volunteer engineers, iSPIRT behaved as the shadow army of Indian digitisation, coding the APIs and persuading the government to adopt them.

This model, resting on an asymmetric partnership with "tech volunteers", nonetheless ended up showing its institutional limits when faced with questions of probity and democratic accountability (accountability).

The rift in the ecosystem: the emancipation of Beckn

Faced with this crisis of legitimacy, which threatened the international export of the model, Nandan Nilekani and Dr Pramod Varma (chief architect of Aadhaar) took a different path from iSPIRT's, by creating the Beckn Foundation (since become FIDE), charged with guaranteeing the strict neutrality of the future protocols.

Today the ecosystem diverges into two visions that confront each other on the economic ground. As the investigative outlet The Ken (2023) documents: iSPIRT is trying to push its own credit infrastructure (OCEN) in a "volunteer" manner, while Nilekani and Varma are deploying the government ONDC network in a more "institutional" manner.

The institutional puzzle: status and governance of the agencies

One of the fundamental rules of India Stack is institutional decentralisation. No single agency controls the whole stack. The parties involved fall into three circles.

The sovereign agencies, the regalian functions, first. MeitY (Ministry of Electronics and Information Technology) is the principal supervising ministry; it steers the National e-Governance Division (NeGD), which directly manages the "Paperless" layer (DigiLocker). UIDAI (Unique Identification Authority of India) is the government agency responsible for the identity layer (Aadhaar); initially a mere executive order, it became a powerful and independent "Statutory Authority" following the passage of the Aadhaar Act in 2016. The NHA (National Health Authority), under the aegis of the Ministry of Health, is responsible for the health layer (Ayushman Bharat Digital Mission - ABDM).

The hybrid actors, next, the Section 8 Companies, an Indian innovation. To combine private agility with a public-service mission, India relies on the status of the Section 8 Company (a private not-for-profit company that distributes no dividend). The NPCI (National Payments Corporation of India) manages the payment layer (UPI, RuPay); its official status describes it thus: "Founded in December 2008, the NPCI is a government organisation registered under Section 8 of the Companies Act 2013 (Not for Profit)... The organisation is owned by a consortium of major banks and has been promoted by the country's central bank, the Reserve Bank of India (RBI)." ONDC (Open Network for Digital Commerce), created at the end of 2021 by the government, is also a Section 8 Company, funded by banks (SBI, HDFC) and the NPCI, charged with breaking the monopolies of Amazon and Flipkart.

The regulators and private foundations, finally. Sahamati is an industry alliance (recognised as a Self-Regulatory Organisation by the RBI) that federates the actors of the data layer (Account Aggregator). FIDE (Foundation for Interoperability in Digital Economy), formerly the Beckn Foundation, is a purely private foundation created by N. Nilekani and P. Varma, which manages and develops the open-source Beckn protocol used by ONDC.

Governance rules and practices: the "Techno-Legal" model

Indian doctrine rests on a very powerful concept: techno-legal governance (Techno-Legal Approach). The architects consider that law alone (such as the GDPR) is slow and ineffective. The rules protecting privacy must be hard-coded into the IT architecture (by design). For example, the entities that transfer your financial data (the Account Aggregators) are under a legal obligation not to read it. The technological rule makes such theft impossible through end-to-end encryption.

A major critique: a "closed" governance disguised as "open"

Despite its Open Source aura, the governance of India Stack is deeply technocratic and undemocratic.

"Unlike open protocols such as HTTP, email, or TCP/IP where anyone can implement independently, India Stack requires permission at every layer. There is no public RFC process, no citizen representation on technical committees, and no mechanism to challenge design decisions before deployment. [...] In India Stack, if a proposal is rejected, there is no alternative—the gatekeeper's decision is final."

— Summary of the critiques (researchers and digital-rights activists cited on Wikipedia)

Schema 5.5 · What openness gives, property by property

Open code gives transparency who can read and use ? Open standard makes compatible who can connect ? Shared governance distributes decisions who decides, who contests ? Capacity to take over makes independent who can take the system over ? these properties may reinforce one another ; they are not the same for Henri Verdier, a commons is a solution whose governance is exercised by its contributors
Henri Verdier reserves the word commons for solutions whose governance is exercised by their contributors : free software decided by a single person is free, and admirable, without being a commons. He reads the Indian preference for open code as a choice of transparency rather than of co-management. The investigation draws a grid from it : open code, open standard, shared governance and capacity to take over are four distinct properties, to be established separately for each infrastructure.After the interview with Henri Verdier (27 August 2026) and the editorial dossier of 11 September 2026. His definition of a commons is narrower than the Ostrom one used elsewhere in the investigation.See also : Schema 7.5, Openness and verifiability · Schema 7.1, Openness comes from underneath.

5.7The assessment: 47 years of progress achieved in 6 years

The quantitative effects of India Stack are without precedent in the economic history of the world, compressing the country's development time.

On the sidelines of the 2023 G20, chaired by India, the World Bank published a document officially validating the Indian "miracle" in financial inclusion, banking costs and budgetary savings.

"JAM Trinity has propelled financial inclusion rate from 25% in 2008 to over 80% of adults in last 6 years, a journey shortened by up to 47 years thanks to DPIs. [...] Banks' costs of onboarding customers in India decreased from $23 to $0.1 with the use of DPI. [...] In the last decade, India has built one of the world's largest digital G2P [Government to Person] architectures leveraging DPI. [...] As of March 2022, this had resulted in a total savings of $33 billion, equivalent to nearly 1.14 percent of GDP."

— World Bank, G20 document relayed by the Press Information Bureau

The success is just as dizzying on the payment layer (UPI), run interoperably across all the banks and the third-party applications (Google Pay, PhonePe, Paytm).

"India accounted for 49% of all global real-time transactions in 2023. [...] Over 10 billion monthly transactions by 2023."

— IndiaStack.org / IMF

The IMF sums up the systemic benefit in its report Stacking up the Benefits:

"India's foundational DPI, called India Stack, has been harnessed to foster innovation and competition, expand markets, close gaps in financial inclusion, boost government revenue collection and improve public expenditure efficiency."

— IMF, WP/23/78

A systemic Fourth Way

From a macroscopic point of view, India Stack reads, beyond a "Third Way", as a Fourth Way in global digital governance. In the first way, that of the United States, the free market dominates and the infrastructure is centralised and privatised (Big Tech). In the second, that of China, the state dominates and the infrastructure is centralised and closed, for social control. In the third, that of Europe, the state regulates, essentially through Law, software infrastructures that it does not code (GDPR, Data Act). In the fourth, that of India and the DPIs, the state regulates essentially through Code: it builds a digital public infrastructure, based on protocols that structure and steer the market.

The dark side of scale (Scale): the violence of algorithmic exclusion

While the macroeconomic statistics validate India's technological success, academic field research sets against this triumphalism a far more nuanced reality. In imposing this model, the Indian government made Aadhaar biometric authentication compulsory for receiving food rations and vital benefits (the Direct Benefit Transfer system).

In the geography journal Area (published by Wiley), the researchers deconstruct the tragedies produced by this technological solutionism. The slightest technical failure translates into immediate physical violence: manual workers whose fingerprints have been worn away by labour, elderly people, or villages lacking the network connectivity to run the e-KYC API find themselves rejected by the system, with no human recourse possible.

"Such projects of algorithmic governance often deepen socio-spatial exclusions, as technological glitches and biometric failures disproportionately affect the poorest and most vulnerable bodies. Aadhaar transforms citizenship into a data point."

— Aadhaar and the making of a welfare data subject, Area (Wiley)

The assessment of India Stack therefore carries a paradox inherent in its engineering: an infrastructure coded for the absolute inclusion of the population simultaneously produces, through the intransigence of its algorithmic verification, a relentless exclusion of the individuals who do not fit the data standard.

Mission carried by the Fondation FAM. Mission portée par la Fondation FAM. ← Back to mission homeRetour à l'accueil de la mission www.fondation-fam.org — Chapter 5 — India Stack, another third way — published 2026-06-02 · updated 2026-10-28 www.fondation-fam.org — Chapitre 5 — L’India Stack, une autre 3ème voie — publié le 2 juin 2026 · mis à jour le 28 octobre 2026