Ask the average Indian what makes the country’s digital story remarkable, and you’ll likely hear about UPI or Aadhaar. Almost no one will mention this: the average Indian smartphone user now burns through roughly 32 GB of mobile data every month — more than users in the United States, the United Kingdom, or South Korea — while paying one of the lowest per-gigabyte prices on the planet, close to ₹9 per GB. Put another way, a resident of Zimbabwe pays close to $43.75 for the same gigabyte that costs an Indian a few cents. That is not a rounding error. It is roughly a 480x price gap for an identical unit of digital infrastructure.
This single fact — extreme consumption at near-zero marginal cost is a remarkable achievement. It didn’t happen by accident. It was engineered, over a decade, through a specific sequence of state-led infrastructure bets, private capital, and regulatory design.
As India marks its 80th Independence Day, it’s worth walking through that sequence — not as nostalgia, but as a case study every CIO and CTO with exposure to India, or ambitions of building at Indian scale, should understand.
The rails: from landline scarcity to the world’s busiest payments network
In FY 2016-17, India’s Unified Payments Interface (UPI) was a rounding error: 21 banks live, and roughly 373 transactions processed in its entire first month. A decade later, the same rail processed 24,162 crore (241.6 billion) transactions in FY 2025-26 — up 30% year-on-year — moving nearly ₹314 lakh crore (roughly $3.8 trillion) in value. Monthly volumes now regularly exceed 22-23 billion transactions, with a single-day run rate above 750 million transactions.
The scale detail that should matter most to technology leaders: India now accounts for roughly 49% of the world’s real-time payment transaction volume — more than three times Brazil’s Pix share. UPI has also started exporting itself, with live acceptance across countries including the UAE, Singapore, France, Mauritius, Bhutan, Nepal, Sri Lanka, and Qatar.
For CIOs building payments, fraud, or reconciliation systems, the operational lesson is less about the payment rail itself and more about what it proves: a thin, interoperable, API-first protocol layer, built and owned as public infrastructure rather than a single vendor’s platform, can out-scale anything a private duopoly could have built.
The identity layer: a billion-plus people, one verifiable number
Aadhaar, India’s biometric digital identity system, has enrolled close to 99% of the country’s adult population, with well over 1.3 billion numbers issued. It underpins nearly every other layer of India’s digital stack — KYC for bank accounts, direct benefit transfers, SIM issuance, welfare disbursement, and increasingly, private-sector onboarding. It is, in effect, the identity primitive that made everything downstream — UPI, DBT, e-KYC-based lending — computationally cheap to build.
For enterprises, this is the underappreciated unlock: identity verification costs that would take days and physical paperwork in most markets happen in seconds in India, at near-zero marginal cost per verification. That single design decision is a large part of why fintech, lending, and insurance products can be built and shipped in India faster than in most OECD markets.
The infrastructure build-out: connecting the last village
Digital India, launched on 1 July 2015, set out three goals: infrastructure as a utility, governance on demand, and digital empowerment. A decade on, the numbers behind that ambition:
BharatNet, the world’s largest rural broadband programme, has made more than 2.21 lakh (221,000) Gram Panchayats service-ready with optical fibre, out of a target of roughly 2.64 lakh, with an eventual reach toward 6 lakh villages.
Telephone connections rose from 93.3 crore in 2014 to over 134.8 crore by June 2026; tele-density climbed from 75.2% to 94.3%.
Internet subscriptions expanded from 25.15 crore in 2014 to over 109.2 crore by March 2026.
Wireless data cost fell from roughly ₹308 per GB in 2014 to ₹7.51 per GB in 2026 — a collapse of over 97%, driven largely by the intense competition that followed Reliance Jio’s 2016 entry.
5G services now reach 99.9% of districts across every state and Union Territory, supported by over 5.6 lakh 5G base stations.
More than 5 lakh Common Service Centres now deliver 800-plus government and business services at the village level, run by local entrepreneurs.
The pattern here is a familiar one from telecom economics — cheap unit costs unlocked by aggressive competition and shared infrastructure — but the scale at which India ran this playbook, across the most linguistically and geographically fragmented consumer base on Earth, is close to unprecedented.
The manufacturing pivot: from importer to exporter
A decade ago, only 26% of smartphones sold in India were made domestically; the rest were imported. Today, roughly 99% of smartphones sold in India are manufactured in India, and smartphones have overtaken petroleum products and gems and jewellery to become India’s single largest export category — a shift that didn’t exist in the country’s top-100 export list just ten years ago.
Behind that shift: the Production-Linked Incentive (PLI) scheme for large-scale electronics manufacturing, launched in April 2020, drew roughly ₹96,000 crore in investment, while total electronics production climbed from about ₹1.9 lakh crore in 2014-15 to nearly ₹12-13 lakh crore in 2024-25 — a roughly six-fold increase, with exports growing even faster. Newer schemes — the Electronics Components Manufacturing Scheme and PLI 2.0 for IT hardware — are now trying to replicate that success further up the value chain, in components, PCBs, laptops, and servers, where India still depends heavily on imports.
For CTOs evaluating supply-chain diversification away from a single-country manufacturing base, this is the concrete evidence that the “China+1” thesis has moved from slide-deck theory to shipped volume — even if component-level value addition (currently around 18-20%) still has a long way to climb.
The export nobody expected: India selling governance, not just software
Perhaps the least understood milestone, even among people who follow Indian tech closely, is this: India has stopped being purely a consumer of digital technology and has started exporting its Digital Public Infrastructure (DPI) as a governance model. Following its 2023 G20 presidency — where all 20 member nations formally endorsed a DPI framework built on India’s principles of open, interoperable, secure-by-design systems — India has signed cooperation agreements with roughly 23-24 countries to help them adopt components of “India Stack”: UPI-style payments, DigiLocker-style document storage, and MOSIP-based identity systems.
Concrete examples: Kenya and Cuba are implementing DigiLocker-style systems; Trinidad and Tobago became the first Caribbean nation to adopt a UPI-like rail; the MOSIP identity platform, incubated in India, is now active in roughly 19 African countries. This is soft power delivered as open-source middleware rather than as aid or arms — a genuinely novel form of technology diplomacy that most global CIOs have not yet priced into their thinking about where the next generation of “sovereign tech stacks” will come from.
What it adds up to
India’s digital economy — spanning IT services, electronics, and digital platforms — now contributes roughly 12-14% of GDP, up from about 4% in 2014, and is projected by government and RBI estimates to approach 20% of GDP by around 2030. The country now hosts more than half the world’s Global Capability Centres, employs close to 15 million people directly in the digital economy, and has built a workforce that is, per government data, nearly five times more productive than the rest of the economy.
Looking ahead: the decade that turns rails into a full stack
If the last ten years were about building the rails — payments, identity, connectivity — the next five look like they’ll be about building the stack on top of them: silicon, compute, and sovereign models, layered directly onto the DPI foundation already in place.
A few threads worth tracking now, because they’re already under construction rather than still on a slide:
Silicon leaves the drawing board. India’s first domestically fabricated chip — Tata Electronics’ Dholera facility, built with Taiwan’s PSMC on a 28nm process aimed squarely at automotive, industrial, and IoT applications — is targeting first silicon by late 2026, with a total of six semiconductor plants expected operational across 2026-27 and a full leading-edge fab envisioned by decade’s end. This won’t rival TSMC’s 2nm lines any time soon, but it plants India’s flag in the “mature node” segment that actually powers most of the world’s cars, appliances, and sensors — the unglamorous majority of global chip demand.
Compute becomes a public utility. Through the IndiaAI Mission, the government has already stood up roughly 34,000 subsidised GPUs at a fraction of hyperscaler pricing, with a stated target of 100,000 public GPUs by the end of 2026 and private buildouts pushing national capacity well beyond that. Expect “GPU-as-a-utility,” priced and provisioned the way electricity or bandwidth is today, to become as central to India’s next decade as fibre and 4G were to the last one.
Sovereign AI models move from lab to production. Homegrown foundation models — trained on Indian data, tuned for India’s 20-plus major languages, and hosted on domestic infrastructure — are already shipping (multiple labs have released open, multilingual large language models trained substantially on India-based compute). Over the next few years, expect these models to slot directly into the existing DPI stack: AI-assisted government service delivery, vernacular-language banking and healthcare interfaces, and agentic commerce layered on top of UPI and the Open Network for Digital Commerce (ONDC).
DPI exports go from pilots to platforms. The 20-plus country cooperation agreements signed so far are early-stage; the more interesting shift is India positioning itself as a neutral “digital infrastructure vendor” to the Global South, at a moment when many countries are wary of building their national data and payment rails entirely on top of US or Chinese hyperscale platforms. Watch for India Stack components — identity, payments, consent-based data sharing — becoming the default reference implementation for digital-government projects across Africa, Southeast Asia, and Latin America over the next five years, the way Linux became a default rather than an alternative.
The physical and digital layers start to merge. BharatNet’s fibre backbone, near-universal 5G, and an expanding satellite broadband push are converging with a manufacturing base that increasingly makes its own chips, sensors, and devices. The plausible endpoint by the early 2030s: a country where connectivity, compute, identity, and payments are treated as a single, low-cost, publicly governed utility layer — with private companies and global enterprises building products on top of that layer rather than having to build the layer itself, the way most of the world still has to.
None of this is guaranteed. Component-level manufacturing depth, chip design-to-fab talent pipelines, energy availability for GPU clusters, and cybersecurity at population scale are all real constraints that could slow the trajectory. But the direction of travel — from rails, to stack, to export — is now visible enough that it belongs on every global technology leader’s five-year radar, not just India’s.
India’s DPI model — thin, public, interoperable rails rather than closed platforms — is quickly becoming the reference architecture that other governments ask vendors to interoperate with, not just in India but across Africa, Southeast Asia, and Latin America.
Systems processing 750 million-plus transactions a day, on infrastructure a fraction of the cost of Western equivalents, are a genuine engineering case study in cost-to-serve at extreme scale — relevant well beyond fintech, into identity, healthcare, and logistics stacks.
The shift from 26% to 99% domestic smartphone manufacturing in a decade is evidence that India’s electronics manufacturing base is no longer a bet — it’s operating capacity, with real export volumes to underwrite it.
With India hosting the majority of the world’s GCCs and a digital workforce growing faster than the rest of the economy, the center of gravity for enterprise engineering and platform teams continues to shift there — not just cost arbitrage, but increasingly product and architecture ownership.
NPCI’s market-concentration caps, MeitY’s open API mandates, and the RBI’s account-aggregator consent framework are worked examples of regulators actively engineering competitive, interoperable markets — a governance pattern many other jurisdictions are still debating in the abstract.
India’s digital transformation is not a finished story — component-level manufacturing, rural digital literacy, and cybersecurity at this scale remain open problems. But as a demonstration of what a large, resource-constrained democracy can build in public digital infrastructure within a single decade, it is arguably the most consequential technology case study of the 2015-2025 period — and one still being underestimated outside the country.