Hello and welcome to the India Today Business Today studio here at the World Economic Forum in Davos. I am Sidharth Zarabi and with me now, Erik Ekudden, the CTO of global telecom major Ericsson. Erik, welcome. Thank you very much for your time.
Erik, I want to begin by asking you about the latest breakthrough that you can talk about to set up this conversation as far as telecom is concerned. Because the big headline is the age of AI, but we must not forget that everything that constitutes the digital and the AI world today has been built on the backbone of absolutely fantastic networks, and you as a company have led that for decades now. How do you talk about the latest breakthrough technological advancements in the world of telecom as of today?
So here in Davos, we talk a lot about return on investment on AI. We talk about AI diffusion, how to get AI out and really create the value. I think we see that certainly this year will be the year when you get more clear return, but you also start to broaden applicability in enterprises. Workflows with agentic AI are coming online and clearly you will see a lot of change in how we approach work. But AI in the networks has actually been there for some time. So we could talk about breakthroughs, but I also think we need to talk about the fact that 5G, to your point, really good 5G networks and networks that can be programmable so that you can set up the kind of network performance that the application needs, networks that are pervasive—that is really the foundation actually also for AI. So I think that now we've come to a point with Gen AI and also agentic AI and soon physical AI that will demand even more from the network. So think about that best network for AI that itself needs to be AI-powered: energy consumption, better user experience, better total cost of ownership, simpler to create new services on the fly. AI is already powering the 5G network, so you have AI-powered 5G networks. So some of these technologies that we now talk about moving across society, they have already started to move into the network. More inference, meaning AI generation rather than training, and that needs to be distributed. You don't only do that in the data center centrally; you do it distributed in the infrastructure, of course in your glasses and the phones, but you also do it in the network. So the network becomes a host for inference, and that inference then needs to be energy-efficient, because obviously this is one of the key criteria to optimize for, but also the cost—the tokens per dollar per watt needs to be optimized. So there are breakthroughs in terms of bringing AI into the network, AI RAN in 5G today, and bringing these AI workloads onto the network. But then I think I'm super excited about what networks will be used for. Today, 5G is used for consumers, for enterprises, for SMBs, governments, public safety, also defense. In the future, the networks will really be used also as a sensor—finding out not to get perhaps run over in a crossing in a city or making drone surveillance. I mean, there are lots of things that the network will do that we haven't thought of before. So some of those breakthroughs we are now showing for lead customers and of course they will be making their way into the network over the coming period. Even simpler things but that impact common users on an everyday basis, which is spam calls for example.
It's an excellent example. Fraud prevention in today's society is a big topic—for consumers, but ultimately it's actually for the reputation of the bank or the financial service. Can you trust that this transaction actually works? Is this payment legitimate? All of those things, the network can now start to serve up. We call it fraud prevention, but it's actually a set of network capabilities served up over network APIs on the Ericsson-Vonage platform, provided to application developers. Whether you're building an app for the bank or financial services, you can now start to tap into networks—not just in India, not just in the US, but across the world—because these networks are becoming open and programmable through that network aggregation or the network API layer. And I think that the value for, for example, banks and financial institutions, but frankly enterprises, everyone—of having strong security, strong identity management, and for example fraud prevention—that's not just a consumer service. Those kind of services will be for basically everything.
In India, there's this very scary trend of cyber arrest. People get a call on a video call. They're made to believe that they are being investigated by a security agency and then they are made to empty their bank accounts. It's actually very, very scary. So that's a very good point that you make.
It is scary. So you actually need to get support from the network in terms of not just weeding out the simple part, but also going into the fake part, where actually AI will help in a big way. Because AI will actually help to detect but also to warn consumers that this might actually be something you should not engage in.
