Pune micro-markets: launches up, unsold up, and how both are true

Pune's H1 2026 numbers argue with each other in public: launches up 17 percent, unsold stock up 19 percent to 57,879 units, premium sales up 54. Commentators pick one number and write the city's obituary or its victory lap. Developers do not get to pick: the contradiction is the operating environment, and the register is where it resolves into decisions. This is the supply-side read.

Resolving the contradiction#

A city where supply grows 17 percent while unsold grows 19 is a city absorbing a lot and launching more, momentum and indigestion at once. The premium barbell explains part: developers chased the Rs 2 to 5 crore band's 54 percent surge, synchronising launches into it, while mid-market corridors quietly accumulated. Geography explains the rest, and it is the part the register can actually settle: the micro-market method run across Pune's catchments finds lanes inside one district that are nothing like each other, the buyer-side version of which we wrote separately. The city series is weather; the field-level truth is filed, quarterly, at the register.

What the dispersion actually looks like#

Pune district carries 13,739 published registrations, of which 4,872 currently read Active. Take the 4,120 of those that file the building-level unit table, and the district as a whole shows 221,148 unsold units against 503,974 sanctioned: 43.9 percent unsold.

Now cut it by pincode. Forty-eight Pune pincodes carry 20 or more such registrations, which is enough to be a market rather than a rounding error, and they do not resemble the district average or each other.

The range is 32.1 percent to 69.1 percent, and the median is 47.5. A quarter of these lanes sit below 40.9 percent unsold and a quarter above 54.7. Wagholi and Fursungi are in the same district, under the same regulator, inside the same H1 2026 headline, and one carries roughly twice the unsold share of the other.

The obvious objection is age: a pincode full of registrations from last year should look heavy simply because nothing has had time to sell. It does not hold here. The share of each pincode's active register that was registered in 2024 or later runs from 21 percent to 85 percent, and its correlation with the unsold share is 0.07, which is no relationship at all. Kothrud is 71 percent newly registered and sits at 36.3 percent unsold; Fursungi is 80 percent newly registered and sits at 69.1. Whatever separates these lanes, it is not vintage, which is precisely why a supply decision cannot be made from the district number.

An analogy: the monsoon average#

A state can record a normal monsoon while one district floods and another rations water: farmers plant by their district's gauge, not the state's average, because crops grow in fields. Pune's 57,879 unsold units are a state-level rainfall figure. Your launch grows in a 2 to 3 kilometre field with its own gauge, the catchment's filed supply and unsold stock, and planting decisions made on the state average are how crowded corridors got crowded.

What the register will not tell you, and what people assume it does#

One limit belongs here rather than in a footnote, because planning around it after the land is bought is expensive. MahaRERA files no unit configuration. The building-level table separates residential units from non-residential ones and stops there: there is no BHK field, no carpet-area band per flat, no unit type, and no price. Nobody has a filed absorption series by configuration for any Indian market, ours included, and a consultant who offers one is modelling it from transaction data and listings rather than reading it off the register.

What the register does carry at lane scale is the thing the city series averages away: how many projects are there, who registered them, when, how many units they sanctioned, how many remain unsold, how far construction has moved, and which registrations have gone quiet or died. That is a supply and competitor picture, not a product-mix one, and it is enough to decide whether a parcel should carry another 300 units. The product mix decision needs the register plus the transaction data, and the honest version of this method says which half is which.

The supply-side playbook for this cycle#

1. Run the field gauge before the land cheque. Catchment supply table, unsold share against the district spread above, construction distribution, distress share: the launch proceeds where filed supply leaves room for the added units, whatever the city mood. The gauge has a forward half as well as a standing one, since every active registration files a promised completion date: stacking the catchment by that date says which quarters your sales window meets three rival completions and which meet none.
2. Count the recent arrivals, not just the standing stock. Moshi's 412105 holds 332 active registrations from 263 distinct promoters, 203 of them registered since the start of 2024. A lane absorbing that much new supply is a different bet from one where the register has barely moved, and density-versus-velocity data hands that gap to whoever checks.
3. Read competitor stress as strategy. Rising inventory strains weak balance sheets first: silent filers, slowing certificates and extension applications in your catchment are pricing intelligence, and sometimes land-acquisition pipelines, visible quarters early.
4. Protect your own record through the cycle. Crowded markets make buyers pickier readers of filings, the record MahaRERA's shelved grading matrix was designed to score; the promoters who keep filing clean through a slow patch convert the next upcycle's first demand.

Anita reads two lanes that share a postcode's reputation#

Anita Rao's Pune decision this cycle (illustrative, as ever): her firm held an option on a parcel in a corridor where the register showed four competitors' registrations arriving inside six quarters and the pincode's unsold share sitting above 60 percent, the heavy end of the district spread. The broker's pitch was the city's premium number. Her field gauge was the lane's own filings, and it said the added supply would land into stock that had already stopped clearing.

She walked, and took a parcel three kilometres away in a pincode reading 34 percent unsold with a comparable count of active registrations behind it. Note what she did not claim to know: nothing in the filings told her which configuration to build, because nothing in the filings carries one. It told her how much competing stock was standing and how much had arrived recently. The product decision came from her own sales data. The city data said caution; the field data said plant, and said nothing whatever about what to plant.

Where to run this free, and what a report adds#

The supply half of this is public and assemblable by hand. The free Pune district page lists every registered project in the district with its status and promised dates and breaks the district down by taluka, and each project page carries the unit table, the construction progress, the extensions and the disclosed disputes. For a catchment of a dozen projects that is an afternoon, and it is the right way to learn what the filings do and do not say before paying anyone for a summary of them.

What costs money is repetition, memory and the joins. The area market report is a flat Rs 2,999 per pincode and assembles this method's output for a Maharashtra pincode: the supply table, the unsold and construction distributions, the distress markers and the competitor set, positioned against the district cohort rather than against a national average. The Rs 2,999 developer project analysis does the same work around a specific parcel, with the two kilometre and five kilometre rings drawn from the filed coordinates.

The one-line summary#

Launches up, unsold up, premium surging: all true, none decisive. Pune's lanes run from 32 to 69 percent unsold and vintage does not explain the gap, so plant by the field gauge, count who registered nearby last year, read competitor stress as intelligence, and remember that the register answers how much is standing and never what to build.

Methodology and sources#

• City-level launches, unsold inventory and the premium band: Knight Frank India, H1 2026 Pune residential market update.
• Pincode unsold shares, the district spread, promoter counts and registration years: ReraGenie analysis of 55,913 published MahaRERA projects, as updated on 21 September 2026, restricted to Pune district registrations currently reading Active. Unsold and sanctioned units are summed from the building-level unit table, which 4,120 of the district's 4,872 active registrations file; the 48 pincodes charted and quoted are those carrying 20 or more such projects.
• The age check is a Pearson correlation across those 48 pincodes between the share of each pincode's active register registered in 2024 or later and its unsold share, r = 0.07.
• Unsold share is stock standing today against units sanctioned, not a velocity measure. MahaRERA publishes current state and no history, so no absorption rate is derived here from the register.
• MahaRERA files no unit configuration, no carpet area per flat and no price, so no figure in this article splits supply or demand by BHK, and the premium-band number above comes from consultancy transaction data rather than from the register.