Revenue
Margin
Demand
Guidance
Tone
E2E Networks - Q1 FY27 Earnings Call Summary Wednesday, July 22, 2026 8:00 AM IST
Event Participants
Executives
2 Nitin Jain, Tarun Dua
Analysts
13 Abhishek Shindadkar, Ashish Ajit Golechha, Bharat Gulati, Bhavya Gandhi, Chirag Satiya, Keshav, Neel Munot, Nishant Joshi, Rohan Nagpal, Shubham Tamrakar, Sucrit Patil, Varun Gandhi, Vedant
Financials & KPIs
| Metric | Reported | Commentary |
|---|---|---|
| Revenue | ₹156.8 crores | Up 334% YoY and 64% QoQ; driven by operating leverage and 1,024 B200 GPUs going live |
| EBITDA | ₹117.9 crores | Strongly up QoQ on capacity additions and near-maximal utilization |
| EBITDA Margin | 75.2% | Expanded 1,450 bps QoQ vs Q4 FY26; operating leverage from higher capacity |
| Profit Before Tax | ₹58.6 crores | vs ₹8.6 crores in Q4 FY26 |
| Profit After Tax | ₹43.9 crores | Driven by B200 cluster going live and operating leverage |
| Live GPU Capacity | ~5,100 GPUs | Current live capacity; excludes next 1,024 B200 lot expected in ~2 months |
| Outstanding Debt | ~₹450 crores | Debt taken to fund first B200 lot; expected to increase with next lot |
Geographic & Segment Commentary
- Sovereign AI Platform: Management emphasized open-weight models running on E2E's AI factory, using TIR and Jarvis Labs software layers. Customers gain control over deployment, fine-tuning, data access, and state retention; open-source models are now close to frontier-model capability, enabling "AI Sovereignty" while avoiding dependence on rented, externally controlled frontier APIs.
- SovCloud Subsidiary: Infrastructure-focused subsidiary for large-scale CPU/GPU cluster build-out; exploring funding arrangements to accelerate AI infrastructure. Management stated plans are early days and specifics will be announced as execution progresses.
- U.S. (Delaware) Entity: Focused on international sales, alliance management, and servicing customers outside India.
- Geographic Revenue Mix: In Q1 FY27, India AI mission contributed ~20–21% of revenue, international ~37%, and the remainder from domestic customers. This compares with ~40% India AI contribution in the prior quarter (reference point from analyst discussion).
Company-Specific & Strategic Commentary
- Blackwell Deployment: 1,024 B200 GPUs went live during Q1 and were put on revenue; utilization described as near-maximal. Another 1,024 B200 lot is expected within the next couple of months.
- Capacity Expansion: Management plans more Blackwell GPUs (including non-flagship variants) and has started preliminary work on next-generation architectures like B300 and Vera Rubin. Funding for announced GPUs is fully arranged via debt, internal accruals, and prior equity raises.
- Pricing Actions: E2E implemented July price increases for both GPU and CPU, citing robust demand and higher input costs—memory cost increases hitting the CPU side harder. Existing contracts are honored; some customers opted for 1–2 year contracts or advances to lock current prices.
- Customer Contracting: Rising demand for 1–3 year contracts is improving revenue predictability; management will maintain a mix of contracted and on-demand/spot capacity.
- L&T Partnership: Arm's-length arrangement—E2E buys data center capacity from L&T, L&T uses E2E cloud, and the two jointly engage customers; E2E may also offtake future L&T compute capacity.
- Operating Philosophy: "Do less and do better"—focus on high-quality partnerships and long-term growth rather than chasing vanity metrics.
Guidance & Outlook
| Metric | Guidance / Outlook | Commentary |
|---|---|---|
| Revenue / MRR | No explicit guidance | Management declined to guide MRR; advised tracking historical quarterly exit MRR |
| B200 Capacity | Next 1,024 B200 lot expected in next couple of months | Deployment as soon as available; utilization expected to remain near-maximal |
| Capacity Roadmap | More Blackwell (incl. non-flagship) and Vera Rubin | No GPU count or timing disclosed; described as "aggressive and judicious" |
| Pricing | July price hikes for GPU and CPU | Growth expected to come from capacity additions rather than pricing alone; contract renegotiation happening with minimal pushback |
| Margins | Current margin level sustainable over medium-to-long term | Supported by utilization, robust demand, and operating leverage; judicious contracted vs on-demand mix |
| Customer Contracts | Increasing share of longer-term (1–3 year) contracts | Improves revenue predictability; exact contracted share not disclosed |
Risks & Constraints
| Risk | Context |
|---|---|
| GPU supply/timing delays | Next B200 lot expected in "a couple of months"; any slippage could delay capacity additions and revenue growth |
| AI pricing/commoditization | Management dismissed bubble concerns but acknowledged demand-supply phases; open-source model efficiency could pressure pricing, though management sees 5–6 year asset lifecycles |
| Input cost inflation | Memory cost increases, especially on CPU side, drove price hikes; sustained cost inflation could pressure margins if pricing power fades |
| Leverage/funding | Debt of ~₹450 crores will increase with next B200 lot; management declined to quantify peak debt or future equity raises; heavy capex needs could strain the balance sheet |
| NVIDIA ecosystem dependency | Platform deeply integrated with NVIDIA; ASIC or non-NVIDIA accelerators could alter demand; management says it is vendor-neutral and open to integrating other platforms |
| Competitive intensity | Hyperscalers scaling aggressively and customers getting harder on pricing (raised by analyst); management leans on 16+ years of cycle experience |
Q&A Highlights
Revenue Growth Drivers
- Question: What drove the revenue/MRR jump—volume/utilization vs pricing? (Neel Munot)
- Answer: Growth was driven mainly by higher utilization and capacity additions; pricing had a moderate impact. The B200 cluster is at near-maximal utilization. (Tarun Dua)
Capacity and B200 Deployment
- Questions: When will the next B200 lot arrive, and can you quantify the capex plan? (Bhavya Gandhi, Shubham Tamrakar)
- Answer: Next B200 lot expected within a couple of months; management is also planning more Blackwell (including non-flagship) and Vera Rubin, with no GPU count disclosed. Funding for announced GPUs is fully arranged via debt, internal accruals, and prior equity raises. (Tarun Dua)
Margin Sustainability
- Question: Are the record margins sustainable? (Varun Gandhi)
- Answer: Yes, supported by increased utilization, robust demand, and operating leverage; management will maintain a judicious mix of contracted vs on-demand revenue and sees this as sustainable over the medium-to-long term. (Tarun Dua)
Pricing and Customer Contracts
- Questions: What is driving CPU/GPU price increases, and are customers pushing back? (Rohan Nagpal, Shubham Tamrakar)
- Answer: Price hikes reflect both strong demand and higher input costs—memory cost increases have hit CPU harder than GPU. Customers wanting to hold prices were offered longer-term contracts (1–2 years) or advances; post-contract renegotiation is happening with minimal pushback. (Tarun Dua)
Revenue Mix and Longer-Term Visibility
- Questions: How is the domestic/international/India AI mix trending, and are customers locking in longer capacity? (Bharat Gulati, Abhishek Shindadkar)
- Answer: In Q1, India AI was ~20–21% of revenue and international ~37%, with rest domestic. Customers are increasingly asking for 1–3 year contracts, improving revenue predictability; management will shift some capacity to longer-term contracts while retaining flexibility. (Tarun Dua)
Subsidiaries and Funding
- Questions: What are the Delaware and GPU-infrastructure entities for, and how will SovCloud be funded? (Keshav, Bhavya Gandhi, Ashish Golechha)
- Answer: SovCloud is an infrastructure subsidiary for large-scale cluster build-out; Delaware entity handles international sales and alliances. Current loan outstanding is ~₹450 crores and will rise with the next B200 lot; peak debt was not quantified, and future equity raises, if any, will be announced. (Tarun Dua, Nitin Jain)
AI Industry Outlook and Competitive Pressures
- Questions: How do you view AI bubble concerns and cheaper open-source models? (Vedant, Sucrit Patil)
- Answer: Management believes AI is at "day zero" of a decadal supercycle; media headlines are exaggerated, and enterprise adoption remains early. E2E has navigated cycles for 16+ years and remains confident of growth through every cycle. (Tarun Dua)
GPU Lifecycle and Platform Flexibility
- Question: How do you protect ROIC on older GPUs and address potential ASIC/non-NVIDIA competition? (Ashish Golechha)
- Answer: Each GPU generation has a minimum 6-year life, aided by open-source fine-tuning that keeps older models competitive; newer generations find their own price-performance sweet spot. E2E is open to integrating any accelerated computing platform and describes itself as vendor-neutral. (Tarun Dua)
Key Takeaway
E2E Networks delivered a strong Q1 FY27 with revenue of ₹156.8 crores, up 334% YoY and 64% QoQ, and EBITDA margin expanding to 75.2% on the back of 1,024 Blackwell B200 GPUs going live and operating leverage. PBT rose to ₹58.6 crores and PAT to ₹43.9 crores. Management is focused on building a Sovereign AI platform atop open-weight models using its TIR and Jarvis Labs software, with subsidiaries SovCloud (infrastructure) and a Delaware entity (international sales). It plans to deploy another 1,024 B200s in the next couple of months, expand into non-flagship Blackwell and Vera Rubin, and is shifting some capacity to 1–3 year customer contracts to improve revenue predictability. July price hikes for GPU/CPU were driven by demand and memory-cost inflation. Key watch points include GPU supply timing, rising debt (~₹450 crores and increasing), and potential pricing/commoditization pressure, though management views current margin levels as sustainable over the medium-to-long term.