At MWC 2025, Kevin Wollenweber, SVP/GM at Cisco, discussed the rapid evolution of AI and its growing demands on network infrastructure. He highlighted a shift from carrier cost-cutting to exploring new revenue opportunities through AI-driven services and applications. With AI moving from cloud to enterprise data centers, Cisco's role in simplifying and speeding up deployments is becoming more crucial than ever.
Wollenweber also forecasted a potential optical networking boom, driven by the vast volume of AI-generated data traveling across networks. The future, he suggested, would see data dispersed further throughout the network, with more AI applications interacting across devices. He also spoke about the industry's embrace of Industry 4.0, where AI will transform various verticals, driving innovation and operational efficiency at an accelerated pace. With physical AI on the horizon, Cisco is at the forefront of this transformation, enabling the seamless integration of AI into real-world applications.
Steve Saunders:
Kevin, it's great to see you again.
Kevin Wollenweber:
Good to see you too.
Steve Saunders:
Yeah, last time we were talking it was about ethernet, so it must have been a little while ago. And the industry has certainly moved on a bit since we last had the chance to talk about the carrier business. It seems like there's been a shift recently from carriers looking to save money on OpEx to working out new ways to make money to increase their revenues with new types of service or application. Do you think that's true? And how's Cisco working with those carriers?
Kevin Wollenweber:
Yeah, it is definitely true, and if you kind of look at the evolution of AI technologies, we had the major builds with the hyperscalers and the LLM builders, but now we're kind of moving into that world of inference and usage of the models. And so I think companies like Cisco and service providers can have a huge play in that space. So as the models move outside of the cloud and really into the enterprise data centers, service provider networks, and even out into the world itself. The solutions that we're building and the simplifications we're driving to those solutions can help them deploy much, much faster.
Steve Saunders:
So one of the things which I'm hearing here at MWC 25 is that obviously AI is putting increasing demands on the network. As a provider of network solutions, that's obviously good for you. Do you think there's a possibility both inside the data center and between the data center that we're going to see a sort of new optical networking boom being caused by this vast volume of AI-generated data traveling over the network?
Kevin Wollenweber:
Yeah, so we've seen a huge increase in data in the hyperscale networks that we've been building out. And so as we see more and more usage of the AI applications and AI infrastructure being built, I expect to see a lot more data dispersed further in the network. But also, you kind of think of this futuristic world of agents and moving to physical AI and robots. I expect to see a lot more agents and AI applications talking to each other, so we'll drive more diversity in where the data actually is and more communication between different devices. So we definitely do expect to see a growth from this.
Steve Saunders:
Let's stay on data for a moment. I mean, a lot of people are talking about distributed data, and at the same time, hyperscalers and others are building these huge centralized data centers. Which approach wins out? Because the data, it seems, would work better if it was closer to the end user.
Kevin Wollenweber:
Yeah, and I'll use my product answer and just say, I think it's going to be a combination of all of those. And so if you think about massive training workloads, those are definitely very centralized and being built by a small number of very, very large companies. But as you start to think about fine-tuning and inference, it's really about data gravity and where the data resides. And so as more and more of proprietary, enterprise, and customer data is being used, we're going to see a lot of these AI applications pushing deeper into the network and a lot closer to where those customers reside and where that data resides.
Steve Saunders:
Let's think about the future of the communications industry. I think that within three or five years, a show like this will be dominated by digital industrialization, industry 4.0. I really see these technologies, automation, predictive analytics, AI moving into the industry verticals and being transformative there. Do you agree with that? And how important is that to Cisco as a potential market?
Kevin Wollenweber:
Yeah, and actually it may not even be three to five years. If you look at the acceleration that we're seeing in this space last year, we talked a lot about AI, but it was really AI for the sake of AI. We didn't really know exactly what things would look like. I think this show, you're seeing a very different vibe and you're seeing a lot of people looking at real-world use cases and examples of how they're going to leverage these technologies.
And so in three to five years, or potentially even faster, I definitely expect to see usage of AI and actually AI enhancing everything that we do and not just being the thing we talk about, but actually being built into everything that we do. And so acceleration of new service offerings, getting to innovation faster, and the operational savings a lot of people are enjoying today from some of these AI applications is absolutely going to be part of that next industrial revolution.
Steve Saunders:
I agree with that. It seems to me that the best way for a service provider to plan around AI is to assume it's going to be in every device and it's in their network, to look at it as a table stakes kind of commodity and then assume it's there and then work out how they're going to use it. I mean, that's a different way to the way we talked about AI last year, but that's very much in line with Cisco's new philosophy of AI, I think, right, Kevin?
Kevin Wollenweber:
Yeah, a hundred percent. And I think the faster AI moves, the more we learn about it, the more it's going to be used and how it's going to be used. And you're moving from generative AI and building the large rankings models to agentic AI and all the talk we've been having around a mixture of intelligent agents working together to solve problems to that view of physical AI.
Physical AI is interesting because when we get to the world of AI being outside the data center and in the physical world, robots and manufacturing floors, and a lot of the industrialization that we talked about, I actually think it's a domain of companies like Cisco and our service providers to actually bring that to market. So it's an exciting time and it's perfect for shows like this.
Steve Saunders:
Kevin, it's great to see you.
Kevin Wollenweber:
Great to see you too.
Steve Saunders:
Always a pleasure.
Kevin Wollenweber:
Appreciate this. Happy to do this anytime.
Steve Saunders:
Thank you.