The short version.

• This piece now covers 24 of the 26 possible quarters between October 2019 and March 2026 — every DPIIT Quarterly FDI Newsletter's "State/District-wise FDI Equity Inflow" table (Table No.16) except two: Jan–Mar 2020 and Apr–Jun 2020, the two COVID-lockdown-era quarters, which could not be located as standalone tables. Nothing is interpolated or estimated for those two; they're reported as a genuine gap.
• Cumulative total captured: ₹25,80,189 crore (≈$321,847 million) across 3,320 (quarter, state, district) line items and 371 distinct districts.
• Four districts absorb 72.6% of everything: Navi Mumbai (₹6,92,082 cr), Bengaluru Urban (₹5,20,831 cr), Delhi (₹3,31,975 cr) and Ahmedabad (₹3,27,518 cr). Every other district in India — urban or rural, in any of 28 states and 8 union territories — splits the remaining 27.4% between them.
• Maharashtra is the single largest state (31.7% of the national total), and within Maharashtra, Navi Mumbai alone is 84.6% of the state's own total — a concentration inside a concentration. Pune (8.9%) and Thane (5.2%) round out 98.7% of the state; every other Maharashtra district, Nagpur and Aurangabad included, is individually under 0.3% of the state total.
• This dataset was supplied directly, not fetched by this session (dpiit.gov.in remains unreachable from this environment's network). It was independently spot-checked against three quarters this piece had already read directly from primary PDFs in an earlier pass — every district row, state sub-total and quarter grand-total matched exactly. It was further cross-checked against five years of separately, independently WebSearch-sourced annual and state-split figures (FY2021-22 through FY2025-26) — every single one matched to within rounding, several to the exact decimal. See §7 for the full verification record.

Why the series starts in October 2019, and where the gap is

DPIIT's district-level breakdown — "Statement on State/District-wise FDI Equity Inflow," printed as Table No.16 in most issues of the quarterly Newsletter — is not a series that runs back indefinitely. A footnote reproduced on the earliest table in this dataset states plainly: state-wise data is maintained with effect from October 2019; before that, DPIIT tracked FDI inflow by which RBI regional office processed the transaction, not by the investee company's actual state or district. That's a structural cutover in DPIIT's own data collection, not a gap in anyone's research — a district-level FDI series for India cannot be built before October 2019 from DPIIT's own sources, full stop.

Of the 26 quarters between then and March 2026, 24 were located and parsed as standalone Table No.16 documents. The two missing — January–March 2020 and April–June 2020 — are the two quarters spanning India's first COVID-19 lockdown; no standalone district-level table for either could be found. No values are estimated or interpolated for them anywhere in this piece. One useful cross-check: summing the three quarters of FY2020-21 that are present (Jul–Sep, Oct–Dec 2020, Jan–Mar 2021) gives $53.07 billion against an independently-reported full-year FY2020-21 total of $59.64 billion — leaving roughly $6.6 billion implied for the missing Apr–Jun 2020 quarter, which is plausible for a nationwide lockdown quarter rather than suspicious, but it remains an inference, not a measured figure.

The RBI-regional-office-to-state-wise cutover footnote and the two-quarter gap are as documented in the supplied consolidated dataset's own methodology report, which states the gap was reached "after an exhaustive search" of DPIIT's site. This piece did not independently re-search for those two specific quarters; see §7 for what was and wasn't independently re-verified.

The national trend, quarter by quarter

This is the actual quarter-by-quarter analysis: district-attributed FDI equity inflow, summed nationally, for each of the 24 available quarters in sequence. Two things stand out. First, the Jul–Sep 2020 spike (₹1,74,793 crore) — covered in this piece's earlier version, driven almost entirely by a single Ahmedabad district figure — remains the single largest quarter in the entire 24-quarter run, nearly matching two full quarters combined either side of it. Second, the Apr–Jun 2025 and Jul–Sep 2025 quarters are the strongest sustained pair since 2020-21, consistent with the national year-on-year growth (+18%) independently reported for FY2025-26.

Figures are each quarter's district-level rows summed, which is this dataset's own reconciliation target against DPIIT's printed quarterly "Grand Total" row (agreement reported at 0.00–0.03%, i.e. rounding-level, across all 24 quarters — see §7). QoQ change is calculated from the two adjacent available quarters; where a quarter follows the Jan–Jun 2020 gap (i.e. the Jul–Sep 2020 row), the "change" therefore spans roughly nine months, not three, and should be read accordingly.

The state dashboard: top 18 states/UTs, all 24 quarters

The 18 states and union territories below account for 99.67% of the entire national total; the remaining 19 states/UTs (mostly small north-eastern states and DPIIT's own "Others/State Not Indicated" catch-all) are omitted from this table for readability, not from the underlying data. Scroll right to see all 24 quarters for each state.

Figures transcribed from the supplied consolidated dataset's Raw Data sheet (one row per quarter/state/district, 3,320 rows total), itself parsed from DPIIT's own Table No.16 PDFs and reconciled against each quarter's own printed sub-totals and grand total. This piece independently re-summed every cell in this table directly from that raw data (not copied from any pre-aggregated summary) as part of its own verification pass — see §7.

The district dashboard: the same four districts, at 24-quarter scale

Ranking all 371 distinct districts by cumulative inflow across the full window confirms, at much larger scale, what the five-quarter version of this piece already found: a small number of specific metro districts — not entire states — are where India's FDI equity is actually recorded as landing.

"Navi Mumbai" and "Bengaluru Urban" reflect DPIIT's own district-name labelling (Bengaluru Rural and Mumbai City/Suburban are separate, far smaller line items reported on their own). These figures describe where the recipient company's registered office sits, per DPIIT's own collection method — not necessarily where the underlying investment is physically deployed, a standard caveat for this series repeated from the earlier version of this piece.

Inside the leading state: Maharashtra is itself dominated by one district

Maharashtra's 31.7% share of the national total is, on its own, unsurprising — it's consistently the top or near-top state in every independent year-by-year ranking this blog has seen. What's less obvious until you look inside the state: Maharashtra's own FDI is even more concentrated than the national picture. Navi Mumbai alone is 84.6% of Maharashtra's entire 24-quarter total; add Pune (8.9%) and Thane (5.2%) and three districts account for 98.7% of everything the state has recorded. Every other Maharashtra district — Nagpur, Nashik, Aurangabad, Kolhapur, all of them — is individually under 0.3% of the state's own total.

Top 8 Maharashtra districts by cumulative total shown; the remaining ~21 Maharashtra districts appearing in this dataset collectively make up the balance. Figures independently re-summed by this piece from the supplied dataset's raw per-quarter rows.

What moved underneath the big numbers

Below the top three, Maharashtra's own secondary industrial cities show real, if small, divergent trends across the window — comparing the first 12 available quarters against the last 12. Kolhapur essentially dried up (−98.7%), almost entirely driven by a single ₹112.5 crore quarter in Oct–Dec 2020 that never repeated. Parbhani shows the mirror pattern: its entire ₹244 crore cumulative total landed in one quarter (Apr–Jun 2025). Nashik and Satara both declined (−57.2% and −42.6%), while Aurangabad and Nagpur both rose (+86.4% and +58.0%) — a mild shift of Maharashtra's secondary-city FDI away from the Nashik/Satara belt toward Aurangabad/Nagpur, though every figure here is small relative to the Navi Mumbai/Pune/Thane core in §5.

"First 12 quarters" and "last 12 quarters" split the 24 available quarters at the midpoint (Oct 2019–Jan 2022 vs. Apr 2022–Mar 2026); because the Jan–Jun 2020 gap falls in the first half, that half technically spans a longer calendar period. At the state level, several small north-eastern states and union territories show large relative swings quarter to quarter in the underlying data (a state whose typical inflow is near zero can show a 10x move on a single mid-sized deal) — these are real but are noise around a small base rather than a trend, and are not itemised here.

How this was verified

This dataset did not arrive from a fetch this session could make — dpiit.gov.in, its mirrors, and every proxy/reader-service workaround tried remain blocked from this environment's network (documented in this piece's own prior revision). It was supplied directly, built by parsing DPIIT's own PDFs with pdftotext -layout and a custom line-reconstruction parser, per its own methodology report. Two independent layers of verification were run before using it here, neither of which simply trusted the supplied report's own claims:

1. Direct row-level match against primary PDFs this piece had already read itself. An earlier version of this piece read five DPIIT newsletter PDFs directly (Oct–Dec 2019, Jul–Sep 2020, and Oct–Dec 2022 among them, matching three of this dataset's 24 quarters). Every figure this piece checked from those three overlapping quarters — the Ahmedabad Jul–Sep 2020 row (₹1,16,379.032 cr / $15,585.19M / 66.58%), the Bengaluru Urban and Navi Mumbai Oct–Dec 2019 rows, the Delhi Oct–Dec 2022 row, the Gujarat state sub-total for Jul–Sep 2020, and all three quarters' grand totals — matched the supplied dataset exactly, to the fourth decimal place in several cases.

2. Independent cross-check against a completely separate research pass. A prior version of this piece separately compiled national and state-level FDI figures via web search (not from any PDF), covering FY2021-22 through FY2025-26. Summing this dataset's own quarterly rows into fiscal years reproduces every one of those five years almost exactly: FY2021-22 $58.77bn (search-sourced figure: $58.77bn, exact), FY2022-23 $46.03bn ($46.03bn, exact), FY2023-24 $44.42bn ($44.42bn, exact), FY2024-25 $50.02bn ($50.02bn, exact), FY2025-26 $58.85bn ($58.84–58.85bn, matches). State-level splits matched too: Karnataka's FY2021-22 share recomputes to 37.55% against an independently-sourced 37.55%; Maharashtra's FY2024-25 share recomputes to 39.16% against an independently-sourced 39.16%, with Karnataka (13.23%), Delhi (12.18%) and Gujarat (11.42%) all matching to the same two decimal places. This is two different research passes, run separately, converging on the same numbers from different methods — about as strong a corroboration as this piece can produce without personally reading all 24 underlying PDFs.

What this verification does not cover: this piece did not personally re-read all 24 source PDFs end to end, and relies on the supplied dataset's own account of its parser-validation process (its report states every quarter's district rows were checked against DPIIT's own printed sub-totals and grand totals, with 0.00–0.03% agreement, and describes specific parser failure modes it found and fixed — multi-line state names, a form-feed page-break character, a watermark that split one document's text). Those claims were not independently re-derived from the raw PDFs by this piece; they are reported as the dataset's own documented process, corroborated only indirectly via the two checks above.

What doesn't follow from any of this

The concentration finding — four districts at 72.6%, three Maharashtra districts at 98.7% of their own state — describes where DPIIT's own collection method attributes FDI equity inflow, which per its own convention is the recipient company's registered office, not necessarily where a factory, data centre, or other physical asset actually gets built. A holding company or India headquarters registered in Navi Mumbai or Bengaluru Urban can direct capital deployed anywhere in the country; this dataset cannot distinguish that from genuinely local investment, and neither could the five-quarter version of this piece. The Jan–Jun 2020 gap means no quarter-over-quarter comparison spanning it is a clean three-month step; it's closer to nine months, and is flagged as such wherever it appears in §2. The Ahmedabad Jul–Sep 2020 spike and the Parbhani/Kolhapur single-quarter swings are reported as facts from the data, not explained — nothing here identifies the specific transaction(s) behind any of them. Finally, like the five-quarter version, this remains equity inflow only, per DPIIT's own table convention; total FDI (which also includes re-invested earnings and other capital) is a larger, differently-distributed figure this piece does not attempt to break down by state or district.

Sources and caveats

The full dataset behind every table on this page — all 3,320 quarter/state/district rows, plus the state and district cumulative totals — is a single JSON file, about 1 MB, served from herrrickshaw.github.io/masaladeutsch/data/dpiit_fdi_district.json. It is plain and freely downloadable; every table in this piece was built from it directly, so it carries no rounding or curation this page's own tables don't already reflect — take it if you want to slice the data yourself (e.g. a state or district this piece didn't have room to feature).