AI: Missori (Ericsson), Italia lags behind on 5G Standalone, public support needed

How is Ericsson's present and future changing because of AI? We have to distinguish between two dimensions: AI for the network and the network for AI. Let's start with the network for AI. This revolution will move to the periphery: AI will enter cars, robots, visors, advanced devices and sensors. The common trait? None of these objects have a wire. AI cannot fully spread without a 5G - and in perspective 6G - network capable of supporting it. To use a metaphor, without roads there would be no automobile revolution.

Similarly, when AI moves to the edge, it requires a profoundly different network from the current one.

In Italia, except for specific initiatives such as the recent launch with WindTre, most networks are still not 5G Standalone but an evolution of 4G in best effort mode, with characteristics not adequate to support advanced AI applications. What is needed is a programmable, resilient network with always-on connectivity and low latency. A robot cannot lose its connection to the cloud that enables its functionality at any time. New services require radically different performance.

A quantum leap is needed. But how can this be achieved with telcos struggling with investment?  

Public support is needed at a crucial stage. One example could be the non-burdensome renewal of frequencies expiring in 2029, in exchange for infrastructure investments. Many operators today are restrained in their investments and are asking for support. Trade unions have also called for government intervention in favour of the sector.

Can themarket consolidation, which started in the tlc sector, have negative effects for you as a vendor?

Theoretically, yes, because you reduce the number of customers. In reality we prefer fewer solid operators rather than many struggling ones. It is better to have one player who invests heavily than many who do not invest. We look carefully at the Ran sharing agreement between Telecom and Vodafone-Fastweb: it may represent an opportunity, depending on the technological choices. If the goal is to guarantee high performance for both with the highest quality, we will be able to offer solutions that, at least in Europe, no competitor has.

Are you satisfied with the impact the new European regulations, starting with DNA, might have on vendors?

The Digital Networks Act goes in the direction outlined by Draghi and Letta. The time for implementation in individual countries will be long, but it is good that the Commission recognises telecommunications as a strategic infrastructure.

Would you have preferred more direct support to favour European champions?

We are a European champion. All over the world, local champions are supported; more favourable legislation would be desirable. I very much appreciate the Italia approach, which in public procurement rewards Western bloc technologies - European, American and allied countries. After the European guidelines, some countries adopted a ban, others a moral suasion, others intermediate solutions. Italia has chosen a rewarding approach, I hope that other European countries will follow this example. We in Italia, like Ericsson, have contributed to bringing 5G to the centre of the main sporting event of the year, connecting healthcare facilities with WindTre and supporting Tim in high-traffic contexts. The railways are also considering the use of 5G both for internal processes - signalling, operational maintenance, station management - and to improve connectivity on high-speed trains. The conditions to accelerate are there. I expect operators to invest in this direction.

After the agreement with WindTre are you talking to other operators?

We dialogue daily with several operators in the region I manage. The problem is that in order to activate significant investments, guarantees and a stable framework are needed. Many operators say they can no longer sustain these efforts alone and are eroding cash. I therefore hope for a long-term plan to accompany the investment in critical national infrastructure. Getting to 6G and fully enabling AI requires intermediate technological steps. The speed so far has not been what was hoped for and in the meantime a gap has been created between the pioneer countries and those that are following.

So far we have talked about Network for Ai, will there be an employment impact?

Every technological revolution has effects on labour. Automation in the mines has reduced some occupations but created many others. According to established research, 65% of children attending primary school today will do jobs that do not yet exist.. At Ericsson we have always taken a long-term view. Personally, I am more concerned about macroeconomic factors than the direct impact of AI on employment. A concrete example is AI applied to coding: those who develop software with AI tools work faster and often better. However, our investment in research and development - around USD 5 billion a year - is not diminishing. While less resources are needed for traditional code writing, we invest the same energy in developing the network-integrated AI and sensing capabilities that will be central in 6G. But the important thing is to have a healthy ecosystem, with companies growing and investing. The telco sector has lost about 40 per cent of its value in ten years: theoretically it could have lost the same share, but the reduction has been less. If business does not grow and the trend is not reversed, employment will not grow either and may even be negatively affected. To reverse the trend, therefore, we need new services, sustainable prices and an ecosystem that supports investment.