Extensions across 966 registrations
| Extension records counted | 852, mirror read 29 July 2026 Counted |
| Registrations carrying at least one | 526 of 966, 54.5 percent Counted |
| Median length | 6 months, quartiles 6 and 12 Counted |
| Longest single span | 120 months Counted |
| Records stating any reason | 455 of 852; the other 397 give a date and nothing else Counted |
| Of those, citing covid force majeure | all 455, in identical wording Counted |
| Why any individual date moved | Not on the register for 397 of 852 records. Not established |
The column everyone misreads
Open any UP-RERA filing and there is an extension column. Most buyers read it as a black mark, and almost nobody knows what a normal entry in it looks like, because nothing published anywhere reads the whole set at once. So here it is read at once: every 852 extension records on the Gautam Buddha Nagar register, sitting on 526 of 966 registrations, 54.5 percent of them.
An extension is a revised completion date entered on the public register, which the Act provides for. The figures below are counts over our mirror of that register, read 29 July 2026 and computed 23 September 2026. No developer is named on this page and nothing here ranks anyone.
A normal extension is six months
| Length | Records | Share |
|---|---|---|
| 6 months or less | 518 | 61.4 percent |
| 7 to 12 months | 186 | 22.0 percent |
| 13 to 24 months | 62 | 7.3 percent |
| more than 24 months | 78 | 9.2 percent |
Across the 844 records whose two dates both parse, the median is 6 months and so is the twenty-fifth percentile, which is another way of saying that six months is not the short end of the distribution. It is the distribution. The seventy-fifth percentile is 12 months.
The tail is real and it is small. 78 records, 9.2 percent, run beyond two years, and the longest single span on this register is 120 months. If you are looking at a filing whose extension runs years rather than months, you are looking at something the register treats as unusual, and that is worth a question. If it runs six months, you are looking at the most common entry on the register.
455 records state a reason. It is the same reason.
455 of the 852 records carry any stated reason at all. The remaining 397 give a revised date and nothing else, which is the first thing to know: on nearly half the extensions on this register, the public filing does not say why.
Of the 455 that do, every single one carries the same parenthetical, identical word for word with only the date in front of it differing:
Six months extension in view of force majeure caused by covid-19 pandemic
It sits on 452 distinct registrations, at most once on almost every one of them, and the recorded span is exactly six months on 452 of the 455. That is the shape of a standard allowance applied across many filings rather than 455 separately argued applications. Treat it as a regulatory allowance the sector received, not as a finding about the project in front of you.
A covid extension does not date a project to 2020
This is the part that misleads, and it misleads in a specific direction. The six months attaches to whatever completion date the filing already carried, so the revised date lands wherever that date already was. On this register the covid-worded records put the new date anywhere from 2020 to 2029.
| Year the revised date falls in | Covid-worded records |
|---|---|
| 2020 | 53 |
| 2021 | 150 |
| 2022 | 83 |
| 2023 | 86 |
| 2024 | 39 |
| 2025 | 30 |
| 2026 | 6 |
| 2027 | 5 |
| 2028 | 1 |
| 2029 | 2 |
252 of the 455 covid-worded records, the highlighted rows, put the revised completion date in 2022 or later. Seeing that wording on a filing tells you the allowance was applied to it. It does not tell you the project was behind in 2020, it does not date the project, and reading it either way will lead you to the wrong question at the site visit.
The six months is often not the end of it
| Extensions on one registration | Registrations | Share |
|---|---|---|
| 1 extension | 320 | 60.8 percent |
| 2 extensions | 124 | 23.6 percent |
| 3 extensions | 55 | 10.5 percent |
| 4 extensions | 19 | 3.6 percent |
| 5 extensions | 5 | 1.0 percent |
| 6 extensions | 3 | 0.6 percent |
320 of the 526 registrations with an extension carry exactly one. The other 206 carry two or more, up to 6. And here is the thing that changes how you read them: on 198 of those 206, every extension begins on the day the last one ended. No gap.
A chain of 3 six-month grants is not three separate six-month events. It is eighteen continuous months, entered in three rows. On a registration with more than one extension, the number to read is the date at the end of the chain, against the completion date originally proposed. That single comparison is worth more than the extension count, and both numbers are on the public filing.
What an extension does not establish
Not build quality, not financial standing, not bad faith, and not, by itself, that a project is late today. Older registrations have had more years in which to need a revised date, so a raw count of extensions tracks the age of the filing at least as much as it tracks anything about the promoter. That is why this page counts extensions and ranks nobody, and why a league table of promoters by extension count would be a worse document than this one.
What the column is genuinely good for is a short list of questions you can answer from the public filing before you pay anyone anything: how many extensions, how long the chain runs in total, what date sits at the end of it, and how far that date is from the completion period originally proposed. Our read of completion dates across 957 filings covers that last comparison across the register, what a normal filing looks like covers every other field, and who registers projects here covers the entity on the other side of the agreement.
