E2E Networks Introduces Blaze, Making AI Model Access Pay-Per-Token

Indian cloud and AI infrastructure company E2E Networks has launched Blaze, a pay-per-token AI inference service that allows developers and enterprises to access multiple open AI models through a single OpenAI-compatible API. Founded by Tarun Dua, the company has built the service on NVIDIA GPU infrastructure hosted in India, supporting models including NVIDIA Nemotron, Sarvam, DeepSeek, Qwen, GLM and Gemma. Under its consumption-based pricing model, customers pay only for the tokens they use, while E2E Networks manages model hosting and scaling.

Built using components from NVIDIA DSX OS, Blaze is designed to support applications such as AI coding agents, voice services, document processing, OCR, and image and video generation. The launch adds a managed inference offering to E2E Networks' existing TIR AI platform, which supports model training, fine-tuning and self-hosted inference. Through these offerings, the company aims to provide developers and businesses with infrastructure to move AI applications from experimentation into production.

#ArtificialIntelligence #AIInfrastructure #IndianStartups #Startup77