NTT's Vikhroli IV data centre carries 3 MW of critical IT load and 4 MVA of total facility load, with 3 to 6 kVA per rack, on the Godrej and Boyce Industry Estate in Vikhroli West. It sits about a kilometre from Vikhroli, the other 2012-vintage NTT site, and the two together are 5.4 MW — 3.4% of what NTT counts in Mumbai. The published numbers do not agree with each other. Four megavolt-amperes behind three megawatts implies a ratio of 1.20 to 1.33, better than the operator’s own 1.48 hyperscale halls. A third-party listing puts total power at 10 MVA, which implies 3.00. Anywhere between 1.20 and 3.00 is possible, and the truth is probably 2.4. The page publishes all of it and resolves none of it.
1At a glance
2The pair that is 3.4% of Mumbai
Vikhroli and Vikhroli IV are two buildings on the same stretch of LBS Marg in Vikhroli West, about a kilometre apart, both operated by NTT, both small, both with the same 3 to 6 kVA rack specification.
Five point four megawatts across two buildings, against 155.3 across the Chandivali five. The pair exists for a reason that has nothing to do with capacity: it gives NTT a second Vikhroli site for recovery, on a different estate, inside the same city and away from the Chandivali campus.
| Measure | Vikhroli (VIK DC2) | Vikhroli IV (VIK DC4) |
|---|---|---|
| Critical IT load | 2.4 MW | 3 MW |
| Facility load, operator | 6 MVA | 4 MVA |
| Facility load, third party | not listed | 10 MVA |
| Rack power | 3 to 6 kVA | 3 to 6 kVA |
| Estate | Mehra Industrial Estate | Godrej and Boyce Industry Estate |
| Positioning | general colocation, built 2012 | disaster recovery and continuity |
| Counted here | 2.4 MW | 3 MW |
Neither building can be sold on density or efficiency. Together they can be sold on something no hyperscale campus in Navi Mumbai offers: two separate Mumbai-city buildings, under one operator, for recovery.
That is 5.4 MW of edge product carrying 3.4% of a 161 MW book — and it is the only part of NTT’s Mumbai estate that is genuinely scarce.
3Four numbers, three ratios
Vikhroli IV is the clearest example in this registry of a record where the published numbers cannot all be describing the same thing.
| Reading | Ratio | Plausible? | Why |
|---|---|---|---|
| 4 MVA at 0.9 pf = 3.6 MW | 1.20 | no | better than NTT’s own new hyperscale halls, on a DX-cooled legacy site |
| 4 MVA treated as 4 MW | 1.33 | no | same problem, slightly less extreme |
| 4 MVA is a per-feed rating, dual feed | 2.40 | probably yes | consistent with the sister building at 2.25 |
| 10 MVA at 0.9 pf = 9 MW | 3.00 | possibly | but 10 MVA does not come from any operator statement |
Two of these readings are effectively impossible. A 3 MW hall with direct expansion cooling and 3 to 6 kVA racks cannot run at 1.20, because the fixed plant alone is a larger multiple of that. So the 4 MVA figure is almost certainly not a site total — it is a per-path rating on a dual-feed design, which makes the real facility load about 8 MVA and the ratio about 2.4.
That is an inference, and this page labels it as one. The registry does not correct the operator’s figure. It counts the 3 MW of IT, publishes the 4 MVA as stated, shows the 10 MVA listing alongside it, and lets the reader see the gap.
A tenant evaluating Vikhroli IV against Chandivali would conclude it is a third more efficient than a 2023 hyperscale hall, and price accordingly.
On the 2.40 reading it is 62% less efficient. The difference on 3 MW of IT is roughly Rs 12 crore a year at Rs 8 per kWh and 80% loading.
4The disaster recovery claim
NTT describes Vikhroli IV as offering the highest levels of disaster protection and recovery in India. That is a marketing statement, and there is no published standard behind it — not an Uptime Institute certification number, not a TIA-942 rating, not an audited RTO.
| Attribute | What is published | Verifiable? |
|---|---|---|
| Separate building from Vikhroli | yes, different estate, about 1 km | yes |
| Separate power feed | dual feed stated | partly |
| Geographic separation from Chandivali | about 20 km, different part of Mumbai | yes |
| Certification level | not published as a rating | no |
| Stated recovery time objective | not published | no |
| “Highest levels in India” | operator marketing language | no |
The defensible part of the claim is structural, and it is enough to sell on: two independent Mumbai-city buildings on two separate industrial estates, both inside the city rather than out in the Navi Mumbai belt. For financial-services tenants that must stay close to the Fort and BKC corridors, that is a real constraint and a real product.
The indefensible part is "highest levels in India". Nobody can check it, which is exactly why the operator says it.
5Procurement
6Limits and open questions
- The four numbers on this page do not reconcile. 4 MVA and 10 MVA cannot both be the total facility load of a 3 MW building, and neither supports a credible ratio on its own.
- The 2.40 ratio is an inference, from a per-feed reading of a dual-feed design. It is labelled as an inference, not a published figure.
- "Highest levels of disaster recovery in India" is unverifiable and is treated here as marketing language, not data.
- Rack count, floor area and PUE are all unpublished for this facility.
- The 10 MVA figure comes from a directory listing with no operator confirmation.
- No dated commissioning or occupancy evidence.
- What would close it. A power factor, a per-feed clarification and a rack count. Three lines on a facility page would collapse this record from four ratings to one.
7Sources
Sources: NTT Mumbai Vikhroli IV facility page, NTT DATA global data center locations index, Inflect facility listing for NTT DC-4, and Datacenters.com facility listing.