E2E Networks, the NSE- and BSE-listed Indian AI cloud provider, on 2 November 2026 launched Blaze, a pay-per-token inference service that gives developers and enterprises a single OpenAI-compatible API to open-weight models such as NVIDIA Nemotron, Sarvam, DeepSeek, Qwen, GLM and Gemma, all served from NVIDIA GPU infrastructure hosted in India.

The company disclosed the launch to the stock exchanges under Regulation 30 of SEBI’s listing rules, listing Blaze for both domestic and international customers. It comes a week after E2E said it would add 5.4 MW of NVIDIA Blackwell capacity to its AI cloud, and turns that hardware into a metered service: customers pay only for the tokens they consume, while E2E handles model hosting and scaling.

What is E2E Networks Blaze?

Blaze is a managed inference API. Instead of renting GPUs and running their own serving stack, teams send requests to one endpoint and choose from a catalogue of open models covering text, voice, OCR, multimodal and image generation. According to the product page, the service runs in Indian data centres and is billed in rupees.

• Model choice: NVIDIA Nemotron, Sarvam, DeepSeek, Qwen, GLM and Gemma through one API.
• Target workloads: coding agents, long-context jobs, voice and chat, document AI and OCR, image and video generation, and Indian language applications.
• Pricing model: pay per token, with no reserved GPU capacity required.

The NVIDIA stack underneath

E2E said Blaze is built on open-source components of NVIDIA DSX OS, the software layer of NVIDIA’s AI factory platform. NVIDIA Dynamo handles distributed serving with KV cache-aware routing and separate prefill and decode stages, which matters for agents that resend long prompts repeatedly. NVIDIA Grove and the KAI Scheduler coordinate jobs across E2E’s Kubernetes fleet, and BlueField DPUs with DOCA enforce hardware-level isolation between customers.

That isolation is aimed squarely at banks, healthcare providers and public-sector bodies that face strict data rules. “Blaze is where you can run your production inference,” said Tarun Dua, Founder and Managing Director of E2E Networks, adding that customers want frontier-class open-weight models running on sovereign compute in India. Vishal Dhupar, Managing Director for South Asia at NVIDIA, said access to AI computing and inference services is essential to advancing India’s AI ecosystem.

From training to tokens on one platform

Blaze sits alongside TIR, E2E’s existing AI platform for pre-training, fine-tuning and self-hosted inference on NVIDIA Blackwell clusters. Customers who want control can serve their own or fine-tuned models on TIR, while those who want simplicity can call Blaze. E2E said its AI factory capacity is expanding to about 14.6 MW, including a 1,024-GPU Blackwell cluster live since May 2026.

Why it matters for Indian CIOs

As AI agents move from pilots into production, token bills, latency and data residency are becoming the real constraints, not model access. An India-hosted, rupee-billed API for open models gives regulated enterprises a middle path between global hyperscaler APIs and running their own GPUs. It also lands as the government works to widen compute access through the IndiaAI Mission’s 10,000-GPU tender, and as startups such as Simplismart work to run AI models on fewer GPUs. Analysts have also warned that India’s new capacity must be used well, with one report finding GPU utilisation in India near 20%; metered inference services are one way to keep that hardware busy.

Key takeaways

• E2E Networks launched Blaze, a pay-per-token, OpenAI-compatible inference API for open models, on 2 November 2026.
• Models run on NVIDIA GPUs in Indian data centres, with hardware isolation aimed at banks, hospitals and government users.
• Blaze pairs with E2E’s TIR platform, giving one India-hosted path from training and fine-tuning to production inference.

FAQ

What models are available on E2E Networks Blaze?

E2E says Blaze offers open-weight models including NVIDIA Nemotron, Sarvam, DeepSeek, Qwen, GLM and Gemma through a single API.

How is Blaze priced?

Blaze is pay-per-token: customers are billed only for the tokens they use, in rupees, with E2E managing hosting and scaling.

Source: E2E Networks press release and stock exchange filing, 2 November 2026; E2E Networks Blaze product page.

Featured image: Tarun Dua, Founder and Managing Director, E2E Networks. Photo: E2E Networks

Be a part of Elets Collaborative Initiatives. Join Us for Upcoming Events and explore business opportunities. Like us on Facebook , connect with us on LinkedIn and follow us on Twitter.

"Exciting news! Elets technomedia is now on WhatsApp Channels Subscribe today by clicking the link and stay updated with the latest insights!" Click here!