Glance accelerates Visual Gen AI stack with NVIDIA infrastructure

Bangalore: Glance, a global consumer technology company pioneering AI-led agentic commerce, today announced it is accelerating the next phase of its Visual Generative AI stack with NVIDIA AI infrastructure. The collaboration enables real-time, high-quality image and video generation at scale and efficiency levels designed to make generative AI accessible to millions of consumers.

Glance has positioned itself at the forefront of applying deep learning to visual and art-driven use cases that require both performance and efficiency at scale. Unlike most Visual Gen AI platforms that charge users per creation, Glance’s AI engine is built on a commerce-led model where generative creativity is freely accessible to consumers, with value unlocked through downstream commerce. Delivering on this model requires a fully integrated, end-to-end Visual Generative AI stack engineered for speed, operational efficiency and cost control.

By leveraging NVIDIA RTX PRO™ 6000 Blackwell Server Edition along with NVIDIA TensorRT and NVFP4 inference, Glance’s agentic shopping platform aims to significantly enhance inference speed and efficiency for visual Gen AI workloads. The optimised stack supports multiple production-grade workflows, including personalised image generation, text-based image editing and text-to-video generation.

Key use cases include transforming user selfies into high-quality avatars for instant exploration of styles and identities; enabling seamless visual edits using natural language prompts; and powering short-form video creation through advanced generative video models.

According to Glance’s internal testing, these optimisations have enabled up to 20× faster image generation and up to 16× faster video generation, while significantly lowering per-unit compute costs.

“Our collaboration with NVIDIA is core to how we are building Glance’s agentic shopping platform,” said Arvind Jayaprakash, Senior Vice President Of Technology, Glance. “To make generative AI truly consumer-grade, it must be fast, visually compelling, and economically sustainable. Building our own Visual Gen AI stack and optimizing it on NVIDIA’s accelerated computing platform allows us to operate complex visual models efficiently at scale — delivering real-time consumer-grade image and video generation with the performance and cost profile required for global adoption.”

“As Visual Gen AI moves from experimentation to production at scale, performance and efficiency becomes critical,” said Vishal Dhupar, Managing Director, Asia South, NVIDIA. “Glance’s agentic shopping platform demonstrates how accelerated computing can meet those demands by delivering real-time image and video generation.”