Infrastructure Built for What's Next

Two integrated service lines, engineered to serve the most demanding digital workloads on the planet, deployed across India's most strategic locations.

Build to Suit
01Mega-Scale. Giga-Scale. Free-Scale.

Build to Suit

We design and deliver purpose-built, AI-ready data center campuses tailored to your technical, operational and commercial requirements, from free-scale 50 MW buildings to giga-scale 1 GW+ ecosystems. Our full-lifecycle model spans site acquisition, energy planning, design and engineering, construction and data-driven operations, backed by renewable energy and energy-as-a-service.

Free→Giga
Campus Scale
2 GW+
Pipeline by 2032
10–300 kW
Rack Density
99.999%
Operational Uptime
  • Free-scale, mega-scale and giga-scale delivery models, up to 1 GW+ ecosystems
  • Full lifecycle ownership: site acquisition, energy planning, design, build, operate
  • Energy-as-a-service: solar/wind generation, private transmission, on-site microgrids
  • AI-ready, future-proof design with programmable whitespace and hardware-agnostic flexibility
Discuss This Solution
Colocation
02Space, Power & Cooling at Hyperscale Density

Colocation

Carrier-neutral colocation engineered for the full spectrum of workloads, from Colo+ cabinets to dedicated cages and private suites. Flexible power densities from 3 kVA to 300 kW per rack, precision cooling, biometric dual-authentication security and 24×7 remote-hands support, all monitored through our DCIM platform.

3–300 kW
Power / Rack
42U–52U
Rack Sizes
Tier III
Design
24×7
Remote Hands
  • Colo+, dedicated cages and private suites for CPU and GPU workloads
  • Flexible power densities from 3 kVA to 300 kW+ per rack
  • Dual PDUs, fiber/copper cross-connects and biometric dual-authentication security
  • Carrier-neutral meet-me room with 60+ network providers
Discuss This Solution

400 kV × 2 / 220 kV

Two independent grid sources, with a 220 kV interim bridge

Every campus is planned around a 400 kV connection from two independent sources, with both the central (CTU) and state (STU) transmission routes kept open. Early phases energise on a faster 220 kV interim supply while the full-build connection is completed.

400 kV substation and transmission towers at golden hour
01
400 kV × 2
Independent grid sources, CTU or STU
02
400/33 kV GIS
On-campus switching station
03
33 kV ring
MV ring main to every building
04
33 kV → 415 V
Direct-to-pod, two fewer stages
05
2.2–2.5 MW pod
UPS + Li-ion, BESS, DG N+1
06
≤160 kW rack
Air today, liquid when you need it
2 × 400 kV
Independent sources per campus
220 kV interim
Bridges Day-1 energisation
24 months
EHV line and substation delivery
3–4% lower
Losses with MV direct-to-pod

One platform, sized for the way you buy capacity

Hyperscalers buy in campuses and availability zones. AI clouds buy in pods and need them energised fast. TeraNova's standard pod-floor-building template serves both from the same land, power and supply chain.

Campus-scale capacity to your standard
01
Hyperscale cloud

Campus-scale capacity to your standard

  • 108–150 MW buildings on 48 to 170-acre campuses, at 22–25 MW of IT per acre
  • Multi-parcel clusters in Mumbai and Hyderabad, planned as availability zones within metro distance
  • We build to your basis of design and reference architecture, with no redesign between variants
  • Two independent 400 kV sources per campus, with CTU and STU routes both kept open
  • Renewable supply structured for 100% RE commitments through inter-state open access
  • Powered land, powered shell, build-to-suit or BOT, with long-tenor certainty
GPU-ready pods, energised on your ramp
02
Neocloud & AI cloud

GPU-ready pods, energised on your ramp

  • Start with a single 2.2–2.5 MW pod and add pods as contracts land, with no idle capex
  • Structure, cooling and power sized for GB300 today and Vera Rubin next
  • 42 kW air-cooled to 160 kW liquid-cooled per rack, with CDU power and manifolds pre-laid
  • OCP Open Rack V3 (21", 48 VDC) envelope with AC + DC hybrid per pod
  • Adjacent pods handed over four weeks apart; new pods hot-tied without shutting live halls
  • India-resident capacity for sovereign AI and data-residency workloads, run by a 24x7 NOC

Five ways to take capacity

Choose how much TeraNova delivers and how much you own. Every model sits on the same entitled land, the same grid connection and the same tested building template, so you can start with one model and move to another as your programme grows.

Prepared industrial land for a data center campus
Powered land

Entitled land with a path to power

Clear-title industrial land with zoning, statutory approvals, power sanction and the EHV route in hand. You design and build.

Best fit

Hyperscalers running their own design and construction programmes

Powered shell

Core, shell and power to the building

A GPU-grade structure (2,500 kg/m², 7 m+ clear) with substation, MV supply and site utilities delivered. You fit out the halls.

Best fit

Hyperscalers with a standard in-house MEP fit-out

Build-to-suit · BTS

Turnkey to your specification

We design, procure, build and commission a facility to your basis of design, then lease it to you long term. Operations by you or by us.

Best fit

Hyperscalers who want capex-light capacity built to their own standard

Build-operate-transfer · BOT

We build and run it, then hand it over

TeraNova delivers and operates the facility under SLA for an agreed period, trains your team, and transfers ownership on agreed terms.

Best fit

Platforms and sovereign programmes that want eventual ownership with delivery risk carried by us

Wholesale colo

Ready-to-fit pods, floors and halls

Powered, cooled and operated space in a TeraNova building, from one 2.2 MW pod up to a 24–30 MW floor. You bring the IT.

Best fit

Neoclouds, GPU-as-a-service, AI labs and hyperscaler burst capacity

Scope Matrix
Who delivers what, model by model
Scope of delivery01Powered land02Powered shell03BTS04BOT05Wholesale colo
Land, title & approvalsTeraNovaTeraNovaTeraNovaTeraNovaTeraNova
EHV connection & substationTeraNovaTeraNovaTeraNovaTeraNovaTeraNova
Core & shellClientTeraNovaTeraNovaTeraNovaTeraNova
MEP fit-out & commissioningClientClientTeraNovaTeraNovaTeraNova
Facility operationsClientClientEitherTeraNovaTeraNova
Asset ownershipLong leaseTeraNovaTeraNovaClientTeraNova
Typical unitPlot or campusBuildingBuildingBuildingPod or floor
Your capexHighestHighLowestDeferredLowest
TeraNova TeraNova scopeClient Client scopeEither Agreed per contract

Designed from the slab up for GB300 today and Vera Rubin next

Liquid cooling is in the structural model from Day 1. Structure, cooling and power are sized for three GPU generations, with no retrofit through 2030.

Liquid-cooled server rack manifolds and precision piping
42 → 160 kW
per rack, air-cooled today, liquid-ready
~2,000 kg
Vera Rubin-class rack weight carried by design
7 m+
Slab-to-slab clear
Room for overhead manifolds, busway and an immersion-ready plenum
2,500 kg/m²
Floor loading
About twice the industry norm; SEOR-certified with ETABS
2,500 kg
Point load on 4 casters
24" × 42" footprint; GB300 ≈1,400 kg, Vera Rubin ≈2,000+ kg
Slab-first
Raised floor optional
Flooded-room default with 50% concentrated-load capacity
CDU + SFN
Liquid cooling in the model
CDU and manifold weight designed in, never added later
4" CHW
Pre-laid for liquid
Blinded chilled-water flanges, SFN pathways and CDU power in place
ORv3
Open Rack V3 envelope
21" rack, 48 VDC, with AC + DC hybrid per pod
IS 1893
Dynamic & seismic analysis
Rolling-rack load paths and VC-C vibration criteria included

Design once, build many

One standard pod scales to a floor and a floor scales to a building. The same template delivers G+5, G+6 and G+7 formats, so a new site starts with a proven design and a qualified supply chain.

Prefabricated data center modules being lifted into place by a crane
Pod
2.2–2.5 MW

Sized to GPU generation and tenant ramp. Fault isolation (N+2C / N+C+M) stays inside one pod.

Floor
21.6–24 MW

9–10 pods on one repeatable template, scaling linearly.

Building
108–150 MW

G+5 to G+7, with height and floor count set by the site.

14–22 mo
Design start to first pod ready-for-service
DFMA ≥ 60%
Factory-built modules cut site interfaces and improve quality
One floor / 10 days
Jump-form core; shell complete in 16 weeks or less
4-week twin-pod handover
Adjacent pods go live four weeks apart

About 40% faster deployment than conventional builds. Comparison indicative, from design start to first pod RFS.

Let's build your next campus together

Speak with our team about build-to-suit or colocation capacity.

Contact Us