Frontier AI lab showcases how hyperlocal network infrastructure can power a new generation of real-time AI experiences

SAN FRANCISCO, CA / ACCESS Newswire / March 18, 2026 / Decart is collaborating with Comcast and NVIDIA to demonstrate how NVIDIA's AI Grid architecture, where GPUs deployed across distributed facilities, could transform the way AI-powered experiences are delivered using Comcast's network architecture. The work offers a look into how advanced AI systems may eventually be delivered closer to users and devices. As part of the collaboration, Decart is deploying its Lucy real-time model across NVIDIA-powered infrastructure within Comcast's nationwide, deeply distributed architecture to explore how these applications perform in a geographically distributed environment.

The joint work highlights how interactive AI workloads, such as real-time generative video and responsive digital environments, can operate with the low latency required for live, user-driven experiences. Through benchmarking conducted with Comcast, Decart demonstrated how its Lucy real-time model can operate on NVIDIA-powered edge infrastructure with sub-35 millisecond end-to-end latency. For Decart, the collaboration offers a window into how real-time AI systems might be experienced in the future.

"World models are going to be one of the defining infrastructures of the next decade of AI," said Dean Leitersdorf, CEO and co-founder of Decart. "What makes these systems powerful is their ability to run in real time. Instead of generating static outputs, they create environments that respond instantly to user input and evolve as people interact with them. That kind of responsiveness opens the door to entirely new experiences across gaming, interactive media, robotics, and physical AI."

Decart is a vertically integrated frontier AI research lab focused on building real-time world models, AI systems designed to simulate environments rather than simply generate content. Unlike traditional video generation models that produce frames independently, world models create environments that maintain spatial continuity, react to user interaction, and evolve according to internal rules.

As AI-native applications grow more complex and interactive, researchers and infrastructure providers are exploring how these workloads can be delivered efficiently to users around the world. Telecommunications networks already operate some of the largest distributed infrastructures in the world, with tens of thousands of network facilities globally spanning regional hubs, mobile switching offices, and central offices.

"Real-time AI experiences require both powerful models and infrastructure capable of delivering inference close to users, especially when it comes to scaling thousands of workloads concurrently," said Chris Penrose, global vice president, Telecom, NVIDIA. "Running Decart's Lucy model on NVIDIA accelerated computing, over a distributed edge AI grid, enables a new class of interactive AI applications that respond instantly and operate at global scale."

With AI grid, NVIDIA aims to leverage highly distributed network infrastructure, like Comcast, as a distributed computing layer, allowing AI inference to run closer to users, devices and data. By moving computation to the network edge rather than relying solely on centralized cloud data centers, these systems can help reduce latency, minimize network congestion and support real-time AI workloads at scale.

About Decart
Decart is a fully vertically integrated frontier AI research lab, building SOTA real-time world models. Backed by Sequoia and Benchmark and valued at $3.1B, Decart enables real-time generation, transformation, and simulation for use cases across entertainment, advertising, robotics, gaming, and enterprise AI. Its models power millions of users and represent a step-change in speed, cost, and scalability for AI video.

Media Contact:
Mia Balaban
[email protected]

SOURCE: Decart



View the original
on ACCESS Newswire


Information contained on this page is provided by an independent third-party content provider. XPRMedia and this Site make no warranties or representations in connection therewith. If you are affiliated with this page and would like it removed please contact [email protected]