Methodology, and what is missing

The unit of record on this site is a measured quantity with a date. Not a project, not a milestone, not a status. A row exists here only if an official document prints a number, a unit and a date that the number refers to. If a document says work is progressing, that sentence is not a row, and this register will not convert it into one.

That rule is why some obvious questions have no answer here. How many kilometres of road have been laid in the Activation Area is the most common one, and the honest answer is that none of these documents publishes a length of road. The gap list below says so in the source's own words rather than filling it with an estimate.

Two dataset files stand behind the site. The monthly series was read from the Dholera section of eight Delivery Monitoring Unit proformas plus one parliamentary answer, giving 115 records. The company dataset was read from two audited annual reports, giving 225 records.

Every text_layer record's verbatim string was asserted to occur on the cited PDF page. Every vector_render subtotal was re-summed from its own components, and the FY2024-25 headline P&L was matched against the text-layer summary on p19. 157 records were checked that way with 0 failures, and 12 arithmetic cross checks were run with 0 failures.

The check column

• checked: the figure and its date are carried exactly as printed, and the page is cited.
• transcribed from a page render, because that table has no text layer, then arithmetically cross checked.
• a target stated in the source, which is not a reading.

The mark travels with the figure everywhere it appears, including in the downloadable files, so a reader can see at a glance whether a number came from a text layer or from a transcription of an image.

No figure on this site is tagged by us. The tag is carried in the dataset against the record, and it reflects what the source document itself is doing: stating an outcome or stating an intention.

Every quantity in the monthly series has its own page, because a monthly series is the thing this register exists to publish. An annual report line gets its own page when it is a statement total or a physical quantity. The remaining line items are published in full on their statement sheet, with the same dates, pages and extraction marks, because a single line lifted out of a statutory account is misleading without the statement around it. That gives 26 series pages and no line item that is hidden. The index lists all of them and says where each one lives.

These 12 limitations are declared in the dataset and reproduced here without edit.

• ROAD KILOMETRES: no km-of-road figure for Dholera in any of the eight DMU reports or the Lok Sabha reply. The reports describe Dholera trunk infrastructure only as 'nearing completion' (30.06.2025) then 'complete' (31.08.2025 onward). No length was published, so none is emitted.
• WATER AND SEWER: no km of water main, sewer line or drainage commissioned, and no MLD capacity, anywhere in the Dholera section of any report. Not emitted.
• HECTARES SERVICED: the reports state area PLANNED (920 sq. km), area CARVED OUT as Phase I Activation Area (22.5 sq. km) and area TRANSFERRED to DICDL (48.31 sq. km). None of these is 'hectares serviced'. No serviced-area figure exists in the sources.
• PERCENTAGE COMPLETION OF A NAMED PACKAGE: the DMU reports do carry 'Cumulative Physical Progress of X%' lines, but only for other nodes (for example IMC Khurpia and IMC Rajpura). The Dholera section carries no percentage in any of the eight editions. Not emitted, and must not be borrowed from another node.
• EXPENDITURE RELEASED / SPENT: no expenditure-to-date figure for Dholera. Rs. 2,784.83 Crore appears twice, as (a) the value of APPROVED tender packages and (b) matching EQUITY RELEASED against land transfer. Neither is money spent on construction. Recorded as two distinct metrics with that distinction stated in each record's note.
• MW ENERGISED beyond solar: the only energised-capacity figure is 300 MW of the 1,000 MW Solar Park (Tata Solar Power Ltd.), and it is identical in all eight reports from 30.06.2025 to 30.06.2026. No substation, transmission or grid MW figure is published.
• AIRPORT, EXPRESSWAY AND RAIL PROGRESS: the Dholera International Airport is described only as 'construction works are underway'; the Ahmedabad-Dholera Expressway only as a 6-lane greenfield expressway 'being implemented by NHAI'; the Bhimnath-Dholera rail line only as 'approved' on 21 September 2021. No km, no percentage, no cost, no date of completion. Only the 6-lane design figure is emitted, tagged as a specification rather than progress.
• MISSING REPORT MONTHS: the eight editions available are 30.06.2025, 31.08.2025, 30.09.2025, 31.10.2025, 30.11.2025, 31.01.2026, 31.05.2026 and 30.06.2026. July 2025, December 2025, February 2026, March 2026 and April 2026 are absent from the supplied artifacts, so movement inside those windows cannot be dated more finely than the surrounding editions. Not interpolated.
• TATA SEMICONDUCTOR FAB: not mentioned by name in the Dholera section of any of the eight DMU reports. The only named investor is TATA Chemicals, described as the anchor investor. No fab investment figure, area or milestone is available from these sources.
• NO DHOLERA-SPECIFIC INVESTMENT FIGURE: the Rs. 1,52,757 crore / US $18 billion figure in the Lok Sabha reply is the combined investment POTENTIAL of all 20 NICDP projects. It is emitted with tag TARGET and a scope field saying it is not Dholera-specific. There is no Dholera-only investment number in these sources.
• PLOT ALLOTMENT DETAIL: the reports give a count, a total acreage and an industrial acreage, but never the allottee names (beyond the anchor), plot numbers, allotment dates or rates. Those fields are null by absence.
• TWO FILES ARE DUPLICATES BY NAME ONLY: 'dmu.pdf' is the edition as-on 31.01.2026 and 'dmu-latest.pdf' is the edition as-on 30.06.2026. Neither filename carries its date, so as_of_date was read from inside each PDF, not from the filename.

These 11 limitations apply to the audited source.

• Per-allottee land area is NOT disclosed in either report. Both give only a cumulative total (530 acres at 2025-03-31, 522 acres at 2024-03-31) against a named but explicitly non-exhaustive list of allottees. No acreage or price can be attributed to Tata Semiconductor, Renew Photovoltaics, Polycab, Inox Air Products or Tata Chemicals.
• Per-allottee consideration is NOT disclosed, with one exception: the FY2023-24 report discloses a 13,646.74 lakh GSEM land-allotment subsidy receivable relating to the allotment made to Semifab Private Limited (AR_FY2023-24.pdf p112). No other allottee has a monetary figure attached.
• Land price per acre / per sq m is NOT disclosed in either report. The FY2023-24 report states only that land was allotted to anchor tenants 'at discounted rate' (AR_FY2023-24.pdf p117) and does not quantify the discount or the rate.
• Land monetised DURING each financial year (as an area figure) is NOT stated. Both reports give cumulative allotted acreage only. The 530 minus 522 difference is not presented by the source as a year's allotment and has not been computed here.
• Project-wise physical progress percentages for DICDL's own eight social infrastructure projects are NOT disclosed. The FY2024-25 report gives contractor, award date, cost, areas and construction start date, and describes each as 'currently underway', but publishes no percentage completion. Only the Ahmedabad-Dholera Expressway (an NHAI project, not DICDL's) carries a percentage figure.
• Dholera International Airport physical progress percentage is NOT disclosed in either report; only target dates are given.
• Total project cost of the Tent City is NOT disclosed. The report gives only the 1.55 crore management fee receivable by DICDL under the PPP arrangement.
• The FY2022-23 figures in this dataset are taken from the comparative columns of the FY2023-24 annual report, not from the FY2022-23 annual report itself, which was not supplied.
• The FY2024-25 cash flow statement (dicdl_ar25.pdf p69) exists in the source but has not been transcribed into this dataset; it was outside the requested field list.
• Segment-level or project-level revenue attribution is NOT available. Note 33 states the company reports a single segment, 'Development of industrial and residential township'.
• DICDL employee headcount is NOT disclosed in either report.

These 9 caveats are places where the source documents are internally inconsistent, or where an extraction decision had to be made and recorded.

• All FY2024-25 financial-statement tables in dicdl_ar25.pdf are vector-outlined artwork with no extractable text layer, so those figures were transcribed from 190-200 dpi renders of the cited pages. Each transcription was then validated two ways: every subtotal was re-summed from its own components, and the headline P&L figures were matched character-for-character against the Board's Report summary table on p19, which does have a text layer. All checks passed. Records carry extraction='vector_render' so this provenance is visible downstream.
• The two reports disagree on the Ahmedabad-Dholera Expressway lane count: the FY2023-24 report calls it a 6-lane Greenfield Expressway (p19); the FY2024-25 report calls it '110 km, 4-lane' (p21). Both statements are recorded as written. Neither has been corrected or reconciled.
• The expressway completion target moved between reports: January 2025 (FY2023-24 report) to end of 2025 (FY2024-25 report). Both are stored, each tagged TARGET against its own source year.
• The FY2024-25 report restates two FY2023-24 comparatives under its Note 43 reclassification: employee benefit expenses 465.41 becomes 466.34 and other expenses 1,121.46 becomes 1,120.53. The dataset keeps the original figure from each report rather than picking a winner.
• Source-document inconsistency, left uncorrected: Note 44 (dicdl_ar25.pdf p111) lists land development cost sub-lines for FY2023-24 that differ from Note 10 (p86) for the same items (Roads & Services 1,88,346.61 vs 1,58,497.79), and its own FY2023-24 column does not sum to the total it prints. The dataset uses the Note 10 figures, which tie to the balance sheet and sum correctly. The variance is flagged, not silently resolved.
• The FY2024-25 board composition table (p30) prints names, designations and DINs as three parallel columns, and DIN 09403694 appears against Shri Ashokkumar Khandubhai Patel elsewhere in the same report. Director records for FY2024-25 carry a note that the DIN-to-name mapping on that table is not fully reliable. Names and designations are as printed.
• Rupee figures are in lakh unless the unit field says otherwise; the reports print land and infrastructure headline values in crore, and those are stored in crore as printed rather than converted. Paid-up share capital is the one converted value (Rs. 56,83,32,65,310 as printed, stored as 5683.326531 crore) and the original is kept in the note.
• The FY2024-25 named-allottee list drops M/s TATA Chemicals Limited, which the FY2023-24 report named. The source gives no explanation. Both lists are stored against their own year; no inference has been drawn about whether that allotment lapsed.
• The eight social infrastructure projects include award and construction-start dates falling after 2025-03-31 (up to 2025-06-13), reported as post-balance-sheet activity in the FY2024-25 Board's Report. They are tagged OFFICIAL because they are statements of fact in an audited annual report, not forward targets.

Write to corrections@dholera.today with the document, the page and the figure. A correction that cites a page will be published with the date it was made. A correction that cites a news report, a brochure or a listing will not, because those are not sources here.

The whole register is downloadable as JSON and CSV from the data page under an open licence, so the extraction can be checked against the original documents without asking us for anything.