
India's groundwater story got better in 2025, on average. That qualifier is the whole point. The latest Dynamic Ground Water assessment from the Central Ground Water Board puts national extraction at 60.63 per cent of what can safely be taken out each year — comfortably inside the safe band. Yet more than a quarter of the country's assessment units sit outside that band. For a bank pricing an agri portfolio, an insurer writing crop covers in a dry district, or an input company planning territory stock, the block category is the number that matters. The national average is not.
Start with the balance sheet. Total annual groundwater recharge is assessed at 448.52 billion cubic metres (BCM) for 2025. Of that, 407.75 BCM is extractable, and actual extraction across irrigation, industry and domestic use came to 247.22 BCM. Divide one by the other and you get the stage of extraction of 60.63 per cent.
The direction is genuinely encouraging. Recharge has risen from 432 BCM in 2017 to 448.52 BCM in 2025. The share of assessment units classed as safe has climbed from 62.6 per cent to 73.14 per cent over the same period, and the over-exploited share has fallen from 17.2 per cent to 10.8 per cent. Eight years of recharge structures, aquifer mapping and demand-side work show up in those three lines. Nobody should talk that down.
The assessment grades 6,762 units — mostly blocks, tehsils, taluks and mandals — by their stage of extraction. Below 70 per cent is Safe. Between 70 and 90 per cent is Semi-critical, 90 to 100 per cent is Critical, and anything above 100 per cent is Over-exploited: more water leaves the aquifer each year than returns to it.
| Assessment units by category, 2025 | Units | Share |
|---|---|---|
| Safe | 4,946 | 73.14% |
| Semi-critical | 758 | 11.21% |
| Critical | 201 | 2.97% |
| Over-exploited | 730 | 10.80% |
| Saline | 127 | 1.88% |
| Total assessed | 6,762 | 100% |
Dynamic Ground Water Resource Assessment 2025. Source: Central Ground Water Board, via the Ministry of Jal Shakti reply in the Lok Sabha (PIB). Farmioc.
Read the table from the bottom up. There are 730 over-exploited units where extraction exceeds recharge outright, and another 959 units in the semi-critical and critical bands on the way there. Add the saline units, where the water exists but is not usable, and 1,816 of 6,762 units — roughly 27 per cent — carry some form of water stress flag. A portfolio spread evenly across India would still have about a quarter of its exposure sitting in those units. Real portfolios are not spread evenly. They cluster, and clusters follow cropping intensity, which follows water — until it doesn't.
The annexure makes the point with two large agricultural states. Maharashtra looks comfortable at state level: 306 of 359 taluks are Safe, 85.24 per cent. Inside that average sit 7 over-exploited and 5 critical taluks — small counts, but a lender's exposure in those taluks does not experience the state average. It experiences the taluk.
Madhya Pradesh tells the opposite story. Only 69.72 per cent of its 317 blocks are Safe, and 64 blocks are Semi-critical — one block in five is already in the 70–90 per cent band, with 26 blocks over-exploited. Semi-critical is the category that matters most for forward-looking risk work: it is where a couple of dry years, or a shift into a thirstier crop mix, tips a block over the line. The 2025 assessment is a snapshot. Credit decisions run for seasons, not snapshots.
Government has stopped treating the average as the unit of action. Under the National Initiative on Water Security, launched in September 2025, at least 65 per cent of MGNREGS funds are to be spent on water conservation works in over-exploited and critical blocks, and 40 per cent in semi-critical blocks. The money follows the category, not the state total.
Regulation does the same. The Central Ground Water Authority's guidelines prescribe higher extraction charges and heavier environmental compensation in over-exploited, critical and semi-critical blocks, and prohibit new large-scale industries in over-exploited blocks altogether. For a processor siting a plant, or a lender financing one, the block category is now a compliance and cost variable, not background colour. Two supporting programmes matter for how fast categories can move: aquifer mapping under NAQUIM has covered about 25 lakh square kilometres, and the Atal Bhujal Yojana's community-led work in 80 water-stressed districts built over 83,000 recharge structures and brought 9 lakh hectares under efficient irrigation. Those are mechanisms that can shift a block's grade over time — which is exactly why the grade needs tracking, not a one-time check at loan origination.
One caution on timing. The 2025 assessment describes the resource position assessed for that year; it is not a live gauge. This season is a reminder of how unevenly the resource refills. The India Meteorological Department's extended range release of 5 November records post-monsoon rainfall from 1 October to 4 November at 30 per cent below the long period average nationally — and, in the same release, a single week ending 4 November that ran 96 per cent above average as two cyclones crossed the east coast. Recharge is lumpy. A block graded safe on an average year can still face a hard rabi if the rain that refills it arrives in one week, in the wrong district.
Three practical moves follow. First, tag agri exposure — loans, insured area, dealer stock, procurement catchments — by assessment-unit category, and refresh the tag when each year's assessment lands. Second, watch the semi-critical band hardest: 758 units nationally, a fifth of Madhya Pradesh's blocks, is where tomorrow's over-exploited list is currently forming. Third, treat category upgrades as credit-positive evidence only when they persist across two assessments; one wet year flatters everyone.
This is the kind of slow-moving, block-level signal that belongs inside a data platform rather than an annual PDF review — Farmioc's DataStore is built to carry exactly these layered datasets, from groundwater categories to sowing and price series, against the districts where a portfolio actually sits. The national average improved. Price the block anyway.