The View from Davos with Ericsson’s Niklas Heuveldop and Åsa Tamsons

The AI chatter in Davos is loud. The hard part is still quiet: getting the physical world connected fast enough for AI to matter.

Ericsson’s Niklas Heuveldop and Åsa Tamsons join Daniel Newman for a grounded look at what enterprise and agentic AI actually require to deliver ROI. In ports, airports, mines, factories, and other industrial environments, 5G and advanced connectivity are no longer optional. They are the control panel for real-time automation.

The discussion moves quickly from ambition to execution. From, why large-scale industry transformation has lagged to what China’s momentum reveals about adoption at scale, and why enterprises keep circling back to the same demand: trusted, reliable connectivity that’s as simple to consume as the cloud. The goals are clear: stop optimizing for tools. Build the stack. Connect the devices. Make the data usable. Then move, fast.

Key Takeaways Include:
‍
🔷 Enterprise AI needs real-time connectivity to leave the demo stage: Sensors, machines, and edge devices create the workload, and networks determine whether that workload can run reliably at business speed.
🔷 5G was built for “many device types,” not just smartphones: Industrial AI depends on connectivity that is smart, secure, and performance-tuned to device requirements.
🔷 Physical AI success is a stack problem, not a single-tool problem: Cloud, connectivity, compute, and model access have to work together, then teams can iterate on what creates value.
🔷 Adoption and culture decide ROI: Faster feedback loops and “learn fast” execution matter more than steering-committee perfection when the transformation window is shrinking.
🔷 Edge + wide-area is the real operating environment: Many use cases start inside factories and then extend beyond them, which raises the bar for consistent networking and applications across contexts.

Subscribe to our YouTube channel so you never miss an episode.

Listen to the audio here:

Disclaimer: Six Five Media’s The View from Davos is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.

‍

Daniel Newman:
Hey everyone, the Six Five is on the road with a view from Davos. Here we are, it is day three. We are outside, and this is the busiest Davos that I've been to over the last handful of years. A lot going on here, a lot of things in focus, and today is expected to be a really big day. But I'm excited to start the day here this morning on the street talking to a few friends, Nicholas Ossa from Ericsson, who have Been with me, we've been together at least half a dozen times now, right?

Åsa Tampson: 

Yes.

Daniel Newman: 

I mean, not always, but I still remember one of my first Davos. I had you and it was bitter cold and we had snow up to our ankles. Nicholas, we usually go inside and talk somewhere nice. I don't know what the difference is here, but it's great to be… It's more 4x4. I had to actually climb out of the last conversation. No, we had a great walk and talk last year, actually up and down this promenade, but this year has been something really interesting. I'd love to start off with just kind of hearing from both of you. Every year you come here, big event brings together policy. It brings together government, brings together, of course, technology is a big focus, but we knew this year is a little different. Without getting into the weeds and the politics, what were the big things that, you know, also start with you? You were here in Davos trying to get done.

Åsa Tampson: 

So obviously I run enterprise wireless solutions for Ericsson. We see 5G and advanced connectivity as a critical part of the network stack. And if you want to scale AI automation, it's fundamental. We have a strong network in the telco industry. This is the space to meet a lot of people, a lot of executives, decision makers across industries. And since our solutions are suited for ports, airports, mines, factory, across all industry, retailers, financial services, public sector, this is my hunting ground. I go and meet with either current customers, partners, but a lot of new people that are not working with us yet, that I want to get to know better, want to understand what they're trying to do with automation AI and how we can help solve their connectivity issue.

Daniel Newman: But you said it at the end there, automation AI. I know, Nicholas, that you're very focused on bringing AI out to the edge. We know that there's a massive boon in data center. That tends to be a little bit of the focus. But we've been at least saying that this year, we believe it's the year of enterprise AI. It's the year of agentic AI. And it's also the year of very verticalizing more to get more ROI. Because realistically, all this large language models, very general, they do some very neat things. But I keep saying they're the parlor trick. When you start taking that really high value data that's inside an enterprise and then apply it to some of these big language models, there's a lot that can be done. So what are you zoning in on here?

Niklas Heuveldop: 

Yeah, we've been talking about that for a while, haven't we? Yes. How do you drive the industrialization of the enterprise transformation at scale? So we discussed 4G kind of being a great platform for the consumer digital transformation. And then we took stock in the fact that, well, enterprises are not going to transform at the back end of a smartphone. There is going to be sensors, there's going to be humanoids, there's going to be autonomous guided vehicles, drones. It's a very different ecosystem of devices that you need to connect. So we created 5G. 5G basically can cater to the individual needs of different devices with very specific connectivity capabilities. Smart, secure and high performance depending on what the device needs. So we launched 5G 10 years ago. We've been on stage together talking about the potential to transform industries and here we are 2026 And there is very few examples across the globe where this industry transformation has really happened at the back end of that connectivity opportunity.

Åsa Tampson: 

And at scale. And it's really at scale.

Niklas Heuveldop: 

If we're honest about it, I mean, China is kind of leading, leaps and bounds, in, I mean, what is it, 42% of the global lighthouse factories, according to World Economic Forum, in China. over 50% of, 55% actually, of IoT devices in China, right? So they have really figured out how to drive industry transformation and scale. And then you have AI on top of that. And for the last 12 months, agentic AI has really taken off at massive scale. And then you have to ask yourself, well, how do you then get all this data collected to feed AI for genetic AI to make the right decisions, then you need to connect the devices and you need to do it in real time at scale. And that's the big lift this year. And that's been kind of the focus of our conversations here. So I'm here, I don't know if I think about it as a hunting round, but I'm kind of trying to rally the different players in the ecosystem to come together so that we can accelerate the availability of devices to support all these use cases, and then build applications that can run across private networks, wide area networks to support these.

Daniel Newman: 

I'm glad you mentioned AI, because I kind of come to the conclusion that the IoT era, the edge IT era, kind of hit a bit of a wall. You said basically never scale it. I think AI brought it back to life.

Niklas Heuveldop: 

Yes.

null: 100%.

Daniel Newman: 

And it's accelerating it really quickly. Osa, you talk to enterprises. You're focused on enterprise wireless for the business. You're having those conversations, you're hunting here. What are the enterprises telling you about what they need that maybe they're not getting either from the hyperscalers? Because we know a lot of workloads are going to the cloud. And what are they telling you about how they're planning to build out their edge in their enterprise networks to be able to support what they're doing in the cloud?

Åsa Tampson: 

You know, to be honest, not always they have actually a full understanding of what they need. What they do know is they need a trusted, advanced, I would say trusted, reliable connectivity for what they need to do. And for massive video streaming, real-time of something that moves, there's certain capabilities they need, right? Sometimes it's just that sensor in the right place and it can easily run maybe the elements out at the edge. So it's more of those questions, right, than where they figure out, well, we're actually building a new factory. So I spoke with a steel company the other day. They're building a new melting plant, and they say, well, we only do this once every 40 years, right? We're going to sit with this for a long time. We want to make sure we have the best infrastructure to support that operation. So there are a lot, actually, of conversations that happen even before they think about, how is my architecture going to look like at the edge? They're more thinking, how can I solve so my devices are connected? How can I solve so? And I think that's interesting. We talk a lot about AI, genetic AI. The moment you want to combine that with the physical world, physical devices, that's where we have a lot of conversations. So my big takeaway, if we really want physical AI to happen, which we're already seeing happening at scale, in industries in China, you need cloud, you need also wireless connectivity. And you need that to work for whatever use cases. And enterprises don't want to go into the details of understand how that is being solved. They want it to be as easy to access, integrate, as you would have from cloud resources. And I think this is where really exciting, where the two businesses we're running, how that really comes together and solving that.

Daniel Newman: 

It's interesting, I hosted a WEF panel with IBM yesterday and it was, there was a topic, Rob Thomas, one of the execs there, he said, boring is beautiful. And if I'm listening to you, and no knock on anything here, but there is a bit of kind of returning to boring and execution. And that's kind of what I'm hearing here is like, look, you need the infrastructure. You need to connect to cloud. You need to have software and platform flexibility. You need to be able to make sure all your devices are connected and all your data is accessible. And at that point, then we can start to see AI payoff. And we only have a minute here, but I would love to get your read on that, Nicholas. How are you, what is the kind of the big recommendation, the big picture that you're giving to your enterprises and your clients and your partners to basically execute To start to deliver and get to the ROI part, because we're not going to stop talking about a bubble until the businesses, the manufacturers, and the healthcare companies start to say, we're making money from AI.

Niklas Heuveldop: 

And that is the question for the last year. All these investments in AI, what's the ROI? And I think it comes down to adoption. And again, we take a page from what we have seen in China. I mean, they are leaps and bounds ahead. Small language models at the edge. Certainly open source. And open source. So they have really cracked the formula for how you help industries leverage technology. at the edge to drive industry transformation. And that's the big conversation. How do we now replicate that here with our customers? So the big conversation is, how do we get the device ecosystem invigorated? How do we make sure that we build applications that can leverage these networking capabilities in private networks, but then also in the wide air network? So we're working with smart manufacturing. So you have a number of use cases inside our factory. You were at our smart factory. Yes, I was. And then that device or component leaves the factory and is on a wide area network and needs to be connected, right? So we know exactly what it's going to take. We've been playing around with this for 10 years. And now with Agentica AI, we have a whole new tailwind that should help us accelerate this whole journey. So I'm super excited about where we are.

Åsa Tampson: 

I think one thing we heard yesterday morning in a session which really stuck with me was, you know, don't spend too much time thinking about what is the best tool to take. There are certain evident layers you need. You need connectivity, you need cloud, you need compute. And then you need to be able to access a couple of different language models, right? But you need, especially if you're a large company, you need a stack. And maybe not spending too much time, should I use this tool or this tool? But you need a complete stack. Once you have that, I think it's all about culture and learn fast. Start trying it out. We were listening to the team that built Maven for NATO yesterday. They were forced, it's interesting, such an organization, they were forced to put out their new AI-driven operating tool before it was kind of ready, which is very uncomfortable for them to do. And they did it in a test environment and started dropping improvements within one week, two weeks. start to have that agile development between people in the field and people developing product. I think it's just an example. If you have that sense of urgency, you start to try out and you kind of break the wall. So the classic ways are working. We go into a meeting, we agree on a plan, we do a project, and then we go back in another steering committee. I think this is all about that, right? And I think that's interesting. So experiment, try out the technology, decide what's actually making value for your business. and then move fast. And that's a massive cultural change, I think, for any company that is not AI native.

Daniel Newman: 

This transformation is just faster than any in the past, giving us a lot less time to basically digest and make decisions. I do have to wrap up. What I'll say is, you said something very poignant. I want to point that out to the audience. Right now, having the access to the compute and the connectivity is really the winning formula. And I think you're seeing it with Meta and others that have sort of pivoted away from saying, we want to build the best model. And they're pivoting to, we want to have the most compute. Because once you have that, you have a lot of flexibility in terms of which model, how you build applications, where you deploy them, and how you basically monetize and build the business around it. Ossa, Nicholas. Great to see you both. Thank you so much for the time. So nice seeing you. And thank you everybody for joining us for this Six Five On The Road with a View From Davos. Great conversation there. Hope you take it all to heart and I hope you subscribe and be part of our community. We appreciate you. We'll see you all later.

‍