Ninety-Seven Per Cent Built, One-Third Working

Blitz Bureau

NEW DELHI:India laid optical fibre to almost every gram panchayat it set out to reach. An audit tabled in Parliament this month establishes what happened next, and the gap between the two numbers is the most useful thing written about rural broadband this year.

The document is Performance Audit — Civil Report No. 19 of 2026 of the Comptroller and Auditor General of India, “Performance Audit on BharatNet”, covering the Department of Telecommunications in the Ministry of Communications. It was tabled in Parliament on 12 August 2026. It examines the period 2017-18 to 2022-23 with its focus on Phase-II, records instances found during a test audit conducted between October 2023 and August 2024, and updates the project’s status to November 2025 where that was relevant.

BharatNet was conceived in October 2011 and executed through a special purpose vehicle, Bharat Broadband Network Limited, incorporated in 2012, with funding from the Universal Service Obligation Fund, now the Digital Bharat Nidhi. The design intention was straightforward and large: broadband to roughly 2.5 lakh gram panchayats. The CAG describes it, without qualification, as “one of the world’s largest telecommunications initiatives”.

What was built

The construction record is genuinely substantial and the audit says so first. As at November 2025, expenditure of ₹39,888 crore had been incurred against an approved budget of ₹42,068 crore — ₹11,148 crore in Phase-I and ₹30,920 crore in Phase-II. Blitz India recomputes that as 94.82 per cent of the approved budget spent. Against that, BharatNet had achieved service-ready status in 2.19 lakh gram panchayats — 96.70 per cent of a revised scope of 2.26 lakh. (Paras 1.1, 1.5, 3.3.2)

Very few infrastructure programmes anywhere deliver ninety-seven per cent of physical scope for ninety-five per cent of sanctioned money. That sentence is the honest opening, and the audit’s own conclusion begins in the same place: BharatNet “has achieved substantial infrastructural success”.

BharatNet, as the auditor found it

Source: CAG Report No. 19 of 2026, paragraphs as cited. Percentage of budget spent recomputed by Blitz India. Photograph note: an audit report is a document and not an event; no copyright-clean photograph of it exists, and a stock picture of fibre-optic cable would be a symbolic image of the kind Circular BIMG/CIR/2026/02 forbids. This in-house data card runs in its place.

Where the gap opened
The audit’s central finding is not about construction. It is that infrastructure “could not be translated into efficient operation, utilisation and revenue generation”. As at November 2025, only 33.20 per cent of service-ready gram panchayats were operational. (Paras 1.5, 3.2.5, 4.2)

The auditor is specific about why, and the reason is unromantic. The main cause of downtime was fibre faults and cuts, which accounted for approximately forty-eight per cent of non-operational cases. The mean time to restore a fibre fault averaged seventeen days. An inability to integrate fibre-fault localisation systems hampered maintenance further — which is to say that when a cable broke, nobody could quickly establish where.

Seventeen days is the number to hold on to. A village school, a primary health centre or a common service centre that loses connectivity for a fortnight at a time has not been connected in any sense a citizen would recognise, whatever the asset register says. And the audit found only 17.91 per cent of available bandwidth in use, with major telecom service providers having withdrawn over service-quality concerns, and dark fibre left unleased in many cases despite demand, for want of a policy or a decision at the level of the state implementing agency. (Para 4.3)
Planning carried its own weight. Phase-II was designed for twenty states, twenty-one circles, across three models — a central public sector undertaking-led model, a state-led model and a private-led model — but was implemented in thirteen states and fourteen circles. Detailed Project Reports, which begin the work, were prepared from desktop surveys where physical surveys were mandated; in Punjab and Bihar under the private-led model, BBNL implemented without preparing DPRs at all. In the satellite component intended for remote areas, Wi-Fi access points were not installed even at service-ready panchayats, so end-users could not reach the service that had been declared available. Six states and seven circles reached full service-ready status, with delays of two to six years across every model. (Paras 3.1, 3.2)

The public sector strand

Because BharatNet was executed through a special purpose vehicle and two central public sector undertakings, the audit carries findings that belong squarely to the public-enterprise record.

The milestone-based payment model was weakest under the CPSU-led model, which the audit associates with lower service delivery, while the private-led model demonstrated stronger financial discipline and better service delivery. The auditor also records that Bharat Sanchar Nigam Limited, under the CPSU-led model, diverted ₹2,001.43 crore of project funds to other purposes — and states in the same breath that the amount has since been recouped, and that submission of utilisation certificates is now required under the Amended BharatNet Programme. (Paras 3.3, 4.4)

The project was designed to be self-sustaining. High maintenance costs set against slight revenue ended that expectation, and the audit says so.

A cable in the ground is an asset. A cable that carries traffic is a service. The audit is about the seventeen days between them.

The Remedy, in the Auditor’s Own Words

• BBNL has been merged with BSNL; BharatNet now runs on a single-agency model under BSNL
• This is delivered through the Amended BharatNet Programme, approved by the Cabinet in August 2023
• The CAG records that the merger decision “reflects an acknowledgement of the governance and execution deficiencies noted”
• ABP expands scope from 2.5 lakh gram panchayats to 6.4 lakh villages and gram panchayats
• The diverted ₹2,001.43 crore has been recouped; utilisation certificates are now required
• Audit recommends its reforms “may be considered during implementation” of ABP

Where the file stands, stated exactly
Two things must be said about status, because an audit finding without its status is half a fact. The remedial position set out above is carried in the auditor’s own text: the merger of BBNL into BSNL, the single-agency model and the Amended BharatNet Programme are recorded by the CAG as the response already in motion, and the recoupment of ₹2,001.43 crore is recorded as accomplished rather than promised.

What this desk cannot establish from the primary document is the parliamentary stage. The press brief issued with the report does not record a separate reply from the Department of Telecommunications, and Blitz India has not been able to establish from the tabled document whether an Action Taken Note has been filed or whether the Public Accounts Committee has taken the report up. That gap is stated rather than papered over, and it will be closed when the departmental reply and any Committee proceedings are available.

What would close the distance
The audit itself frames the task correctly: the reforms it recommends “may be considered during implementation” of a programme that will take BharatNet from 2.5 lakh gram panchayats to 6.4 lakh villages. That is a two-and-a-half-fold expansion of a network whose existing third is working. The order of operations matters more than the ambition.

Three steps would show up in a citizen’s experience quickly. The first is the seventeen-day restoration time, which is an operations problem with an operations answer — fault localisation integrated into the network so that a break is located in hours, and maintenance contracts written against restoration time rather than against uptime averages. The second is the dark-fibre decision: capacity is lying unleased against real demand for want of a policy at state implementing agency level, and a model leasing framework issued centrally would unblock it without a rupee of new capital. The third is publication. Gujarat and Punjab have taken operational gram panchayats above ninety per cent while the national figure sits at a third; whatever those two states are doing differently is knowledge the other eleven are entitled to. A monthly, state-wise operational-panchayat figure on the Department’s own dashboard would put it there.

The asset is built and largely paid for. What remains is the cheaper half of the job.