AI factories, AI-RAN and new revenue streams: Nvidia sharpens its telecom pitch at MWC

Details: Manny Pham; Category: Service Management; 23 March 2026; 20847 views

At Mobile World Congress 2026, telecom operators are under increasing pressure to turn years of network investment into meaningful returns. For Nvidia, that challenge is becoming central to its telecom strategy, as the company positions AI as both a growth engine and an operational reset for the industry.

Speaking to Developing Telecoms, Ronnie Vasishta (pictured), SVP of Telecom at Nvidia, frames the opportunity in straightforward terms. "There's top line growth, there's productivity enhancements, there's network operations," he says. "If you're not getting on that journey, you risk missing those opportunities."

While Nvidia's presence in telecom is not new, the way it engages with operators has evolved. Rather than selling technology in isolation, the company is leaning into its role as a platform provider, tailoring its message to the specific needs of telecom customers.

"We're a platform company," Vasishta explains. "We develop compute platforms, cloud software stacks, libraries, but the platform itself supports multiple different vertical applications. When a telco looks at us, they want to see things that are meaningful to them, and that we understand the challenges they're going through."

That understanding is increasingly focused on how AI can move beyond experimentation and into core network functions. Early deployments centred on customer care and internal tools, but attention is now shifting to areas where the impact is more immediate and measurable.

"Network operations and management is a very big thing for telcos," he says. "A lot of that lends itself towards AI optimisation and agentic implementations."

At the same time, operators are looking for ways to generate new revenue, something that has remained elusive in the 5G era. Here, Vasishta points to a wave of emerging AI-driven services that could finally change the equation.

One example is the use of connected cameras combined with video language models. "Think of them as sensors that can react to what they see," he says. "You can train them in the environment they're in and make real-time decisions based on that. When you run that over a low latency network, you can use it for public safety, for enterprises, for industrial environments."

Voice is another area where AI is opening up new possibilities. "One operator announced live translation where you're speaking one language and the other person hears it in their language, but in your voice," Vasishta notes. "It's a meaningful conversation without long delays, and people are willing to pay for that."

He also highlights the potential for AI-driven call assistants, particularly for small and medium-sized businesses. "You call a store and ask what products they have, or ask for recommendations," he says. "That's a difficult conversation for a person to handle at scale, but it lends itself to AI."

Reworking the economics of the RAN

Many of these opportunities are tied to Nvidia's vision for AI-RAN, which integrates artificial intelligence directly into the radio access network.

Vasishta describes three layers to this approach, ranging from using AI to optimise network performance through to running AI applications over the network itself. However, it is the ability to share infrastructure between RAN and AI workloads that could have the biggest financial impact.

"Radio access networks are normally provisioned for peak capacity, but they're only around 30% utilised," he says. "Now you have infrastructure that can also be used to generate tokens, and that creates monetisation opportunities."

In effect, this turns the network into a dual-purpose platform, supporting both connectivity and compute. "Historically, you had single-purpose equipment generating revenue through a very traditional connectivity model," Vasishta adds. "This adds a lot of new opportunities."

He argues that this shift could significantly change the economics of running a network, particularly as operators look to justify continued investment in 5G and beyond.

AI factories and the push into infrastructure

Beyond the RAN, Nvidia is also seeing growing momentum around so-called AI factories - dedicated infrastructure that operators can use to deliver AI services.

"We have over 20 operators that have already decided to build AI infrastructure," Vasishta says. "That's not for the radio access network per se, but to create new business model opportunities."

Much of this activity is taking place in emerging markets, where operators often play a central role in the digital economy. In these regions, the ability to provide AI services locally is particularly valuable.

"Telcos are trusted in their region, and data sovereignty is important," he explains. "They can provide that AI factory resource to the region, whether it's for healthcare, education or government services, and do it in local languages and dialects."

This creates a potential pathway for operators to move into AI infrastructure, even as hyperscalers dominate globally. In sectors where data sensitivity is critical, local providers may have an edge.

"In some cases there's an imperative," Vasishta says. "In banking or healthcare, where data protection is very important, there's always a trade-off. Some organisations want to use a local telco they already have a relationship with."

Ecosystem and execution

Nvidia's ability to execute on this vision depends heavily on its ecosystem. The company works across the full stack, but relies on partners to bring solutions to market.

"We absolutely rely on our ecosystem," Vasishta says. "There's no way we can address everything, whether it's building the platform, deploying it, or developing applications."

Traditional telecom vendors remain central to that effort. Companies such as Nokia are responsible for integrating and delivering solutions that operators can deploy at scale.

"They're very important partners that telcos want to buy from," he says. "We need to support them so they can take these solutions to market."

Closing the gap between AI and telecom

Despite the momentum, a fundamental challenge remains. The pace of innovation in AI is accelerating rapidly, while telecom networks have historically evolved more slowly.

"The pace of innovation in AI is creating new capabilities every month," Vasishta says. "Telecom hasn't moved as fast, and those two things have to match."

The answer, he argues, lies in software-defined platforms that allow operators to adopt new capabilities without waiting for hardware cycles. This will be critical as the industry moves towards 6G, which is expected to be AI-native from the outset.

However, Vasishta is clear that operators do not need to wait. "You can run many of these applications over 5G today," he says, pointing to 5G Advanced as a stepping stone that enables faster upgrades through software.

A new experience for users

For consumers, the impact of these changes will become visible through new types of services and experiences. Vasishta points to real-time language translation, immersive media and augmented reality as just the beginning.

"You'll see robots you can talk to, augmented and virtual reality experiences, and the ability to put yourself into a sports event," he says. "The quality of those experiences will continue to improve as AI and networks evolve."

He also highlights a broader shift in how networks are used, with more uplink-heavy traffic driven by AI-generated content and real-time interaction.

A question of timing

For operators, the opportunity is clear, but so is the risk of moving too slowly. While some are already investing heavily in AI infrastructure and services, others remain cautious.

Vasishta expects that to change as early adopters begin to demonstrate results. "Some will move quickly, others will take time," he says. "But once they see success, many more will follow."

For Nvidia, the goal is to ensure the platform is in place to support that transition. For operators, the challenge is deciding how quickly to act.

As Vasishta puts it, the industry is at a point where innovation is no longer the limiting factor. The real question is whether telecom can keep up.