E2E Networks is India’s first and largest GPU-focused cloud provider - a company that rents out the specialised chips needed to train and run artificial intelligence, much like a power utility sells electricity to factories. Founded in 2009 by a computer-science engineer and bootstrapped through seed funding before listing in 2018, the business originally sold conventional CPU-based cloud services but made a decisive pivot in 2019, becoming the first to offer advanced NVIDIA cloud GPUs in the Indian market. Everything the company does now flows from a single, recurring bet: build massive GPU capacity ahead of confirmed demand, control the full technology stack with open-source software to keep costs down, and let customers serve themselves through a web portal - a model designed to make high-performance AI compute as easy to buy as electricity.
Business segments
A single engine - AI-focused cloud infrastructure - delivered through one self-service platform that serves startups, enterprises, and government buyers from Indian data centres.
1. GPU Cloud Infrastructure: renting the picks and shovels of the AI gold rush
The company buys thousands of NVIDIA GPUs, installs them in Indian data centres, and sells access by the hour - a pure infrastructure-as-a-service model where customers bring their own software and pay only for what they use.
- Hourly billing with no lock-in - customers provision GPU-powered virtual machines through a self-service web portal and are invoiced monthly based on hourly usage, with no hidden costs or long-term contracts, which is how the company converts trial users into recurring revenue.
- Speculative capacity, not just confirmed orders - management deliberately builds GPU infrastructure ahead of demand, betting that the chips will be absorbed because “capacity is running out everywhere”; this strategy lets the company capture customers who cannot find GPU supply elsewhere.
- Sovereign-by-design platform - the cloud runs on open-source software built in-house across all five layers of the stack, which means the company controls its own costs and can price flexibly without paying licence fees to large software vendors; it also meets Indian data-localisation requirements because all infrastructure sits in Noida, Chennai, and Mumbai data centres.
- TIR: the AI workbench that sits on top - the company’s flagship software layer gives users a single environment to develop, train, deploy, and run machine-learning models, with pre-configured setups for popular open-source models; it was the first platform in India to offer H200 GPUs and has been extended into media tools for script writing, animation, and 3D rendering.
2. Government AI contracts: becoming the national compute backbone
E2E has been selected as the lowest bidder to supply GPU hours for India’s sovereign AI mission, turning a startup-focused cloud into a supplier of critical national infrastructure.
- IndiaAI Mission empanelment - the company was empanelled by the Ministry of Electronics & Information Technology in February 2025 and named the lowest bidder in certain AI compute and network categories, which qualifies it to supply GPU capacity for building India’s foundational AI models.
- MeitY cloud empanelment opens the public-sector door - a separate empanelment secured in June 2024 makes E2E an approved cloud provider for all public-sector units, nationalised banks, and government agencies, creating a pipeline beyond the AI mission itself.
Group structure and partners
E2E Networks is a standalone public company with no operating subsidiaries, but its fortunes are now structurally tied to India’s largest engineering conglomerate through a strategic investment and a growing commercial alliance.
- L&T is both investor and landlord - in November 2024, Larsen & Toubro invested Rs. 1,079 crore for a 19% stake and gained the right to nominate two board directors; the partnership gives E2E prioritised access to L&T’s data-centre capacity, which it integrates into its cloud infrastructure, while E2E powers L&T’s digital-transformation projects with AI-optimised GPUs.
- A newly incorporated subsidiary for future expansion - in June 2026 the company set up a wholly owned subsidiary, Sovcloud Technologies, to run data-centre services and license data-centre technology, though it had no operations as of the latest filings.
- NVIDIA is the critical supplier, not a partner - the company has a long-standing vendor relationship with NVIDIA and brings the chipmaker into customer conversations, but management has clarified that no binding partnership or new agreement exists beyond routine procurement and deployment of GPU clusters.
- A lean board with deep operating and financial experience - the eight-member board includes the founder as Managing Director, two Whole-Time Directors who are chartered accountants running operations and finance, and independent directors drawn from Blume Ventures and the technology industry; the senior team adds a CTO steeped in open-source cloud computing and a Chief Revenue Officer with over two decades of B2B startup sales experience.