This record did not exist until 26 October 2026, and its absence was the clearest single defect in the registry. PDG’s MU1 campus publishes 150 MW across five buildings, of which two buildings were live at a combined 48 to 50 MW — DC1 delivered in Q4 2022 and DC2 in 2024-25. The registry held a campus ceiling (PRJ-026, 150 MW, not countable) and an expansion phase (PRJ-239, 100 MW, countable) and no record at all for the capacity that is running. The result was a registry that counted a hundred megawatts of steel that is not finished and not the fifty that customers are using today. PRJ-242 fixes it with 48 MW — the lower of the two published figures, taken from the more recent source. The effect is not cosmetic: PDG moves from absent to fourth in the national operating ranking, and Maharashtra’s operating total rises from 328.7 MW to 376.7 MW, a 14.6% increase from one record.
1At a glance
2What this record contains
| Field | Value | Why |
|---|---|---|
| Record type | Facility | A building, not a campus and not a programme |
| Status | Operational | Both buildings are running and serving customers |
| Capacity | 48 MW | Two buildings combined; the lower published figure |
| Basis | IT capacity | The operator describes the campus in IT megawatts throughout |
| Additive flag | No | Child of PRJ-026: adding 48 to the 150 MW campus ceiling would double-count |
| Counted by the reconciliation | Yes, operating tier | The most granular record carrying this capacity |
| Combined rather than split | Two buildings, one record | No per-building capacity is published; PRJ-095 and PRJ-161 set the precedent |
| Related | PRJ-026, PRJ-239 | Campus ceiling, expansion phase |
The additive flag is “no” and the record is still counted, and that is not a contradiction. The flag answers one question — may this be added to its parent? — and the answer is no, because the parent is a 150 MW campus that already includes it. The reconciliation counts the most granular record carrying the capacity, which is this one. PRJ-239 works the same way: flagged non-additive, counted at 100 MW.
What the registry now holds for MU1 is 150 MW of campus claim, 48 MW operating and 100 MW under construction, of which 148 MW is counted and 2 MW is the gap between the operator’s live figure and the one counted here.
3The 48-or-50 decision
| Source | Live capacity | Date | Wording |
|---|---|---|---|
| Data Center Dynamics | 48 MW | 14 Aug 2026 | “MU1 currently offers 48MW across two buildings” |
| Princeton Digital Group | 50 MW | 30 Jul 2025 | “the first two buildings with 50 MW capacity are operational” |
| Data Center Dynamics | 50 MW | May and Jul 2026 | “the first two buildings at the site are live with a combined capacity of 50MW” |
The registry counts 48 MW and publishes both. Three reasons for taking the lower figure:
| Reason | Detail |
|---|---|
| It is the more recent statement | August 2026 against July 2025 |
| It is the more specific statement | “Currently offers”, in a report about the campus footprint, rather than a programme announcement |
| The registry’s convention is conservative on a live-capacity conflict | The same choice is made on STT GDC Ahmedabad 1, where 4 MW is counted against a 4 MW published page on a 3 MW figure |
| The cost of being wrong is asymmetric | Counting 48 understates the national total by 2 MW; counting 50 overstates it. A registry should err towards the floor |
Two megawatts is 4% of this record and 0.26% of the national operating total. It is not the point. The point is that the disagreement is recorded, attributed and dated, so that anyone who prefers 50 MW can see exactly what this registry did and why.
The same two buildings are 48 MW and 50 MW in the same year. That is a 4% measurement error on a live, operating facility, and it exists in the published record of one of the best-documented operators in India.
Scale it. If a 50 MW live campus can be reported 4% apart, the 400 MW company aggregate on PRJ-234 and the 150 MW rating-agency figure for CtrlS on PRJ-238 should be read as ranges, not points.
That is the whole argument for a record-level registry. Building records do not eliminate the error; they localise it.
4What 48 MW changes
| Measure | Before 26 Oct 2026 | After | Change |
|---|---|---|---|
| National operating capacity | 730.69 MW | 778.69 MW | +48.00 MW |
| National operating and building | 1,229.69 MW | 1,277.69 MW | +48.00 MW |
| All counted records | 3,465.69 MW | 3,513.69 MW | +48.00 MW |
| Counted records | 129 | 130 | +1 |
| Maharashtra operating capacity | 328.73 MW | 376.73 MW | +14.6% |
| Maharashtra total, all stages | 1,503.9 MW | 1,551.9 MW | +3.2% |
| PDG operating capacity | 0 MW | 48 MW | Absent to fourth nationally |
One record moved a national figure by 6.6% and a state figure by 14.6%. That is what a coverage defect costs: not a rounding difference but a market description that was wrong.
The operator ranking is the more interesting change. PDG had no operating capacity in this registry at all — a company with a live campus in Navi Mumbai, a 100 MW expansion under construction, a 120 MW second campus in Mumbai, a 90 MW Hyderabad campus, a 72 MW Chennai campus and about US$2.5 billion committed to India. It now sits fourth nationally at 48 MW, ahead of every operator with fewer than five counted buildings.
That is a description of the registry’s weakness, not PDG’s strength. A company that large should not have been at zero. It was at zero because its live capacity was published inside a campus aggregate and nobody had separated it out.
5Why the record was missing
- MU1 announced and DC1 delivered in Q4 2022 on a campus planned at 150 MW.
- PDG announces the India programme: three new buildings, 100 MW, campus to 150 MW. The record set captures the campus, the expansion and the funding — not the two buildings already running.
- DC2 goes live. Coverage continues to describe 50 MW operational without a per-building record.
- The expansion phase PRJ-239 is counted at 100 MW. The reconciliation audit notices that the live capacity has no record and that the operating tier is therefore understated.
- PRJ-242 created at 48 MW. National operating capacity moves from 730.69 MW to 778.69 MW.
The mechanism is mundane and it is everywhere in public-data registries. Research passes track announcements, because announcements are what gets published. A building that opened two years ago and has been quietly selling capacity since generates no new announcement, so it generates no new record.
The registry’s reconciliation found the same defect in a different form in 45 container records: campuses and portfolios marked non-additive, with the capacity underneath them never counted. Between the two — orphaned children and missing records — 1,234 MW of real capacity was absent from the published total, against a pre-existing headline of 2,420.37 MW.
PRJ-242 is the first record in this registry created specifically to close that kind of gap, and it will not be the last. The review queue names the rest: Nxtra at 238.2 MW and AirTrunk at 601.0 MW, both parked in the pipeline tier for want of a status confirmation.
6How this record is counted
| Rule | Applied here |
|---|---|
| Scope | IT capacity, because PDG describes the campus in IT megawatts |
| Level | The most granular record carrying this capacity is this one; the campus above it is a ceiling |
| Stage | Operating tier — both buildings are commissioned and serving customers |
| Macro | Not a programme, MoU, portfolio or network record |
| Additive flag | No, because the parent campus already contains it. The flag governs summation to the parent, not counting in the national total |
One risk is disclosed rather than resolved. If PDG later publishes a per-building split that sums to 50 MW, this record should be replaced by two records and the count should move by 2 MW. The combined form is the correct one until a split exists, and it is flagged combined_facility in the master so that a future pass finds it.
| Flag set on this record | Meaning |
|---|---|
combined_facility | Two or more buildings in one record because no split is published |
capacity_conflict_48_vs_50 | Two published live figures; the lower is counted |
no_per_building_split | Neither capacity nor floor area is attributed per building |
7What would change this record
| Item | Effect |
|---|---|
| A per-building capacity split | Replaces this record with two; the national total moves by 0 to 2 MW |
| Confirmation of 48 against 50 MW | Settles the live figure and the per-building average |
| Occupancy or contracted share | The difference between energised and sold, which is what an acquirer prices |
| A date for the first 100 MW expansion phase | Moves part of PRJ-239 from building to operating, the single largest pending upgrade in Maharashtra |
The fourth item is worth more than the first three. MU1 has 100 MW under construction and a first phase scheduled for 2026. As that delivers, Maharashtra’s operating capacity moves again — and unlike this record, that one will be visible in the announcements.
8Sources
Sources: Data Center Dynamics, 14 August 2026, stating that MU1 in Mumbai offers 48 MW across two buildings with a 150 MW target across 15 acres; Princeton Digital Group newsroom, 30 July 2025, confirming two operational buildings at 50 MW with the first 100 MW phase under construction and the US$160 million green loan from Axis Bank, Kotak Mahindra Bank and Canara Bank; Data Center Dynamics, May, July and February 2025 to 2026 coverage of the MU1 campus, the 50-acre Airoli park, the DC2 launch and India’s first hourly carbon-free energy matching scheme with Tata Power Renewable and Flexidao; Princeton Digital Group OCP Ready v2 release describing MU1 as 150 MW of critical IT capacity across five buildings with a dedicated 220 kV GIS; Gridvik records PRJ-026 and PRJ-239. Record created and verified 26 October 2026.