Media and Entertainment, Telecommunications
Rearchitecting Live Media Delivery With Verizon’s Private 5G Network With Enterprise AI
Objective
The broadcast industry is undergoing major change, driven by growing demand for high-quality live content, AI-powered workflows, and fast, reliable connectivity. Traditional systems often struggle with wireless reliability, data security, workflow latency, and deploying computational resources on location.
To solve these challenges, Verizon introduced its Private 5G Network with Enterprise AI. Built with NVIDIA AI Enterprise and Holoscan for Media, the solution delivers reliable wireless performance, manages high data volumes smoothly, and reduces workflow delays—giving broadcasters a portable, efficient, and cost-effective way to improve quality and expand creative possibilities.
Customer
FanDuel TV
Verizon
Partner
Haivision
Monks
Ericsson
Dell Technologies
Key Takeaways
Dramatic Latency Reduction and On-Site Control
- FanDuel’s pilot using Verizon’s Private 5G Network with Enterprise AI running on NVIDIA AI Enterprise and Holoscan for Media achieved over 50% lower latency and enabled immediate editorial control near the source for live race production.
AI-Driven Live Content Optimization
- The solution enabled real-time AI inference to dynamically prioritize camera feeds for FanDuel, identify jockeys by jersey color, and automate feed adjustments, leading to faster editorial decisions, improved picture quality, and streamlined workflows.
Portable and Scalable AI Infrastructure
- Verizon’s Private 5G with Enterprise AI offers a smaller, portable hardware footprint and rapid scalability for FanDuel’s remote production sites. This infrastructure supports on-site training and deployment of custom AI models (LLMs/VLMs), ensuring data security and lowering deployment costs.
Breaking Through the Limits of Traditional Broadcasting
Today’s broadcasters face a range of challenges that impact efficiency, drive up costs, and affect the quality of live productions.
Traditional wireless systems often struggle in crowded environments like live events, where RF congestion leads to dropped signals and poor video quality. This impacts not just camera feeds, but also wireless monitors, lighting controls, and crew communications. Long-distance wireless links between teams and equipment are also limited by outdated infrastructure.
Live productions now rely on dozens—or even hundreds—of camera feeds. Processing all that video requires expensive on-site infrastructure, like production trucks packed with servers. Existing wireless and edge solutions can’t scale easily and often fall short on bandwidth, while wired setups limit mobility and camera placement. Production teams also face the creative challenge of monitoring multiple feeds in real time to pick the best shots.
Cloud platforms offer scalability but struggle with two major issues in live production: latency and security. Transmitting real-time data to and from distant cloud environments can introduce delays and security concerns—especially for IP-based workflows like SMPTE ST 2110. Deploying compute-intensive tools like AI on location often requires complex, costly connections back to remote data centers.
As demand for remote production, AI tools, and flexible IP workflows grows, it’s clear that traditional systems can’t keep up. Broadcasters need a faster, smarter, and more reliable solution built for the future.
Verizon’s Private 5G Network With Enterprise AI
To tackle the challenges of modern broadcasting, Verizon developed the Private 5G Network with Enterprise AI—a portable, on-site solution designed specifically for the media and entertainment industry. It brings the power of data center-grade computing and private cellular connectivity directly to production environments.
There are two core components that power the solution:
Verizon Private 5G Network
At the foundation is a dedicated private cellular network, delivering fast, reliable, and secure wireless connectivity. With Ericsson RAN Compute processors, it uses a mix of mmWave, C-band, and CBRS radios to support high-performance wireless communications. The network is vendor-agnostic and configurable, with the ability to tune uplink/downlink speeds to match production needs—ideal for real-time video, monitors, lighting, and comms.
Enterprise AI Compute Module
On the compute side, Verizon’s AI module brings advanced processing power to the edge. Built on NVIDIA Holoscan for Media, it enables real-time AI inference for video and audio and supports multi-vendor workflows via container orchestration. The platform is fully IP-based and compliant with broadcast standards like SMPTE ST 2110, and includes timing (PTP), control (NMOS), and NVIDIA AI Enterprise tools like NIM™ microservices. It runs on bare-metal infrastructure leveraging NVIDIA L40S GPUs for high-performance compute, networking, and storage.
Together, these components enable broadcasters to manage and process high volumes of camera feeds directly on location—without relying on the cloud. This significantly reduces latency and simplifies setups by cutting down on bulky gear and external dependencies.
In addition to improving wireless reliability and efficiency, the system supports on-site training and deployment of custom AI models—including large language models (LLMs) and vision language models (VLMs)—for intelligent automation, all while keeping sensitive data secure and lowering deployment costs.
Implementation
Setting up Verizon’s portable Private 5G with Enterprise AI starts with a site survey to assess network conditions. One major advantage is Verizon’s use of licensed spectrum, which boosts wireless capacity and allows uplink/downlink tuning based on production needs.
Because the setup is environmentally controlled and software-defined, it offers greater reliability and flexibility than traditional systems—while reducing on-site hardware and complexity. It easily supports scalable, AI-integrated workflows with fewer physical resources.
Deployment is typically faster than traditional infrastructure, making it ideal for high-volume or remote productions with minimal existing setup. In short, it delivers connectivity, compute power, and app control—wherever production happens.
Real-World Performance: FanDuel Deployment
FanDuel TV operates one of the most demanding live sports broadcast environments in the industry, streaming over 70,000 live horse races annually across three linear television networks. These broadcasts originate from more than 130 racetracks nationwide and are supported by a hybrid infrastructure strategy that includes remote production kits, production truck deployments, and third-party partnerships.
In a pilot scenario using Verizon’s Private 5G Network with Enterprise AI, FanDuel simulated a live race production. Multiple video camera feeds were ingested wirelessly over a Private 5G Network and processed at the edge using NVIDIA Holoscan for Media. AI inference models ran in real time on the video streams to dynamically prioritize camera feeds. In the future, FanDuel will extract metadata, such as horse positioning and jersey recognition, and dynamically generate a live, on-screen “Run Order” graphic.
The results were significant:
- Over 50% lower latency versus traditional cloud workflows
- Immediate editorial control near the source
- AI-driven automation for smarter, faster production decisions
- AI-driven feedback for dramatically improved picture quality on key sources
- Improved reliability with dedicated private network connectivity
- Smaller hardware footprint, condensed from multiple racks to one portable unit
- Rapid scalability for remote or temporary production sites
This pilot shows how Private 5G and edge AI can transform high-volume live sports production. FanDuel now has a proven path to expand real-time, AI-assisted workflows across its entire racing operation.
Partnership Ecosystem: Amplifying Capabilities
Verizon partnered with leading technology and media companies to extend the functionality of its Private 5G Network with Enterprise AI and showcase its real-world impact:
Haivision
Enabled seamless camera connectivity and live video distribution over Private 5G. By integrating StreamHub as software on Verizon’s platform, customers gain flexible access to encoding/decoding and can apply NVIDIA AI for tasks like live captioning and object detection.
Monks
A marketing and technology services partner, its LiveVision solution was used to deploy a general-purpose AI model for real-time object detection on the racing footage—without custom training—demonstrating quick, adaptable use of AI tools.
Ericsson
Supplied the RAN Compute processors and radio infrastructure for Verizon’s Private 5G Cellular Core, ensuring high-performance, reliable connectivity.
Dell Technologies
Provided Dell PowerEdge R660 servers and Dell PowerEdge R760 servers equipped with Sapphire Rapids Chipsets, NVIDIA L40S GPUs, and NVIDIA® ConnectX®-7 network interface cards.
These partners collaboratively demonstrated intelligent, AI-assisted video prioritization in a live horse race scenario. Using the Mistral 12B model for object detection, the AI identified specific jockeys by jersey color. When a designated jockey appeared, automated API calls increased that camera feed’s bitrate and highlighted it in Haivision’s StreamHub multi-viewer.
This demo showcased how NVIDIA Holoscan for Media, running on Verizon’s platform, powers:
- Smart bandwidth allocation
- Faster editorial decisions
- Streamlined live production workflows
Verizon
These partnerships underscore the open, platform-based architecture of Verizon’s solution. Third-party developers and content producers can easily deploy their own applications on top of the Private 5G + AI stack.
In collaboration with NVIDIA, the ecosystem is built to support:
- Flexible, developer-ready infrastructure
- Faster innovation in AI-powered content creation and remote production
- Reduced time to market for broadcasting tools
- Broader access to a growing network of media technology partners
Together, these collaborations are driving the future of live media—where wireless flexibility, edge AI, and open innovation converge.
Future Evolution and Potential
Built for the future, Verizon’s platform can easily expand to support emerging formats like extended reality (XR) and volumetric video, thanks to its GPU-accelerated, edge-powered infrastructure. This opens doors not only in media but also in industries like healthcare, live events, and manufacturing, where high-speed video processing and real-time AI are increasingly critical.
As 5G and AI continue to evolve, the system could take advantage of autonomous network capabilities—automatically adjusting compute and bandwidth resources to match production needs. This would drive even greater efficiency and control. Over time, the solution can also reshape how and where content is created by enabling fully remote and hybrid productions—reducing the need for centralized facilities and broadening access to talent and resources globally.
Learn more about NVIDIA technologies for live media.