E2E Networks Limited has released its Investor Presentation for Q3FY26, providing an update on its operational performance and strategic initiatives. The company is significantly expanding its GPU capacity, with orders for approximately 1,024 Blackwell B200 units, totaling 184TB of GPU RAM. The first batch of 1,024 B200 GPUs was procured at the Chennai data center on January 9, 2026. The company anticipates its total GPU capacity to exceed 5,000 units by March 2026.

Strategic developments include a new contract with L&T for one-year GPU services valued at ₹8.49 crore, commencing January 2026. This partnership aims to de-risk and reduce volatility through long-cycle enterprise deals, positioning E2E Networks as an enterprise AI infrastructure provider via the L&T channel. Additionally, the company has secured IndiaAI Mission contracts totaling ₹265 crore for its AI/ML platform, with phased deployment initiated in December 2025. The acquisition of Jarvis Labs was completed in December 2025, positioning the company for rapid global scaling.

Financially, for Q3FY26, E2E Networks reported operational revenue of ₹700 million, a 68.3% year-on-year increase and a 59.8% quarter-on-quarter increase. EBITDA stood at ₹396 million, up 60.9% year-on-year and 120.2% quarter-on-quarter, with an EBITDA margin of 56.6%. However, the company reported a Profit After Tax (PAT) of ₹(57) million, a decrease of 149.2% year-on-year, resulting in a PAT margin of (8.1)%. This was accompanied by a Diluted EPS of ₹(2.80).

The company has also raised substantial funds through preferential equity share issues, with ₹14,849.34 million raised in total, of which ₹2,168.20 million was utilized in FY25 and ₹10,280.75 million in 9MFY26, leaving a balance of ₹2,400.39 million. The balance sheet as of September 25 shows total assets of ₹17,785 million and total equity and liabilities of ₹17,785 million. The company emphasizes its purpose-built platform for AI/ML workloads, sovereign-by-design infrastructure, demonstrated financial growth, and a cost-efficient computing model.