- An AI-enabled, computer vision-centric private network is now in Vegas
- Applications will include public safety and traffic automation
- The AI will be able to process data from consumer mobile phones, as well as cameras around the city
Improved public safety and better traffic automation will be two of the key applications enabled by the new 5G private network rolled out by the city of Las Vegas just ahead of the annual Mobile World Congress (MWC) event this week.
Vapor IO – using Nvidia hardware – has built a 5G private network in Las Vegas that utilizes artificial intelligence (AI) to deliver computer vision-based camera services to the city. The network is also combined with a large language model (LLM) for AI inference, which allows the private network to draw conclusions from new data that the LLM hasn’t seen before.
The Vapor IO CEO Cole Crawford told Fierce on a call that the deployment is a software programmable network that can run “next-gen AI centric workloads.” Much of the data used by the network will come from computer vision cameras at intersections around the city, he explained.
Missing persons
The CEO said that the network – running alongside an LLM for inference tasks – could be used in helping to locate lost children and other missing persons.
“Show me a three-year old that was lost at this intersection at this time of day and tell me what direction that toddler was travelling in,” Cole said, illustrating a sample query. “The LLM integration is going to allow the city itself to be much more intelligent about everything from public safety to transport automation.”
Soma Velayutham, general manager of AI, 5G and telecoms at Nvidia, noted on the call that such public safety tasks could include data “from not only cameras but also consumer devices,” using “citizen engagement through mobile phones” to bolster public safety applications.
Roy Chua, principal analyst at AvidThink told Fierce that safety-related monitoring with computer vision – known as CV powered by AI – has been one of the main use cases for smart cities, construction, utilities, healthcare, transportation and manufacturing, often paired with a 4G or 5G private network. “This is a continuation of that trend,” Chua said.
“Many of the smart cities, smart buildings, etc., use cases relied on computer vision and to the extent that the AI models get smarter and can reasons, more automation can be applied,” he added. He expects that public safety, health and safety in the workplace, patient safety in healthcare environments will be a high-value, easy-to-sell use case, followed by others like traffic management.
“As the costs drop for private networks and our understanding of best practices improve, and vertical applications (particularly CV-related) at the edge mature, then I would expect to see more of these. The AI-enabled edge inferencing does have business value — the uptake has been slower than we expected but perhaps with the GenAI hype, that could change,” Chua concluded.