Most governments digitize services. They take existing paper processes and put them online — portals, dashboards, digital forms. The result is faster versions of the same analog processes, not fundamentally new capabilities.
India did something different. Instead of digitizing services, it built digital infrastructure — reusable primitives that banks, hospitals, transport systems, and citizens all plug into. The result is not just faster government services. It is a transformed economy that leapfrogged to a $3 trillion digital economy in under a decade.
The architecture behind this transformation offers lessons for any leader building platforms at scale — whether for a nation or an organization.
The Architecture: Three Layers
India’s Digital Public Ecosystem follows a deliberate structure:
- Digital Public Infrastructure — universal rails for identity, payments, consent, and data. Population-scale APIs, the backbone that everything else builds on.
- Digital Public Platforms — sector-specific layers built on top of the infrastructure. Multi-sided systems where governments, companies, and citizens interact on a shared backbone.
- Digital Government Services — citizen-facing applications that deliver billions of transactions monthly. They look simple. Underneath them run identity, payments, consent, and registries at national scale.
The genius is in the ordering. Infrastructure first. Platforms second. Applications last. This is the opposite of how most organizations build — and it is why India’s system scales while most government digitization efforts stall.
1. Start With Infrastructure, Not Applications
India did not begin with portals or dashboards. It began with reusable primitives — foundational capabilities that every subsequent system would need.
- Aadhaar — digital identity and authentication. A way to prove who you are without documents.
- UPI / BBPS — real-time, interoperable payments. A way to move money between any bank account instantly.
- DigiLocker / eSign — paperless documents and secure digital signatures.
- DEPA / Account Aggregator — consent-driven data sharing. A way for citizens to control who accesses their data and for what purpose.
Each of these is a Digital Public Infrastructure block. They are not applications. They are capabilities. They do not solve a single use case. They enable every use case that depends on identity, payments, documents, or data sharing.
The decision to build infrastructure before applications is the most important architectural choice in the entire system:
- Infrastructure is hard to build and invisible when it works.
- Applications are easier to build and visible immediately.
- The temptation is always to start with the visible layer.
- India inverted this — and the applications that came later were dramatically simpler and more powerful because the infrastructure existed beneath them.
What works better:
- Before building any application, identify the reusable capabilities it depends on.
- Build those first. The application will be easier to build, and every subsequent application that depends on the same capabilities will be faster to ship.
- Infrastructure amortizes. Applications consume.
2. Layer Domain Platforms on Top of the Rails
Once the infrastructure existed, India added sector-wide platforms that operate on top of the identity, payments, and data-sharing rails.
- GSTN — transformed tax compliance into a unified digital system.
- ONDC — created an interoperable commerce protocol. Not a marketplace, but a protocol that any marketplace can plug into.
- ABDM / UHI — created a health data and services layer connecting hospitals, clinics, labs, and insurers.
- GeM — digitized government procurement.
- e-NAM — connected agricultural markets.
These are Digital Public Platforms. They are multi-sided systems where governments, companies, states, and citizens interact on a shared backbone. They are not applications — they are protocols and marketplaces that enable applications to be built by anyone.
The key design choice: these platforms are open.
- Any participant can join.
- Any application builder can integrate.
- Competition happens at the application layer, not the platform layer.
- The platform provides the rules and the rails. The market provides the innovation.
What works better:
- Design for openness. The platform should define the protocol and the rules of participation.
- Do not try to control the applications built on top of it.
- The more open the platform, the more innovation it attracts — and the more value it generates for the ecosystem.
3. Deliver Citizen-Scale Applications Last
Only after the infrastructure and platforms were in place did India build the applications that citizens recognize.
- BHIM — mainstreamed UPI payments into a simple mobile app.
- UMANG — single-window access to hundreds of government services.
- CoWIN — managed over 2 billion vaccine doses with transparent, real-time tracking.
- Parivahan — digitized transport services.
- DigiYatra — facial-recognition airport entry.
- e-Sanjeevani — telemedicine at national scale.
These applications look simple. That simplicity is deceptive. Underneath each one runs identity, payments, consent, registries, and data infrastructure at national scale. The applications are the visible tip of a massive architectural iceberg.
The order matters. If India had started with these applications:
- Each one would have required building its own identity system, payment system, and data-sharing mechanism.
- They would have been slower to build, more expensive to operate, and impossible to integrate with each other.
- By building the infrastructure first, every application benefited from capabilities that already existed.
What works better:
- Let the application layer be the last thing you build.
- Invest in infrastructure and platforms first.
- When you finally build the application, it will be dramatically simpler, faster to ship, and easier to connect to other systems.
- An application on shared infrastructure is not just cheaper to build — it is more powerful because it participates in an ecosystem.
4. Sovereignty Is the Invisible Layer
Underpinning everything is a design principle that is invisible to users but essential to the system’s longevity: sovereignty.
The infrastructure is designed for:
- Data residency — citizen data stays within India’s regulatory jurisdiction.
- Regulatory control — the state can set rules for participation, privacy, and security.
- Peak-event resilience — systems like CoWIN handled billions of transactions without failure.
- Cost-sharing across ministries — the same identity infrastructure serves tax, health, transport, and commerce.
- Avoiding single-vendor dependence — the architecture is open and interoperable by design.
None of this runs on spreadsheets or ad-hoc SaaS. This is infrastructure built to last decades, not funding cycles.
What works better:
- Sovereignty is not just a national concern. Every platform investment should consider who controls the critical layers.
- If your platform depends on a single vendor, a single cloud provider, or a single technology stack that cannot be replaced, you have traded long-term control for short-term convenience.
- Build control into the architecture, not just the contract.
5. The Real Breakthrough Was Structure, Not Technology
The most important lesson from India’s digital transformation is not technological. It is structural.
The architecture — foundational infrastructure first, domain platforms second, citizen applications third — is the breakthrough. This structure:
- Prevents monopolies at the rail layer by keeping critical infrastructure open and interoperable.
- Accelerates startups at the edge by giving them access to world-class identity, payments, and data infrastructure.
- Lowers costs for government by sharing infrastructure across ministries instead of building duplicative systems.
- Creates competition in services by allowing multiple applications to compete on the same platform.
- Keeps the state in control of critical infrastructure while enabling private-sector innovation at the application layer.
This is strategic architecture. It is designed for scale, trust, longevity, and innovation — not for the next budget cycle.
What works better:
- The structure matters more than the technology. The ordering of investments — infrastructure before platforms before applications — determines whether the system scales or fragments.
- The openness of each layer determines whether it enables an ecosystem or creates a bottleneck.
- The control over critical infrastructure determines whether the system serves its intended purpose over decades.
What I’ve Learned
Five things from India’s digital infrastructure that apply to any platform strategy:
Build infrastructure before applications. The visible layer gets the attention. The invisible layer creates the leverage. Invest in reusable primitives first — identity, payments, data, consent. Every application that follows will be faster and cheaper to build.
Design the layers deliberately and keep them separate. Infrastructure, platforms, and applications serve different purposes and need different governance models. Mixing them creates bottlenecks and monopolies. Keeping them separate enables competition and innovation at every layer.
Openness at the platform layer attracts innovation. A platform that controls everything limits what can be built on top of it. A platform that defines protocols and gets out of the way enables an ecosystem. The platform captures value through participation, not control.
Sovereignty is an architectural concern, not just a policy one. Control over critical infrastructure must be designed into the system. If you cannot leave a vendor, replace a component, or set the rules of participation, the architecture has failed — regardless of how well it works today.
Structure determines outcomes more than technology. The ordering of investments, the openness of each layer, and the control over critical infrastructure matter more than the specific technologies chosen. Get the structure right, and the technology choices become easier. Get the structure wrong, and no amount of technology will fix it.