Sector Guides
How to Analyze Data Center & Cloud Stocks in India: MW & AI Mix
Team QuartrlyData center and cloud stocks in India live or die on power, not floor space — MW Capacity (IT Load) sets the real ceiling on revenue, and colocation versus cloud pricing per MW can differ by 13x. Anant Raj reported cloud revenue of ₹12 crore per MW monthly against ₹90 lakh per MW for colocation, while Netweb Technologies' AI Systems Revenue Mix rose from 15% of sales in FY25 to 25% in H1 FY26, and Tata Communications' Digital Portfolio revenue grew 71.6% YoY to ₹2,082 crore in Q4 FY24. Reading these metrics together — operational capacity, revenue mix, and order-book visibility — separates operators actually scaling from those merely announcing megawatts.
Key Takeaways
- MW Capacity (IT Load) is the primary measure of a data center's scale — focus on operational MW, not announced targets
- Revenue Per MW varies dramatically between colocation (₹90 lakh/MW) and cloud services (₹12 crore/MW), a 13x difference
- AI Systems Revenue Mix indicates actual exposure to the AI infrastructure boom, not just marketing claims
- Order Book and L1 Pipeline provide forward visibility into revenue, critical in a sector with long sales cycles
- Digital Portfolio Growth tracks legacy telecom companies' transformation into modern technology businesses
Quick Reference
| Metric | Definition | Healthy Range | Warning Sign |
|---|---|---|---|
| MW Capacity (IT Load) | Power available for IT equipment | 85-95% utilization of operational MW | Large gap between announced and operational |
| Revenue Per MW | Monthly revenue per MW | ₹90L-1.2Cr (colo), ₹10-15Cr (cloud) | Declining yield, stuck in pure colo |
| AI Systems Revenue Mix | % of sales from AI hardware | >20% and growing | Stagnant mix despite AI claims |
| Order Book | Confirmed contracts value | 4-6 quarters of revenue coverage | Declining while pipeline grows |
| L1 Pipeline | Lowest-bidder deals pending signature | Converting to orders within 1-2 quarters | High L1 that never converts |
| Digital Portfolio Revenue | Revenue from cloud, security, edge | >20% YoY growth | Single-digit growth, legacy drag |
Understanding Data Center & Cloud Metrics
Data centers are specialized infrastructure facilities that house computing equipment, storage systems, and networking hardware. Unlike traditional real estate, where value is measured in square footage, data center capacity is constrained by power availability. A facility can have ample floor space but still be limited by how much electrical load it can support.
The sector includes three distinct business models: colocation providers who rent physical space and power to customers, cloud service providers who offer computing resources on-demand, and hardware manufacturers who supply servers and systems. Each model has different margin profiles and growth characteristics. Colocation generates stable, lower-margin revenue, while cloud services command premium pricing but require significant technical expertise. Hardware sales tend to be lumpy, driven by large enterprise contracts and government tenders.
MW Capacity (IT Load)
What it is: MW Capacity measures the total electrical power a data center can deliver to IT equipment, expressed in Megawatts (MW). It represents the facility's true operational scale, as power availability — not floor space — is the primary constraint in data center operations.
Why it matters: MW Capacity is the equivalent of "rentable square footage" for data centers. A facility's revenue potential is directly tied to how much power it can sell to customers. Investors should distinguish between operational MW (currently generating revenue) and planned or announced MW (still under construction).
What good looks like: Consistent delivery of operational MW against stated timelines. Anant Raj delivered 28 MW by Q2 FY26 and targets 63 MW by December 2026 and 117 MW by FY28. Leading operators typically achieve 85-95% utilization of operational capacity.
Red flag: Large gaps between "announced" capacity and "operational" capacity, or repeated delays in commissioning timelines. Companies emphasizing gross floor area while avoiding IT Load MW disclosure may be masking inefficient assets.
Example from earnings call:
"Anant Raj delivered 28MW of data center capacity in Q2. The company remains on track for 63MW by December 2026 and 117MW by FY28." — Anant Raj Q2 FY26 Earnings Call
Revenue Per MW
What it is: Revenue Per MW measures the monthly or annual revenue generated per megawatt of operational capacity. It reflects both the service mix (colocation vs. cloud) and pricing power of the data center operator.
Why it matters: This metric reveals the yield on data center assets and indicates business model positioning. Colocation services (renting space and power) generate lower revenue per MW than managed cloud services (providing computing resources and platform services).
What good looks like: In the Indian market, colocation revenue typically ranges from ₹80 lakh to ₹1.2 crore per MW monthly. Cloud services command significantly higher rates — Anant Raj reported cloud revenue of ₹12 crore per MW monthly versus ₹90 lakh per MW for colocation, representing a 13x premium.
Red flag: Declining revenue per MW over consecutive quarters may indicate pricing pressure, customer churn, or adverse mix shift toward lower-value colocation contracts. Pure colocation operators with no cloud migration strategy face long-term margin compression.
Example from earnings call:
"Cloud revenue per megawatt monthly is INR 12 crores." — Anant Raj Q2 FY26 Earnings Call
AI Systems Revenue Mix
What it is: AI Systems Revenue Mix measures the percentage of total revenue derived from AI-specific hardware, including GPU servers, high-performance computing clusters, and AI training infrastructure. It quantifies actual exposure to AI infrastructure demand rather than marketing positioning.
Why it matters: The AI infrastructure buildout represents a significant growth opportunity for hardware manufacturers. This metric separates companies with genuine AI revenue from those merely using AI as a marketing term. Rapid growth in this segment indicates successful positioning in the AI supply chain.
What good looks like: AI Systems revenue growing as a percentage of total sales, with strong year-over-year growth rates. Netweb Technologies saw AI Systems increase from 15% of sales in FY25 to 25% in H1 FY26, with 78% YoY growth in Q2 FY26.
Red flag: Stagnant AI revenue mix despite company claims of AI focus, or AI revenue that remains below 10% of total sales after multiple quarters of supposed pivot. Single-quarter spikes without sustained follow-through may reflect one-time large orders rather than structural demand.
Example from earnings call:
"In H1, AI Systems formed 25% of sales versus 15% in FY25. AI Systems sales in Q2 FY26 grew 78%." — Netweb Technologies Q2 FY26 Earnings Call
Order Book and L1 Pipeline
What it is: The Order Book represents confirmed contracts awaiting execution and revenue recognition. The L1 Pipeline refers to deals where the company has been declared the lowest bidder (L1 status in government tenders) but contracts have not yet been formally signed. Together, they provide forward revenue visibility.
Why it matters: In infrastructure and hardware businesses, revenue is a lagging indicator — it reflects contracts signed months earlier. The Order Book and L1 Pipeline reveal the trajectory of future revenue. A healthy ratio of Order Book to quarterly revenue indicates strong visibility.
What good looks like: Order Book that represents 4-6 quarters of revenue, combined with a growing L1 pipeline that converts to signed orders within 1-2 quarters. Netweb Technologies reported an Order Book of ₹494 crore with an L1 pipeline of ₹348 crore and a broader order pipeline of ₹4,204 crore at end of Q2 FY26.
Red flag: L1 pipeline that remains elevated without converting to Order Book, indicating deal losses at the final stage. Order Book declining while pipeline grows suggests execution or pricing issues. High pipeline-to-revenue ratios without corresponding Order Book growth may indicate optimistic forecasting.
Example from earnings call:
"Order book stood at INR 4.94 billion with order pipeline worth INR 42.04 billion and L1 pipeline worth INR 3.48 billion at end of Q2 FY26." — Netweb Technologies Q2 FY26 Earnings Call
Digital Portfolio Revenue
What it is: Digital Portfolio Revenue measures revenue from modern technology services — cloud, cybersecurity, IoT, and edge computing — as distinct from legacy connectivity services. It tracks the transformation progress of traditional telecom companies into digital infrastructure providers.
Why it matters: Legacy telecom operators face commoditization in their core voice and data transport businesses. Digital Portfolio growth indicates successful diversification into higher-margin, faster-growing segments. This metric separates companies managing decline from those executing transformation.
What good looks like: Digital Portfolio revenue growing at 20%+ YoY while representing an increasing share of total revenue. Tata Communications reported Digital Portfolio revenue of ₹2,082 crore in Q4 FY24, growing 71.6% YoY.
Red flag: Digital Portfolio growth in single digits, or growth that is entirely offset by legacy revenue decline. Headline revenue growth masking stagnant digital transformation indicates a company trapped in legacy business erosion.
Example from earnings call:
"Digital Portfolio revenues stood at INR 2,082 crore, growing strongly at 71.6% YoY." — Tata Communications Q4 FY24 Earnings Call
Special Considerations
The Announced Capacity Trap
A common pitfall in data center investing is overweighting announced or planned capacity. Indian data center construction faces significant execution challenges: land acquisition delays, power infrastructure approvals, environmental clearances, and long commissioning timelines.
Companies frequently announce ambitious MW targets (300 MW, 500 MW) that may take 5-7 years to materialize. Markets often price in this optimism, then punish stocks when execution delays occur. Investors should focus on operational IT load and capital already deployed, not PowerPoint projections. Anant Raj's 307 MW target, for example, should be evaluated against the 28 MW currently operational.
Colocation vs. Cloud Economics
The revenue spread between colocation and cloud services (approximately 13x based on Anant Raj's disclosed figures) creates strong incentives to shift toward cloud. However, cloud services require significant investment in software platforms, technical talent, and customer support infrastructure. Not all colocation operators can successfully make this transition. Evaluate management's cloud strategy and execution track record, not just stated intentions.