In the AI era, telcos are rethinking infrastructure, accelerating innovation, and redefining revenue models — and Nvidia is positioning itself as a key enabler
Telecom has never been known for rapid transformation. For decades, operators have built networks in predictable cycles, layering new technologies on top of legacy infrastructure while prioritizing reliability and cost over agility and innovation. But according to Ronnie Vasishta, who is responsible for Nvidia’s telecom business, strategy, and products, that paradigm is shifting — fast.
“Every telco has something to say now about how they’re adopting AI,” he told RCR Wireless News, describing a telecom landscape where AI has become a catalyst for true structural change, reshaping everything from customer experience to network planning to how operators think about their business models.
AI — from call centers to core networks
The first wave of AI adoption in telecom was straightforward: customer service. “That was a quick hit and … it’s not about just AI chatbots,” said Vasishta. Operators quickly deployed generative AI tools to improve call center performance, reduce ticket resolution times, and streamline interactions. But the real transformation is happening deeper in the network, where “agentic” AI workflows are being embedded directly into operations.
Agentic AI refers to an artificial intelligence system that has the agency (within guardrails) to go beyond augmenting a process or workflow to automating a process or workflow by accessing data and an application or applications to take your intent and turn it into an outcome. Essentially, these systems don’t just answer questions — they understand goals, make decisions, have access to data, execute across tools, and adjust dynamically, often without direct human input.
Global systems integrators, working on Nvidia’s frameworks, are now building AI-driven tools that automate network monitoring, predict outages, and even plan coverage for large-scale events. SoftBank, for example, is using large telecom models (LTM) to dynamically configure networks for stadiums and public gatherings, ensuring coverage where and when it’s needed. According to the Japanese operator, by fine-tuning the LTM for specific use cases, it will become easier to develop dedicated AI models tailored to various operational scenarios in cellular networks.
And it’s not just operations. Telcos are adopting AI assistants for internal use — from network planning to workforce productivity — signaling that the industry is finally embracing automation at scale.
A shift to software-defined infrastructure
But building AI into telco operations is only half the story. Real transformation will require fundamental changes to the network’s underlying architecture and the physical network itself. Vasishta shared that more than 20 operators worldwide have announced AI data centers, often smaller than hyperscale builds but strategically placed to support regional needs like data privacy and mission-critical workloads. These facilities are laying the groundwork for a future where networks are not just connected but intelligent.
“And while they may not be the 100,000-type large GPU kind of gigawatt-type centers, they are very unique because telco has a unique position within a region. Telcos are always kind of a trusted entity by the governments and by the enterprise and research for data privacy, for mission-critical equipment, for regulations,” he continued.
The linchpin, he said, is the move to software-defined infrastructure. Traditional, purpose-built systems can’t evolve fast enough to support emerging AI applications like integrated sensing, drone detection, or real-time orchestration of low-latency services. A software-defined approach, by contrast, allows operators to update capabilities in near real time — and it’s the foundation for what many now call AI-native 6G.
Breaking barriers to 6G
Challenges remain, particularly in the radio access network (RAN), where legacy architectures have slowed adoption. But the pressure is mounting. Traffic is increasing — and changing — as devices generate richer, AI-driven data. At the same time, the industry’s economics are shifting. “The industry needs 6G to be a software upgrade,” Vasishta said — a necessity if operators want to avoid another costly, full-scale infrastructure overhaul. “We can’t really afford for 6G to be a major capex spend without new business models,” he added.
This transition also reopens an old debate: telcos as “dumb pipes.” Vasishta — and Nvidia — sees an opportunity for operators to monetize infrastructure itself, offering differentiated services at the edge and hosting specialized workloads for enterprise customers.
Accelerating innovation cycles
Historically, telecom research has taken years to materialize in commercial networks. Vasishta claimed AI can compress that cycle dramatically. With 6G on the horizon by 2028–2029, the company is betting that AI’s rapid evolution — from research to deployment — will define the next era of connectivity.
The industry’s “AI moment” isn’t coming. It’s already here — and telcos that embrace software-defined, AI-native architectures today will be the ones shaping the networks of tomorrow.
