The new network is born at the MWC: more automatic and more reliable

In Barcelona, ​​at the Mobile World Congress 2026, “networks” return to the center of the debate. The common thread of the innovations presented by telcos and suppliers is clear: with the expansion of artificial intelligence and connected devices, connectivity must become more responsive, more manageable and more reliable. It is not enough to increase the bandwidth: short response times are needed, the ability to adapt to different uses and tools that make life easier for operators and users.

By putting together the innovations, MWC 2026 talks about a network that is changing its role: it is becoming “closer” to the data, more automated, more open thanks to standards and interoperability and more oriented towards service continuity.

And if the large telco architectures aim to support the A and the industry with reduced response timesthe world of home networking tries to bring the same concepts into homes: more stable coverage, simpler management and connections that don’t stop at the first unexpected event.

FiberCop: the cloud is approaching the territory

The experimentation announced by FiberCop with Microsoft, Dell Technologies and 6WIND goes in this direction. The idea, told in simple words, is shorten the road what the data needs to do: move some of the processing closer to where the traffic originates. This allows applications such as robotics, industrial automation, advanced logistics and digital twins, which must react in near real time, to function better.

The field test revolves around an Edge Cloud node installed with Microsoft Azure Local technology and Dell infrastructure support, with 6WIND virtualized network functions integrated directly into the access network. A key element is Selective Local Breakout: the network chooses to handle locally only traffic that requires high performancewhile the rest remains in centralized management.

The result, according to the company, is a more efficient and flexible architecture, with latencies typically between 1 and 5 milliseconds. An important aspect is the possibility of implementing this architecture even in small-sized exchanges, enhancing the capillarity of the network across the territory as a basis for new generation digital services.

Ericsson: from 5G Standalone to 6G AI-native

Another major theme of the congress is the evolution of mobile networks towards increasingly “cloud-based” and ready-to-use architectures. Ericsson insists on transition to 5G Standaloneconsidered crucial to enable advanced functions and a more “programmable” network.

In parallel, the company is looking further: with Intel it announces a collaboration to accelerate the construction of an ecosystem capable of accompany the transition to 6G AI-nativethat is, networks designed from the beginning to distribute artificial intelligence between devices, edges and clouds. The stated goal is to more rapidly bring research towards implementable and commercial solutions, working on connectivity, cloud and computing and strengthening the ecosystem’s alignment on standards.

Open standards and more autonomous networks: the Ericsson–Nokia agreement

To make networks more automatic, a common language between platforms and suppliers is also needed. It fits in here the collaboration between Ericsson and Nokia: The two companies aim to promote compatibility and open standards for rAPPs (software applications that help manage and optimize the radio network) and for Service Management and Orchestration (SMO) environments.

Concretely, the goal is to facilitate multi-vendor ecosystems – therefore more choice for operators – and accelerate the evolution towards autonomous networksin which many optimization and management operations occur with less manual intervention.

Samsung and AMD: more software, more flexible networks

Samsung is also pushing for “software-first” transformation of network infrastructures. With AMD it announces progress on 5G Core, vRAN and private networks, and above all the transition from the verification phase to commercial implementations. The focus is on AI-enhanced vRAN and multi-cell testing that is intended to make adoption more scalable and less dependent on specialized hardware components. “Network in a Server” is also shown at the fair, proposed as a virtualized edge-AI solution, with use cases that include video analytics and sensor-based services.

TP-Link connectivity as a “service”

On the provider front, TP-Link with Agenet tries to shift attention from connectivity alone to services: a platform that recalls the concept of “app store” for routers, with functions that can be distributed more quickly and tools that, thanks to AI models, promise diagnosis and proactive network troubleshooting. The goal, from the operators’ point of view, is to reduce complexity and operational costs, improving the user experience.

Fritz! instead presents new Wi-Fi 7 solutions, 5G/4G models for fixed wireless access and previews of Wi-Fi 8, with a focus on remote management and continuity of service even in the event of main line failures.