In the quarter that ended in June 2026, E2E Networks turned a net profit of ₹43.88 crore, against a net loss of ₹2.84 crore in the same quarter a year earlier. Revenue nearly quadrupled, up 334% year-on-year to ₹156.76 crore (Free Press Journal; Whalesbook, July 2026).
That turnaround sits awkwardly next to the year just before it. For the twelve months to March 2026, revenue rose 50% to ₹246 crore, and the company still lost ₹16 crore (Screener.in). Same GPUs, same customers, opposite outcomes four quarters apart. The gap between the two is the whole story of what it costs an Indian cloud company to bet on the AI-compute boom before the demand arrives to pay for it.
Quick facts
| Company | E2E Networks Limited |
| Founded | August 2009, New Delhi |
| Founders | Tarun Dua (Managing Director), Mohamed Imran K R (Co-founder & CTO), Srishti Baweja (Co-founder) |
| Businesses | Cloud GPU compute, TIR AI/ML platform, AI Labs-as-a-Service, Sovereign Cloud Platform |
| Latest FY revenue | ₹246 crore (FY26, year to March 2026) |
| Latest FY profit/loss | Net loss of ₹16 crore (FY26) |
| Listed | NSE Emerge (7 May 2018) → NSE Mainboard (2022) → BSE Mainboard (12 June 2026) |
| Market value | Approx. ₹12,500–13,100 crore ($1.3–1.4 billion at $1≈₹96.0) as of September 2026 |
| Key shareholders | Promoters approx. 39.45% (June 2026); Larsen & Toubro approx. 21% (since 2024–25) |
What they do
E2E Networks sells cloud computing by the hour, built around Nvidia GPUs rather than the general-purpose servers most Indian cloud providers started with. Its customers rent H100, H200, A100 and now B200 GPU capacity to train and run AI models, alongside a software layer called TIR for AI/ML and generative AI development, an “AI Labs-as-a-Service” offering for teams that want a managed environment rather than raw machines, and a Sovereign Cloud Platform aimed at regulated and government workloads that must stay on Indian soil. The customer base runs from startups and small businesses, which the company says now number more than 10,000, up to large enterprises and, increasingly, government AI programmes (e2enetworks.com).
The origin
Tarun Dua, a computer science graduate of the National Institute of Technology, Kurukshetra, had worked at Yahoo, Nanocast R&D, GlobalLogic and ValueFirst before he founded E2E Networks in August 2009 with Mohamed Imran K R, an engineer with a master’s degree from Anna University who became the company’s co-founder and chief technology officer, and Srishti Baweja, a chartered accountant and Shri Ram College of Commerce graduate who handled finance, legal and compliance (e2enetworks.com; enrichmoney.in). The insight was narrow but real: Indian businesses were paying for rigid, contract-locked hosting at a time when domestic internet latency was falling and cheaper, SSD-based, pay-as-you-go compute had become technically possible. E2E built exactly that, positioning itself as a leaner, contract-less alternative to both the legacy Indian hosting industry and the international public clouds (enrichmoney.in).
The struggle years
The idea did not attract capital easily. According to an Inc42 profile of the company, most venture investors who looked at E2E Networks between 2009 and 2011 turned it down as either too small or not differentiated enough, and the founders ran the business from their own homes for close to a year to conserve cash. Blume Ventures became the only institutional investor willing to write a cheque, putting in a seed round of about $51,975 (₹35 lakh) in early 2011, followed by further small rounds in 2013 and 2014 that took total funding to roughly $423,225 (about ₹2.85 crore) by 2014 (Inc42; Crunchbase). For a company now valued in the thousands of crores, that is a strikingly small amount of outside capital to have survived on for nine years before its 2018 listing.
The more recent struggle looks different but comes from the same place: too much conviction, too early. Revenue fell 12.6% quarter-on-quarter in the December 2024 quarter (Q3 FY25) as AI-training demand proved “bursty” rather than steady, according to an analyst note on the results (Univest.in). Then, for the full year to March 2026, E2E’s revenue rose 50% to ₹246 crore, yet the company posted a net loss of ₹16 crore — a reversal from four straight years of rising profit (Screener.in). The cause was not weak demand; it was the accounting weight of GPU racks bought and depreciated before enough customers had signed on to use them.
The turning point
The clearest before-and-after in E2E’s history sits across two June quarters. In the quarter ended 30 June 2025 (Q1 FY26), revenue from operations was ₹36.11 crore, the company posted a net loss of ₹2.84 crore, and basic loss per share came to ₹0.14 (Free Press Journal, July 2026). A year later, in the quarter ended 30 June 2026 (Q1 FY27), revenue from operations reached ₹156.76 crore — up 334% year-on-year, and up 63.9% on the preceding March 2026 quarter’s ₹95.64 crore — while net profit came in at ₹43.88 crore and earnings per share flipped to ₹2.14 (Free Press Journal; Whalesbook, July 2026).
What changed in between was not the technology; it was the order book catching up to the capacity. GPU clusters bought through 2025, including a 1,024-unit Nvidia B200 cluster deployed at Larsen & Toubro’s Vyoma data centre in Chennai, began filling with paying workloads from enterprise and government contracts, so each additional compute-hour billed started dropping through to profit rather than being absorbed by fixed depreciation (DataCenterDynamics, 2026).
The money behind it
- Seed capital: Blume Ventures invested about $51,975 (₹35 lakh) in early 2011, the company’s only institutional backer for years, followed by further small rounds that took total pre-IPO funding to roughly $423,225 (about ₹2.85 crore) by 2014 (Inc42; Crunchbase).
- Exit: Blume Ventures sold its stake in 2023, after the company’s 2018 listing, in a deal reported to have returned about 20 times its investment (Blume Ventures commentary).
- First public capital event: initial listing on NSE Emerge, the SME platform, on 7 May 2018, oversubscribed roughly 5 times (Inc42).
- Migration up the ladder: shifted from NSE Emerge to the NSE Mainboard in 2022, then listed directly on the BSE Mainboard effective 12 June 2026, with about 20.56 crore equity shares approved for listing (Inc42; Business Upturn).
- Strategic anchor investor: Larsen & Toubro agreed in November 2024 to acquire a 21% stake for about ₹1,407 crore (roughly $146 million at $1≈₹96.0) — ₹1,079 crore via preferential allotment for 15%, and a further ₹328 crore buying 6% from promoters, to complete by 30 May 2025 (Business Standard; Entrepreneur India; Angel One).
- Current ownership: promoters held about 39.45% as of June 2026, with Larsen & Toubro holding roughly a fifth of the register (Screener.in).
- Latest “valuation” for a listed company is simply its market capitalisation: approximately ₹12,500–13,100 crore in September 2026, varying by the day the figure is pulled (stockanalysis.com; Screener.in).
How it makes money
E2E does not resell someone else’s servers; it owns Nvidia GPUs and data-centre capacity and sells access to them, either by the hour or under committed contracts.
- Pay-as-you-go compute: hourly-billed cloud GPUs, publicly listed from about $1.80 an hour for an H100 up to about $4.90 an hour for a B200, with A100 80GB priced around ₹189 an hour (about $2.10) (e2enetworks.com pricing pages).
- Platform revenue: TIR, its AI/ML and generative AI development platform, and AI Labs-as-a-Service for teams that want a managed model environment rather than raw compute (company description via Screener.in).
- Sovereign Cloud Platform: compute sold specifically to regulated and government workloads that must run on Indian infrastructure (Screener.in).
- Large committed contracts: an ₹8.49 crore GPU cloud services order routed through Larsen & Toubro, and a ₹177 crore GPU order from the Ministry of Electronics and Information Technology’s IndiaAI Mission, won in September 2025 (Angel One; Business Standard).
- Cost side: dominated by GPU procurement and its depreciation, plus data-centre power, cooling and colocation, now spanning Delhi-NCR/Noida and, since the Larsen & Toubro tie-up, the Vyoma facility in Chennai (DataCenterDynamics).
- What people get wrong: rising revenue does not automatically mean rising profit here, because GPUs are depreciated on a fixed schedule from the day they are racked, whether or not a customer is renting them yet. That is why revenue rose 50% in FY26 while the company swung to a ₹16 crore net loss (Screener.in), and why the very next quarter flipped straight back to a ₹43.88 crore profit once utilisation caught up (Free Press Journal).
The numbers
Annual revenue and profit, in ₹ crore, for the four most recent financial years (Screener.in, based on reported annual results):
| Financial year | Revenue (₹ crore) | Net profit/(loss) (₹ crore) |
| FY23 | 66 | 10 |
| FY24 | 94 | 22 |
| FY25 | 164 | 47 |
| FY26 | 246 | (16) |
Revenue has compounded at roughly 47% a year over the five years to FY26, from a ₹52 crore base in FY22, even though profit reversed into a loss in the most recent year (Screener.in). The quarterly data underneath that annual loss is more volatile still: after four consecutive loss-making quarters running from mid-2025 into early 2026, the June 2026 quarter alone produced a ₹43.88 crore profit (Trendlyne; Free Press Journal).
Where the money comes from
E2E does not publish an audited revenue split by GPU, CPU, storage or platform; it reports itself as operating in a single segment, “cloud computing services” (Univest.in; Whalesbook). What can be pieced together instead is a geography and customer picture.
- Geography: effectively 100% India, with data centres in Delhi-NCR/Noida and, since 2026, the Larsen & Toubro Vyoma facility in Chennai (DataCenterDynamics).
- Startup and SMB base: the original customer segment, which the company describes as more than 10,000 businesses (e2enetworks.com, company-stated).
- Consumer-internet enterprises: named customers include Zomato, CarDekho, 1mg, Junglee Games and Purplle (e2enetworks.com, company-stated).
- Strategic and industrial: Larsen & Toubro, which is both a 21% shareholder and a customer that routed an ₹8.49 crore GPU services order to E2E (Business Standard; Angel One).
- Government: the Ministry of Electronics and Information Technology’s IndiaAI Mission, which placed a ₹177 crore GPU order in September 2025 (Business Standard; Angel One).
- The surprise: a company built to serve startups now has its quarterly numbers moved by an engineering conglomerate’s data-centre order and a government AI mission — customer categories that did not exist in its revenue mix a few years ago, and whose individual contribution the company still does not disclose.
The risks
- Capital intensity and depreciation: large GPU purchases are depreciated from day one regardless of utilisation, which is the direct mechanism behind the FY26 net loss of ₹16 crore despite 50% revenue growth (Screener.in; Whalesbook).
- Capacity utilisation: an analyst note put E2E’s GPU utilisation at only around 50–60% in 2026, meaning a meaningful share of expensive, depreciating hardware sits idle at any given time (Univest.in).
- Nvidia dependency: E2E’s entire AI infrastructure runs on Nvidia GPUs, so supply shortages, delivery delays, export restrictions or pricing changes at the chip level flow straight into its capacity plans and costs (Whalesbook).
- Contract lumpiness: revenue has already shown how much large, irregular orders can swing results — the 12.6% quarter-on-quarter revenue decline in Q3 FY25 was attributed to the “bursty” nature of AI-training workloads rather than steady, contracted demand (Univest.in).
- Competition: E2E competes both against global hyperscalers and against other domestic entrants racing into the same GPU-cloud opportunity, a dynamic flagged as a specific risk in coverage of its results (Whalesbook).
The takeaway
E2E Networks’ arc is not really a story about GPUs. It is a story about the lag between spending on capacity and being paid for it, and about how easily that lag gets mistaken for failure. A company that had been quietly profitable for more than a decade chose, deliberately, to post a loss in FY26 by racing to build capacity ahead of confirmed demand, then watched a single quarter’s utilisation catch-up turn that loss into one of its best quarters on record. The transferable lesson sits in that sequence, not in the AI theme around it: when a business is capital-intensive and demand is genuinely uncertain, the income statement will look wrong for a while even if the underlying bet is right, and the only way to tell the difference from the outside is to watch utilisation and order books, not just the profit line.
Frequently asked questions
What does E2E Networks do?
E2E Networks rents out cloud computing built on Nvidia GPUs, billed by the hour or under committed contracts, alongside an AI/ML development platform called TIR, a managed “AI Labs-as-a-Service” offering, and a Sovereign Cloud Platform for regulated and government workloads (e2enetworks.com; Screener.in).
When was E2E Networks founded and by whom?
It was founded in August 2009 in New Delhi by Tarun Dua, Mohamed Imran K R and Srishti Baweja (e2enetworks.com; enrichmoney.in).
Is E2E Networks profitable?
Not consistently in its most recent full year: it posted a net loss of ₹16 crore in FY26 (year to March 2026) despite revenue growing 50% to ₹246 crore, driven by depreciation on new GPU capacity. It returned to profit sharply in the quarter ended June 2026, reporting a net profit of ₹43.88 crore (Screener.in; Free Press Journal).
What is E2E Networks’ market capitalisation?
Approximately ₹12,500–13,100 crore (about $1.3–1.4 billion at $1≈₹96.0) as of September 2026, varying by the specific day the figure is measured (stockanalysis.com; Screener.in).
What is Larsen & Toubro’s relationship with E2E Networks?
Larsen & Toubro agreed in November 2024 to acquire a 21% stake in E2E Networks for about ₹1,407 crore, split between a preferential allotment and a purchase of shares from promoters, and separately placed an ₹8.49 crore GPU cloud services order with the company (Business Standard; Entrepreneur India; Angel One).
Sources
Figures are as of September 2026. Currency converted at $1 ≈ ₹96.0 as of 18 September 2026 (Trading Economics).
- Inc42, “E2E Networks: The Story Of An Underdog With A Rockstar IPO Debut”, accessed September 2026
- Inc42, “AI Cloud Service Provider E2E Networks Debuts On BSE Mainboard”, June 2026
- enrichmoney.in, “E2E Networks Ltd. – History, Latest Updates, Milestones, Subsidiaries & Share Price”, accessed September 2026
- Screener.in, E2E Networks Ltd company page and financial statements, accessed September 2026
- Trendlyne, “E2E Networks – Quarterly Results and Financial Statement as of Jun, 2026”, accessed September 2026
- Free Press Journal, “E2E Networks Net Profit Jumps To ₹43.88 Crore In Q1 FY27, Revenue Up 334%”, July 2026
- Whalesbook, “E2E Networks Posts Strong Q1 FY27 Results; Revenue Surges 63.9%”, July 2026
- Whalesbook, “E2E Networks: ‘Add’ Rating Amid AI Boom, Risks Loom”, 2026
- Univest.in, “E2E Networks Q4 FY26 Results: PAT Falls 51% to ₹8.56 Crore”, 2026
- Business Standard, “L&T to acquire 21% stake in E2E Networks in two-part Rs 1,406 crore deal”, 5 November 2024
- Entrepreneur India, “L&T to Acquire 21% Stake in E2E Networks for INR 1,407 Cr”, November 2024
- Angel One, “L&T Acquires 15% Stake in E2E Networks via Preferential Allotment”, 2024–2025
- Angel One, “E2E Networks Receives ₹8.49 Crore GPU Cloud Services Order Via L&T”, 2026
- Business Standard, “E2E Networks locks in 10% upper circuit on bagging deal worth Rs 177 crore”, 3 September 2025
- Angel One, “E2E Networks Share Price hits 10% Upper Circuit on ₹177 Crore Order from MeitY’s IndiaAI Mission”, September 2025
- DataCenterDynamics, “E2E Networks acquires 1,024 Nvidia B200 GPU cluster for Chennai data center”, 2026
- Crunchbase, “Seed Round – E2E Networks” funding profile, accessed September 2026
- stockanalysis.com, “E2E Networks (NSE:E2E) Market Cap & Net Worth”, September 2026
- e2enetworks.com, company, team and pricing pages, accessed September 2026
Found an error? Write to us and we’ll correct it in the open, dated, on the piece.
