Digital Infrastructure Economy: A Data-Driven Assessment

Every month, a press note from the National Payments Corporation of India lands with a number that would have looked like a typo a decade ago. In July 2026, the Unified Payments Interface handled 23.66 billion transactions worth Rs 29.88 lakh crore — a record, up 22 per cent by volume on the same month a year earlier, and achieved in a month with no festival, no sale season, nothing seasonal to inflate it. That is roughly 763 million payments a day, or about 8,800 every second.

Numbers like that do a lot of rhetorical work. They get quoted at summits, in investor decks, and in the occasional ministerial speech about leapfrogging. What they do not do, on their own, is tell you whether the digital infrastructure economy is a good investment. For that you need the boring documents: demand-for-grants tables, parliamentary standing committee reports, TRAI's quarterly performance indicators, and the odd written reply in the Rajya Sabha. Read together, they tell a more interesting story than the headline — one in which the plumbing is genuinely world-class and the economics are still being argued over.

The base is real

Start with what is not in dispute. TRAI's performance indicator report for January–March 2026 counts 1,092.79 million internet subscribers, up 6.24 per cent in a single quarter. Broadband alone accounts for 1,065.88 million of them. Average data consumption runs at 26.7 GB per subscriber per month, ahead of most European markets, while average revenue per user sits at Rs 196.04 — about $2.25. India is, on the evidence, the cheapest place on earth to be heavily online.

The soft layer built on that base is the part the rest of the world studies. Beyond UPI, the same NPCI data shows Aadhaar-enabled payments at 100 million transactions in July and FASTag at 347 million. This is public infrastructure running at private-sector velocity.

There is a hard-currency dimension too, and it is easy to overlook because it sits in the balance of payments rather than in any digital mission document. India's services exports reached $421.3 billion in FY26, according to Commerce Ministry data drawing on RBI figures, of which telecommunications, computer and information services accounted for $206.6 billion — very nearly half. That single line item does more to steady the current account than most of what gets discussed under the heading of digital policy, and it rests on precisely the connectivity, power and data centre capacity being debated here.

Then there is measurement. MeitY's own estimate, released in January 2025 using OECD and Asian Development Bank methodologies, put the digital economy at 11.74 per cent of national income in 2022-23 — Rs 31.64 lakh crore, roughly $402 billion, employing 14.67 million people. The ministry projects close to a fifth of national income by 2030, growing at almost twice the pace of the wider economy.

Note the date. The most authoritative official estimate of India's digital economy is three years stale, and MeitY itself calls it conservative, since it cannot capture smaller platforms, informal digitalisation, or fast-moving sectors such as health and logistics. We are steering a fairly large ship using a chart drawn some distance back.

“India's digital problem is no longer scarcity. It is that we keep counting what we have built, and rarely what anyone is actually using.”

The hard layer arrives, hungry

The physical build-out is where the past eighteen months have been extraordinary. Estimates of installed data centre capacity differ by source — JLL counted 1,123 MW in the first half of 2025, SBI's April 2026 sector report puts it near 1.5 GW, CBRE reports over 1,700 MW at the close of 2025 — and that spread is itself worth noticing. Analysts cannot agree within 50 per cent on the size of the sector attracting some of the country's largest single foreign investments.

Those investments are real enough. Google's AI hub at Visakhapatnam, confirmed in October 2025, commits $15 billion over 2026-2030 to gigawatt-scale compute, new subsea cable landings, transmission lines and clean energy systems. SBI's report prices a standard Tier III facility at Rs 50-55 crore per MW, rising to about Rs 75 crore per MW for AI-optimised builds using immersion cooling. At those rates, a single gigawatt of AI capacity is a Rs 75,000 crore proposition before anyone switches on the lights.

The electricity is the constraint. In a written reply on 29 July 2026, the Ministry of Power told Parliament that AI data centres could add 26.3 GW to national demand by FY2031-32 — nearly double the government's own March 2026 estimate of 13.56 GW, revised upward within five months. Against a projected peak demand near 388 GW, that is close to 7 per cent of the grid handed over to compute.

Capital intensity is the awkward part. Data centres are among the most expensive assets per job created anywhere in the economy. They generate construction work, property tax, grid demand and a certain amount of civic pride; they do not generate factory floors. Whether the trade is worth it depends entirely on what gets built on top of the compute — and that is a bet, not a projection.

The fibre nobody is using

Which brings us to BharatNet, the cautionary tale in this file. More than Rs 40,635 crore has been committed from the Digital Bharat Nidhi across two phases. On 1 April 2026, the government told the Rajya Sabha that over 2.18 lakh of 2.69 lakh gram panchayats had been made “service-ready”.

That phrase is doing heavy lifting. The Parliamentary Standing Committee on Communications and IT found 78,899 gram panchayats with operational points of presence, and 9,48,237 active rural fibre-to-the-home connections as of 28 February 2026 — fewer than a million households sitting on 7.19 lakh kilometres of laid fibre. The committee's recommendation was refreshingly blunt: measure how effectively the fibre is being used, not how many kilometres of it exist.

Capacity built versus capacity consumed is the single most useful lens for assessing this entire sector.

Who pays for free?

UPI's economics have moved from footnote to open policy question. The RBI estimated in 2022 that processing a person-to-merchant UPI transaction of average value Rs 800 cost roughly Rs 2. A parliamentary committee has cited an industry estimate of Rs 20,700 crore as the annual operating cost of the UPI ecosystem. The FY27 Budget allocates Rs 2,000 crore in incentives — around 11 per cent of that.

The Department of Financial Services has provided Rs 8,276 crore in cumulative support between FY22 and FY25, and the incentive has already been narrowed to transactions up to Rs 2,000 at small merchants, at a rate of 0.15 per cent. In August 2026, Parliament passed the Taxation and Other Laws (Amendment) Bill, which lets the government notify which electronic payment modes enjoy statutory protection from charges. Officials have been clear that UPI stays free for citizens. The working assumption across the industry is that large merchants will eventually pay a merchant discount rate.

This is the point at which a subsidised public rail either finds a revenue model or quietly starts eroding the balance sheets of the banks and fintechs that operate it.

What the Budget is signalling

Fiscal allocations are the least sentimental data available on any sector. MeitY's 2026-27 allocation is Rs 21,633 crore, down 17 per cent from Rs 26,026 crore. The IndiaAI Mission gets Rs 1,000 crore, halved from Rs 2,000 crore, after utilisation that PRS Legislative Research notes has remained below 50 per cent since launch. Semicon India takes Rs 8,000 crore — 37 per cent of the ministry's budget — having used 9 per cent of its allocation in 2024-25 and an estimated 61 per cent in 2025-26. Production-linked incentives for electronics were cut by 78 per cent.

The honest reading is not that ambition has faded. It is that money was allocated faster than institutions could spend it. National compute capacity crossed 38,000 GPUs by December 2025, with a further 20,000 announced, so the mission is delivering something. It is simply not absorbing capital at the pace the headline allocations implied.

The assessment

Three findings survive the data.

First, India's digital infrastructure works at a scale nobody else has attempted at this income level, and the usage statistics are not inflated. A country averaging 26.7 GB a month per subscriber is not performing digitisation for the cameras.

Second, the binding constraint has shifted from a shortage of infrastructure to a shortage of utilisation. Fibre without subscribers, GPUs without absorbed budgets, and payment rails without a revenue model are three versions of the same problem.

Third, the sector is outrunning its own measurement. When the official GDP estimate is three years old, capacity estimates vary by half, and a central power forecast doubles inside five months, forecasting confidence ought to be priced accordingly.

None of this argues against the build-out. It argues for changing the scoreboard — from what has been laid, allocated and announced, to what is being used, spent and earned.

Sources

1.     NPCI — UPI Product Statistics (monthly volumes and values) — https://www.npci.org.in/what-we-do/upi/product-statistics

2.     Business Standard — UPI clocks record 23.66 billion transactions in July 2026 (NPCI data, IMPS, FASTag, AePS) — https://www.business-standard.com/finance/news/upi-transactions-hit-record-23-66-billion-in-july-value-rs-29-88-trillion-126080100548_1.html

3.     PIB / TRAI — Indian Telecom Services Performance Indicator Report, January–March 2026 — https://www.pib.gov.in/PressReleasePage.aspx?PRID=2276780&reg=3&lang=1

4.     TRAI — Reports and Performance Indicator Reports — https://www.trai.gov.in/release-publication/reports

5.     PIB / MeitY — Estimation and Measurement of India’s Digital Economy (released January 2025) — https://www.pib.gov.in/PressReleasePage.aspx?PRID=2095260

6.     PIB / Ministry of Commerce & Industry — India’s Services Exports Rise to USD 421.3 Billion in FY 2025-26 (24 July 2026) — https://www.pib.gov.in/PressReleasePage.aspx?PRID=2288851&reg=48&lang=2

7.     State Bank of India — Data Centres in India, Sector Report, April 2026 — https://sbi.bank.in/documents/57189647/58231607/Sector+Report+DataCentre.pdf

8.     Business Today — India data centre capacity to jump 30% in 2026 (CBRE report, April 2026) — https://www.businesstoday.in/technology/story/india-data-centre-capacity-to-jump-30-in-2026-500-mw-supply-boost-expected-cbre-report-523529-2026-04-01

9.     PIB / MeitY — Google AI Hub at Visakhapatnam, $15 billion investment (14 October 2025) — https://www.pib.gov.in/PressReleasePage.aspx?PRID=2179035

10. Google — Our First AI Hub in India, Powered by a $15 Billion Investment — https://blog.google/intl/en-in/company-news/our-first-ai-hub-in-india-powered-by-a-15-billion-investment/

11. Business Standard — AI data centres to add 26.3 GW to India’s power demand by FY32 (Ministry of Power reply, 29 July 2026) — https://www.business-standard.com/industry/news/ai-data-centres-add-26-3-gw-india-power-demand-fy32-govt-126072900236_1.html

12. PIB / Ministry of Communications — Internet Connections in Panchayats, Rajya Sabha reply (1 April 2026) — https://www.pib.gov.in/PressReleasePage.aspx?PRID=2247709&reg=3&lang=2

13. Universal Service Obligation Fund / Digital Bharat Nidhi — BharatNet Project — https://usof.gov.in/en/bharatnet-project

14. Open — Parliamentary Standing Committee findings on BharatNet utilisation — https://openthemagazine.com/india/bharatnet-has-built-the-highway-but-is-rural-india-actually-online

15. Business Standard — Free for users, costly to run: the economics behind India’s UPI system (17 August 2026) — https://www.business-standard.com/economy/news/free-for-users-costly-to-run-the-economics-behind-india-s-upi-system-126081700067_1.html

16. Business Standard — Govt clarifies UPI to remain free; MDR may apply to large merchants later (8 August 2026) — https://www.business-standard.com/finance/news/upi-to-remain-free-for-users-mdr-may-apply-to-large-merchants-later-126080801212_1.html

17. PRS Legislative Research — Demand for Grants 2026-27 Analysis: Ministry of Electronics and Information Technology — https://prsindia.org/files/budget/budget_parliament/2026/DFG_Analysis_2026-27_MeITY.pdf

18. PRS Legislative Research — Union Budget 2026-27 Analysis — https://prsindia.org/files/budget/budget_parliament/2026/Union_Budget_Analysis-2026-27.pdf