E2E Networks Limited Plans 59% AI Capacity Expansion

E2E Networks Limited plans to add 5.4 MW of NVIDIA Blackwell GPU capacity, increasing its AI factory capacity by nearly 59% from about 9.2 MW to approximately 14.6 MW. The expansion is significant because it supports larger AI training and inference workloads while the company holds a roughly ₹1,000 crore sovereign AI cloud commitment through June 2029.

Why is E2E Networks Limited adding 5.4 MW of capacity?

E2E Networks Limited is expanding its GPU infrastructure to address demand for AI computing services in India. The proposed 5.4 MW addition will support NVIDIA HGX B300 systems and other Blackwell-generation GPUs, which are used for workloads such as foundation-model training, fine-tuning, production inference and multimodal applications.

E2E Networks operates a cloud infrastructure model in which it buys and installs NVIDIA GPUs in Indian data centres and provides customers access to computing capacity. Customers can use the infrastructure on a consumption basis, bringing their own software and paying for the capacity they use. The company has also described its approach as building capacity ahead of confirmed demand to serve customers facing limited GPU availability.

The company said its existing AI factory capacity is about 9.2 MW. The planned addition would take the total to approximately 14.6 MW, representing a nearly 59% increase over the current base. E2E Networks has not disclosed the investment amount or a definitive commissioning schedule for the new capacity.

What infrastructure will the expansion support?

The proposed expansion will support NVIDIA HGX B300 systems and other NVIDIA Blackwell GPUs. These systems are intended for compute-intensive AI applications, including training, inference and deployment of AI models at scale.

E2E Networks had already deployed a 1,024-GPU NVIDIA Blackwell cluster that has been live since May 2026. The additional 5.4 MW capacity would extend the company’s Blackwell infrastructure and support its stated aim of converting India-hosted AI factory capacity into production inference services.

The company also launched Blaze, a pay-per-token AI inference service for open models, globally on November 2, 2026. Blaze runs on NVIDIA GPU infrastructure located in India. The product adds an inference-oriented service alongside E2E Networks’ underlying GPU cloud offerings.

How does the ₹1,000 crore sovereign AI agreement support demand?

E2E Networks signed a binding term sheet with an unnamed India-based sovereign AI company for NVIDIA Blackwell cloud GPUs and related services in a multi-year arrangement valued at approximately ₹1,000 crore, excluding applicable taxes. The agreement runs through June 2029.

Management said the arrangement changes recently deployed Blackwell capacity from a pay-as-you-go model to a longer-term commitment. This provides contracted demand visibility for a portion of the company’s GPU infrastructure rather than relying solely on usage-based customer consumption.

The contract was announced on August 31, 2026, according to the reported details. The customer has not been publicly named in the cited disclosure. E2E Networks did not provide a capacity allocation, annual revenue schedule, or other financial terms beyond the overall approximate contract value and validity period.

What government empanelments does E2E Networks hold?

E2E Networks was empanelled by the Ministry of Electronics and Information Technology in February 2025 under the IndiaAI Mission and was named the lowest bidder in certain AI compute and network categories. This qualifies the company to supply GPU capacity for work related to India’s foundational AI models.

The company separately secured MeitY cloud empanelment in June 2024. That approval makes E2E Networks an approved cloud provider for public-sector units, nationalised banks and government agencies, broadening its addressable government and public-sector customer base beyond the IndiaAI Mission.

These empanelments matter because government-related AI and cloud workloads can require approved domestic infrastructure providers. The reported sovereign AI agreement and the planned Blackwell expansion place E2E Networks’ capacity plans alongside its participation in public-sector cloud and AI procurement channels.

How will the expansion be financed and deployed?

E2E Networks has said the 5.4 MW expansion will be financed through a combination of internal accruals and debt. It has not disclosed a specific capital expenditure figure, debt amount, lender, commissioning timeline or phased deployment schedule for the planned infrastructure.

Deployment is expected to follow the company’s infrastructure plans and anticipated demand for AI computing services. This leaves the pace at which the entire 5.4 MW addition becomes operational unspecified, even though the company has identified the GPU platforms the capacity is intended to support.

For investors, the expansion increases the scale of E2E Networks’ infrastructure base from about 9.2 MW to a proposed 14.6 MW. The long-term sovereign AI commitment provides a disclosed demand reference, while the undisclosed project cost and rollout schedule remain important operating details not included in the announcement.

Conclusion

E2E Networks is planning a substantial increase in AI cloud infrastructure, with 5.4 MW of additional Blackwell GPU capacity set to lift its total AI factory capacity to about 14.6 MW. The plan follows a 1,024-GPU Blackwell deployment, the launch of its Blaze inference service, government empanelments and a reported ₹1,000 crore sovereign AI cloud agreement through June 2029.

The next points to watch are the company’s deployment schedule for the added capacity, its undisclosed investment requirement and the financing mix between internal accruals and debt. Investors will also track execution of the sovereign AI arrangement and further demand for GPU cloud and inference services as the capacity is deployed.