Three of the largest companies in telecom and technology are pushing artificial intelligence out of far-flung data centers and into the infrastructure of everyday life, a shift that lets AI act on information the moment it is created rather than after a round trip to a remote server.
A new set of announcements from telecom and infrastructure companies at Nvidia GTC shows how quickly that model is taking hold. Instead of relying on centralized cloud systems, companies are embedding AI directly into networks and physical systems so it can process information and act in real time.
AI Grid on Telecom Networks
AT&T, Cisco and Nvidia said they are working together to create what they call an AI Grid, a system that allows AI to run directly on telecom networks. The platform combines connectivity, software and computing hardware to process data closer to where it is generated.
By keeping data local, the system reduces delays and makes AI more practical for everyday use. It is designed to support applications such as video monitoring, transportation systems, and industrial operations that depend on fast response times. It also gives companies more control over their data, since it does not need to travel as far to be processed.
The effort demonstrates a broader change in how networks are used. Telecom infrastructure is no longer simply moving data. It is becoming a place where AI systems run.
PYMNTS has previously reported that AT&T is also exploring autonomous AI agents to reduce fraud and customer wait times, underscoring how the company is expanding AI beyond back-end systems into business workflows.
Advertisement: Scroll to Continue
AI on the Network Edge
T-Mobile, working with Nvidia and Nokia, is applying the same idea to its 5G network. The companies are testing systems that enable network sites, such as cell towers, to run AI applications alongside their traditional roles.
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
These systems have been designed for what the companies describe as physical AI, tools that interact with the real world. That includes analyzing video, monitoring infrastructure and helping manage operations in cities, factories and transportation systems.
Because the AI runs at the edge of the network, it can respond quickly. Instead of sending data to a central system and waiting for a response, it can process information and act almost immediately. That makes it more useful in situations where timing matters.
This approach shows how telecom companies are expanding beyond connectivity into computing, using their networks to support new AI-powered services.
As PYMNTS has covered, T-Mobile has also been exploring how 5G and embedded payments can enable new connected retail experiences, pointing to an expanded push to turn network infrastructure into a platform for live services.
AI in Energy and Utility Systems
Itron is applying the same concept to energy and utility infrastructure by embedding AI into its distributed intelligence platform. The system processes data directly within the grid instead of sending everything to a central location.
This allows utilities to respond more quickly to changes, improve efficiency, and better manage resources. For example, systems can monitor situations and adjust energy distribution in real time, helping prevent outages and reduce waste.
Because processing occurs locally, the system can continue to operate even with limited connectivity. That makes it more reliable for critical infrastructure.
Together, these announcements point to a clear direction. AI is being built into the systems that run everyday operations, from telecom networks to power networks. As it moves closer to where data is created, it becomes faster, more reactive and more directly tied to how the physical world is managed.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
Add as Preferred Source
