In the next five years, we will ensure that 50 cities are covered with a strong metro network.

The 2019 Sankalp Patra promises reach and quality in one sentence: "In the next five years, we will ensure that 50 cities are covered with a strong metro network."

Reach did increase. More Indian cities have metro rail now than in 2019, and lines opened in several during the term. Counting cities is not, however, the same as testing whether the networks are strong, and only one metro system outside Delhi has been examined by the CAG in a full performance audit. That report — on the Bangalore Metro, tabled on 10 August 2026 — is what this fact-check uses, and it is worth reading as a case study of what "strong" costs.

Start with the passengers, because a metro is judged by them. India's own rule for choosing between a heavy metro and a bus rapid transit system is peak hour peak direction traffic: above 15,000 for at least five continuous kilometres, build the metro; below it, build buses. On Phase 1 in 2021, that figure "was far below the projections (below 15,000) and ranged between 6,429 and 8,852." The audit's next sentence is the finding: "Despite above, there was no study to show how ridership could be improved to justify investment in Heavy Metro."

The planning documents that would have anchored ridership to the city's actual growth did not exist when they were needed. Phase 2's project reports, written in 2011, were prepared "in absence of the Comprehensive Mobility Plan (CMP), Transit Oriented Development (TOD)/ Land Use Policy (LUP)" — plans Karnataka approved only in September 2020 and November 2022. And the returns used to justify the investment were built on the same optimistic ridership: rates of return "were calculated based on ridership figures which appeared to be overestimated", with no sensitivity analysis for lower numbers and no study of why ridership was low.

Land is where the money went. Requirements were misjudged in both phases, and "improper land estimation coupled with delays in acquisition resulted in increase in cost of land acquisition by ₹6,603.39 crore" by March 2023. Three specific failures sit inside that: because Karnataka issued no preventive orders while Phase 2 reports were being submitted, 46 landowners converted farmland to non-farmland just before or after submission and were paid accordingly, and loose valuation added ₹294.72 crore of excess compensation; delay past the 270-day possession window added ₹186.86 crore of statutory interest; and frontage premiums applied above the state's own prescribed percentages added ₹31.35 crore.

Contracting compounded it. The corporation "did not have a Procurement Manual", so estimates were prepared inconsistently — including nine Phase 2 civil contracts where ₹1,098.28 crore of taxes was left inside estimates while bidders were told to quote without tax, raising project cost by ₹1,222.40 crore. Land was handed to contractors between 12 and 1,305 days beyond the 90-day contractual limit in 13 contracts, and five of those produced ₹98.21 crore of damages claims, ₹12.19 crore already paid.

Then the finances, which is where "strong" is ultimately settled. "BMRCL has been continuously incurring cash losses for the period from 2013-14 to 2021-22", so "the revenue generated by BMRCL is insufficient to meet the External Debt Repayment obligations and BMRCL is completely dependent on GoK to service the debt raised for the project". The non-fare revenue that was supposed to close that gap did not appear: against a projected minimum of 10 per cent of fare revenue, actual non-fare revenue was 1.39 to 8.62 per cent of what had been projected. Phase 2's viability assumed ₹21,282 crore of property development income from 55 hectares — land "yet to be acquired as of March 2023" — and of 2.46 lakh square feet of commercial space already built inside stations, only 0.23 lakh square feet was in use, losing ₹38.53 crore of rent between 2019 and 2022.

Two findings concern the riders directly. Rails were bought with hardness of 260 to 330 Brinell at 34 of 36 locations against a 340-390 standard, which the audit warns raises maintenance cost, noise and vibration. And five Phase 1 stations were built beside national highways with no foot over bridge or underpass to reach them; three were added later and at two the audit records that bridges "were not found feasible due to vertical height clearances."

More cities got metros. On the one system audited in full, the network is carrying roughly half the traffic that justifies the technology chosen, was planned before the city had a mobility plan, cost ₹6,603 crore more than planned in land alone, and cannot service its own debt. Our verdict is lacks context.