E2E Networks Ltd Q2 FY27 Earnings Call: Signs Rs. 1,000 Cr Sovereign AI Deal, Revenue Jumps 411% YoY

E2E Networks Ltd held its Q2 FY27 earnings call on October 16, 2026. Here's a quick read of what management said — performance, strategy, and the outlook ahead.

Hyper-Growth Quarter with Expanding Margins

• Revenue of Rs.2,238 million in Q2 FY 2026-2027 — up 411% YoY and 43% QoQ, driven by Blackwell GPU contribution, better utilization, and higher average realizations.
• PBT of Rs.971 million (up 66% QoQ) — PAT of Rs.727 million with PAT margin expanding to 32.5% from 28% in Q1 FY 2026-2027; EBITDA margin at 76.6% (+141 bps QoQ).
• H1 FY 2026-2027 revenue of Rs.3,805 million (up 376% YoY) — PAT of Rs.1,166 million vs a loss of Rs.163 million in H1 FY 2025-2026.
• Diluted EPS of Rs.3.48 in Q2 FY 2026-2027 — up from Rs.2.10 in Q1 FY 2026-2027.
• Operating costs of Rs.18 crores in Q2 FY 2026-2027 — disclosed alongside the quarter's margin performance.

Rs.1,000 Cr Sovereign AI Deal Anchors Multi-Year Visibility

• ~Rs.1,000 crore contract signed with a sovereign AI company in Q2 FY 2026-2027 — for the Blackwell cluster; management stated the deal improves long-term revenue visibility up to June 2029 and validates the shift from pay-as-you-go to longer-term commitments.
• Contract converts existing B200 usage — Tarun Dua confirmed the Rs.1,000 crore deal is a conversion of current pay-as-you-go B200 usage to a multi-year contract; its contribution to the Rs.75 crore MRR (Q2 FY 2026-2027) varies by usage.
• India AI Mission contributed 13–15% of total revenue in Q2 FY 2026-2027 — per CFO Nitin Jain; management is engaging with India AI Mission to build customer relationships beyond the current contract scope, with no renewal rate or spot pricing guidance provided.
• International customers contributed roughly 20% to 25% of revenue in Q2 FY 2026-2027 — per CFO Nitin Jain, who cautioned this metric is subject to significant quarter-on-quarter variation.
• Management declined a revenue breakdown by service type — no inference/training/storage/CPU split for Q2 FY 2026-2027, citing significant quarterly volatility and calling percentage breakdowns "ephemeral".

Platform Transition, Inference Launch, and Dual-Platform Bets

• Inference products soft-released in Q2 FY 2026-2027 — Blaze, Endpoint, and Jarvis Lab, with an official launch expected in Q3 FY 2026-2027; adoption is at an exploratory stage with no volume metrics disclosed.
• 18-month technology roadmap — over the next 18 months, management's strategy is to flexibly build technology that lowers total cost of ownership (TCO) for customers, focusing on efficiency and team building.
• Dual platform strategy — Tarun Dua described Thir (TIR) and Jarvis Labs as opinionated platforms for distinct customer sets with sufficiently different engineering teams; he dismissed pricing-competition risk, citing a massive $1 trillion TAM (period unspecified) and a multi-million GPU per year marketplace (period unspecified).
• Majority of GPUs run on the company's own platform as of Q2 FY 2026-2027 — with very little bare metal being sold; the transition from raw compute to a technology platform is a medium-to-long-term strategy, with economics (realization, gross margins, ROCs) expected to materialize as customer adoption reduces TCO.
• Chip-agnostic software platform — management stated the platform aims to be compatible with any chip maker (e.g., AMD) and is willing to build customer-required support.
• TIR platform positioned as aspirational world-class (India) product with global potential — management guided that competitive success will only be clear over the medium term due to the rapid pace of AI advancement.

B200 Utilization Full, Next Tranche in Q3 FY27

• Rs.1,000 crore contract fully utilizes current B200 capacity as of Q2 FY 2026-2027 — management sees no fundamental unsurmountable supply chain challenges for GPU procurement (e.g., NVIDIA) in the near/medium term post-Q2 FY 2026-2027.
• 1,024-unit B200 GPU cluster deployment expected in Q3 FY 2026-2027 — with the next lot of B200s going live in Q3 FY 2026-2027 and revenue generation to commence subsequently.
• GPU order-to-deployment cycle typically 3 to 6 months — subject to market conditions; advance payment requirements depend on the vendor and their order book, with no fixed percentage provided.
• No timeline for Vera Rubin or B355 deployments — management indicated updates will be shared as plans develop; no decisions made on the next batch of B200 (Blackwell) capacity (period unspecified).
• Part of GPU capacity allocated to internal platform R&D in FY 2026-2027 — management is balancing long-term product development against pure capacity sales, impacting reported utilization figures; management advocated tracking overall realization across the entire fleet rather than asset-level utilization as the key KPI.

70%+ EBITDA Ambition, Open Funding Posture

• Q2 FY 2026-2027 gross margin of 88% — management characterized quarter-level margins as "moving targets" influenced by capex timing; Tarun Dua reframed the target as a 70%+ EBITDA margin business over the medium to long term (period unspecified), without confirming the analyst's cited 85% gross margin guidance.
• Enabling resolution for a Rs.1,500 crore fundraise — management acknowledged the resolution but provided no immediate timeline, committing only to keep stakeholders informed.
• No predefined upper limit for debt — management expressed comfort as long as cash flows support repayment capacity; openness to off-balance-sheet funding for GPU capex and various capital deployment strategies, without committing to a specific asset-light approach.
• L&T partnership potential flagged by analyst Varun Gandhi — he noted L&T's recent (Q2 FY 2026-2027) acquisition of 10,000 B300s; management confirmed an arm's length relationship with L&T and expressed openness to potential partnerships for utilizing idle GPU capacity, though no specific deal was announced.
• Management strengthened in Q2 FY 2026-2027 — new Chief Revenue Officer, field CTO, and strategic advisor Mr. Ananmoy Roy added to prime for the next growth phase.

Selective Disclosure; No Forward Guidance Offered

• Management declined forward MRR guidance for FY 2026-2027 — Tarun Dua advised analysts to rely on past performance instead.
• Earnings from existing GPU assets expected to grow over the remainder of FY 2026-2027 and into FY 2027-2028 — for uncontracted or short-term contracted capacity, but management declined a specific timeline or numeric forecast for when higher global GPU prices will impact MRR.
• No forward-looking capex or GPU deployment guidance for FY 2027-2028 or FY 2028-2029 — management stated it is "hard to predict the future" and the market will be informed when plans are finalized; a prudential-aggressive capex stance was reiterated, with deployment plans disclosed only once sufficiently solidified.
• Global AI infrastructure TAM framed as ~$1 trillion over the medium term — management called it a "vast blue ocean" with room for 150–250+ players, downplaying comparisons to domestic rivals (one with 90,000 GPUs, another with 8,000 P300s) and stating it is not focused on peer comparison.
• Engagements at the Nvidia AI conference in Q1–Q2 FY 2026-2027 — per Tarun Dua, these reinforce a long-term strategy of bridging AI technology and specific customer sectors via co-designed solutions, rather than generating immediate business impact.
• No new financial guidance, outlook, or risks disclosed at call close — investors directed to submit unanswered queries via Adfactors PR or investors@e2enetworks.com.