The bottleneck isn’t the chips.
It’s the power.
You have the GPUs. You have the demand. What you don’t have is a place to plug them in. Koromanda finds unused industrial power across India and gets it ready for AI — so your hardware can actually run.
Koromanda is the coordination layer between stranded industrial power and AI inference deployment.
Sanctioned megawatts on paper are not the same as deployable megawatts at the rack. We close that gap end-to-end — from discovery and feasibility through utility approvals, modular pod installation, energization, and live workload activation.
Find deployable power
Continuously map underutilized industrial allocations across India and score each pocket on the ten variables that actually determine whether a megawatt will reach a rack.
- MW Registry coverage
- DCI scoring · 10 variables
- Tier-2/3 corridor focus
Coordinate deployment
Drive utility approvals, HT energization, transformer procurement, modular pod installation, cooling retrofit, and fibre redundancy in parallel — not sequentially.
- DISCOM workflows
- Parallel HT + permitting
- Modular pod blueprint
Engineer power certainty
Behind-the-meter batteries, harmonic filtering, N+1 redundancy, and continuous telemetry sit between the grid and the rack so AI workloads run on conditioned power, not utility hope.
- Behind-the-meter storage
- Harmonic conditioning
- Uptime telemetry
OutcomeInference workloads activated on conditioned, deployable power across India’s Tier-2/3 industrial corridors — engineered, not hoped for.
- A deployable AI power network
- Infrastructure coordination middleware
- Engineered power certainty
- Operational telemetry + governance
- A GPU cloud or hardware reseller
- A legacy data centre builder
- A real-estate listing site
- A consulting engagement
Five operational stages from requirement to live workload.
Koromanda sits between AI inference operators, the industrial power scattered across India, and the utilities that need to approve it. The steps below show what happens from the moment an operator sends a requirement to the moment workloads go live.
Infrastructure Requirement
An AI inference operator sends Koromanda a regional deployment request.
Example · A 5MW deployment in the Hyderabad corridor.
Capacity Discovery
Koromanda matches the requirement against the MW Registry — its live inventory of industrial power available across India.
Deployability Validation
Koromanda checks each site: how much of the megawatts allocated on paper will actually be delivered, whether the substation is accessible, transformer lead times, whether cooling can be retrofitted, fibre redundancy, and what permits are needed.
Deployment Coordination
Koromanda coordinates the utility approvals, permits, energization, modular-pod installation, and monitoring setup.
Workload Activation
Inference infrastructure goes live on conditioned, deployable power. Existing sites can be retrofitted faster than building new ones.
The Deployable Capacity Index.
The Deployable Capacity Index (DCI) scores infrastructure deployability across ten weighted variables. Sanctioned megawatts are not the same as deployable megawatts — substation queues, harmonic distortion, transformer lead times, and utility political risk make most stranded capacity operationally unusable for AI inference. The DCI converts that ambiguity into a single score.
The DCI is Koromanda’s internal scoring system. It feeds the deployment work — not a separate analytics product.
Grid stability and the gap between megawatts allocated on paper vs actually delivered together account for 35% of the score. The other 8 variables make up 65%.
∑ w = 1.00FIRST MW REGISTRY DATA · SEEDING Q2 2026
A hypothetical 8MW industrial parcel in an industrial corridor near Pune, scored across all ten variables. The composite reads what the DCI does, not what any specific customer site has been measured at.
A score above 0.70 usually means the site can be retrofitted from an existing shell. A score below 0.50 means the blockers are serious enough that a new build is likely needed.
Programmatic deployment across four phases.
Each phase has defined inputs and outputs. Retrofitting an existing industrial shell is meaningfully faster than building new — exact timing depends on the site.
Intelligence & Advisory
- DCI scoring
- Site validation
- Utility feasibility
Deployment Orchestration
- Energization
- Permitting
- Vendor orchestration
Operational Governance
- Telemetry platform
- Compliance automation
- Uptime analytics
Capital Coordination
- Project-finance vehicle coordination
- Financing orchestration
- Underwriting data licensing
One brief. Four fields. 48-hour turnaround.
A regional capacity brief covers DCI scoring across the regions you select, transformer lead times, substation access conditions, and a deployment feasibility read. No upsell, no newsletter, no sales sequence.