E2E Networks Ltd Q1 FY26 – GPU Dreams, Sovereign Clouds, and P/E That Needs Its Own Cloud Server
1. At a Glance
E2E Networks, the NSE-listed AI cloud kid, has a market cap of ₹6,936 Cr and CMP of ₹3,451 (up 31.9% in last 3 months because, well, AI = magic word). But the P/E? Hold your GPUs – 201x. That’s not valuation, that’s an April Fool’s prank that never ended.
Quarterly sales dipped to ₹36.1 Cr (–12.6% YoY) and PAT crashed into negative ₹–2.84 Cr (–128% YoY). Still, full-year FY25 had 50% OPM and ₹34.5 Cr PAT, proving the business makes money… just not every quarter. Promoters hold 40.6%, but 60% of that is pledged – basically your neighborhood kirana wala putting his shop papers for a Bajaj Chetak loan.
Question for you: would you trust your life savings with a company that owns H200 GPUs but also pledges 60% of its promoter shares?
2. Introduction
Welcome to the great Indian AI GPU bazaar. While Microsoft, Google, and AWS flex trillion-dollar cloud muscles, tiny E2E Networks shows up like the David of Dwarka with ~3,700 GPUs, claiming it’s India’s answer to NVIDIA’s fan club.
The pitch: hyperscale AI-first cloud, sovereign cloud for governments, AI Labs-as-a-Service for colleges, and GPU-as-a-Service for anyone training chatbots that mostly answer, “I’m sorry, I don’t have that information.”
In five years, E2E went from being a no-name cloud infra vendor to the “H100/H200-first mover in India.” Customers include Zomato, Nykaa, Cars24, Healthkart, and other startups who would rather rent GPUs than blow VC money on their own data centers.
But let’s not kid ourselves. This is a high-growth, high-volatility smallcap. One quarter you’re printing 66% EBITDA margins, next quarter you’re bleeding red like an IPL franchise.
3. Business Model – WTF Do They Even Do?
E2E’s model is simple: buy expensive NVIDIA GPUs → rent them out → pretend you’re a hyperscaler.
Cloud GPU Rentals – H100s, H200s, A100s, V100s. The GPU buffet is open, just bring your credit card.
TIR AI/ML Platform – Not a shampoo, but their AI DevOps platform: Jupyter notebooks, LORA fine-tuning, vector DBs, secure resumption – basically GitHub Copilot’s playground.
Sovereign Cloud – For Indian govt & enterprises wanting data to stay in-country. Think “Atmanirbhar Cloud” with an Aadhaar stamp.
AI Lab as a Service – Tie-ups with colleges to let students play with GPUs. Imagine NIT Kurukshetra kids training diffusion models instead of doing assignments.
Self-Service Public Cloud – AWS-style portal, but only in Delhi NCR, Chennai, and a pocket-sized Mumbai DC.
Bottom line: they’re not building rockets. They’re selling GPUs as PG accommodations – fully furnished, no deposit, but pray the power doesn’t trip.
4. Financials Overview
Source table
Metric
Latest Qtr (Q1 FY26)
YoY Qtr (Q1 FY25)
Prev Qtr (Q4 FY25)
YoY %
QoQ %
Revenue
36.1
41.3
33.5
–12.6%
+7.7%
EBITDA
10.5
27.3
13.3
–61.5%
–21.1%
PAT
–2.8
10.1
13.6
–128%
–120.9%
EPS (₹)
–1.42
7.0
6.8
NA
NA
Commentary: One quarter you’re riding GPUs like Virat Kohli in form, next quarter you’re KL Rahul in a knockout match.
