From real-time 5G slice validation to next-gen AI RAN software, Ericsson has highlighted a number of innovations ahead of MWC – spanning network performance, autonomous operations, and 6G readiness; with collaborations from Ookla, Apple, MediaTek, Swisscom, and Mistral AI.
In sum – what to know:
Verified slicing – Ericsson and Ookla have an “industry-first” solution to measure the performance of 5G slices, supporting SLA enforcement for consumers and enterprises.
AI and automation – New AI-ready radios, antennas, and rApps, alongside Swisscom and AWS collaborations, aim to drive Level 4 autonomous networks and AI-assisted operations.
6G ecosystem demos – Ericsson will demo with Apple and MediaTek at MWC to showcase early 6G prototypes, spectrum sharing, and systems for AI-driven network optimization.
Various telco-types loaded up their PR canons last week, and let fly – including Ericsson, which convened press and analysts in London for a MWC curtain raiser. Here is a round-up of the Swedish vendor’s big news from the event, and the week. We go in reverse order, starting with the newest news, and, in ways, the most interesting – which came via a third-party, in fact, and not via Ericsson itself. All the product announcements trail product demos at MWC (March 2-5) in Barcelona next week. Anyway, here goes…
1 | Real-time SLA verification for sliced 5G services
Speed-test specialist Ookla has worked with Ericsson on a solution to measure and validate the performance of 5G network slices. They have called it an “industry-first methodology”, which tackles long‑standing issues to monitor quality of service / experience (QoS / QoE) with 5G SA network slices – to deliver service-level agreements (SLAs) for differentiated consumer and enterprise propositions.
Ookla said: “Traditional speed tests measure the default internet connection. [This] methodology [identifies and tests] specific network slices. This… demonstrates how SLAs… can finally be verified in real-time… Crucially, this puts the power of verification directly into the hands of the [user], allowing them to use the Speedtest app to independently validate that the performance they’ve purchased is being delivered as promised.”
Ookla also called it a “pivotal first step” to guarantee 5G slices. Tibor Rathonyi, senior advisor at Ookla, added: “Network slicing is no longer a future concept; it is a commercial reality. However, you cannot manage what you cannot measure. Our work with Ericsson is a pivotal first step in providing the transparency needed to prove the value of these premium 5G services to both consumers and enterprises.”
As a footnote, here; Ookla also offered a handy description and summary of network slicing, talked about forever, but still in its infancy. “The technology allows operators to create multiple virtual networks (slices) over a single physical infrastructure, each tailored to specific requirements: ultra-low latency for gaming, high-bandwidth for 8K video streaming, mission-critical reliability for industrial IoT and emergency services,” it said.
Its MWC showcase will offer a “side-by-side comparison” (with SLA verifications, reports, and proofs) to show the performance gap between a standard 5G connection and a service-specific slice in a live 5G network (Ericsson’s, at the Fira).
2 | Pre-standard 6G prototypes and spectrum sharing
Ericsson will also show “6G innovation and ecosystem readiness” with device maker Apple and chip maker MediaTek next week at MWC. The trio will deliver live demos showing advancements in pre-standard 3GPP-based 5G and 6G spectrum sharing and 6G prototype systems, it said. The latter will also (see above) include systems for validating “essential 6G capabilities”, it added.
The demo with Apple is around multi-RAT spectrum sharing (MRSS) between 5G and 6G – to help operators migrate from one to the next, and between the two, post-2030. The exercise is about “coexistence, [and] minimizing resource waste and signaling overhead”, said Ericsson. The demo will use two PoC systems – one on 5G and one simulating 6G – that are connected to a base station operating in TDD mid-band spectrum.
Meanwhile, the MediaTek demo will show how 6G cm-wave can meet new data demands, including from sundry AI and URLLC apps. It will use early prototype RAN/UE systems (“radios and baseband processing up to the IP layer”), built around anticipated 6G features. The prototype system integrates an RAN and a MediaTek UE prototype. “The demo includes a conducted data call and highlights contention-based buffer status reporting,” said Ericsson.
3 | Swisscom closes on Level 4 network automation
Ericsson is working with Swisscom to further automate the Swiss operator’s RAN infrastructure – towards ‘Level 4’ autonomy in “certain operational processes”. Level 4, as defined by TM Forum, enables decision-making based on predictive analysis or active closed-loop management via AI modeling and continuous learning. The duo will “share progress” at MWC to “fully integrate” RAN software modules (rApps) into Swisscom’s commercial network.
Specifically, they will show three Ericsson-developed rApps, variously for cell/uplink anomaly detection and root cause analysis (Cell Anomaly Detector, Uplink Anomaly Detector ,and Anomaly Root Cause Explainer). They will also preview a new O-RAN optimization tool, RET Cell Shaper rApp, that dynamically adjusts a cell’s antenna tilt using Remote Electrical Tilt (RET) to improve coverage, capacity, and user experience.
Swisscom is using Ericsson’s automation platform, Ericsson Intelligent Automation Platform (EIAP), for AI management of multi-vendor, multi-tech 4G/5G RAN systems. The EIAP platform includes Ericsson’s non-real time RAN Intelligent Controller (RIC), the Ericsson Intelligent Controller (EIC), and an open SDK for operators to deploy software-based radio-application modules (rApps, in the O-RAN architecture) to automate RAN functions.
Ericsson claims to have corralled an ecosystem of nearly 90 members around its EIAP platform, including Swisscom and 19 other service providers. It supports nearly 90 rApps, it said – with 25 from Ericsson. Philipp Bichsel, in charge of network and services at Swisscom, said: “Ericsson’s Intelligent Automation Platform and its ecosystem of rApps play an important role in helping us explore the potential of AI driven automation.”
4 | Generative AI for network ops – with Mistral AI
Covered in RCR last week, but in brief (as written): Mistral AI and Ericsson are collaborating to apply generative AI (gen AI) tools to telecom network development and operations. Ericsson will act as a design partner, combining its network R&D expertise with Mistral AI’s model customization capabilities. The collaboration focuses on several operational and research use cases, including automation of legacy code translation, AI-assisted development in 6G research, and the development of AI agents to support complex workflows within Ericsson’s Networks organization.
5 | New Japan R&D hub for open programmable 5G
Another quick one: Ericsson’s new telecom R&D centre in Japan, announced in May 2025, will be located in Yokohama, south of the capital Tokyo – close to its facilities in Yokohama’s Minato Mirai 21 district. The site will focus on advanced radio hardware and software – “programmable 5G networks, next-generation mobile technologies, and open network architectures” – for the Japanese and global markets. It will start operations in April, with its official inauguration in the first half of 2027, and recruitment under way.
6 | Reimagining money movement with Mastercard
Also short, and just because it’s not our everyday beat, but… Ericsson has a deal with Mastercard to “reshape how money moves across the world”. It is to integrate its fin-tech platform (Fintech Platform) with Mastercard’s portfolio of money movement solutions (called Move). On the tech side, the project is geared around APIs, and “cloud-native” and “compliance-ready” infrastructure. On the service side, it is around “digital wallet capabilities, new payment services, and unbanked or underbanked communities”.
Some backstory: “Financial inclusion and accessibility are key focuses of the collaboration. Mastercard Move enables money movement across 200 countries and territories, connecting more than 17 billion endpoints, and supporting transactions in 150 currencies. Ericsson’s fintech platform operates in 22 countries, serving more than 120 million active users and processing more than four billion transactions every month across digital wallets, payments, remittances, lending, and loyalty services – all backed by enterprise-grade security.”
7 | AI-first radios, software, antennas for better efficiency
Also linked to (in the newsletter) last week, but probably the headline news: Ericsson has introduced a flurry of portfolio additions, including new RAN and antenna hardware, plus AI RAN software solution. These will integrate AI more deeply into its RAN systems, it said – to optimise performance and energy usage, plus to support greater uplink performance and better TCO.
Ericsson called it an “AI-first approach” to RAN hardware and software – with new software tweaks for beamforming, outdoor positioning, and “instant coverage prediction”, and new AI‑ready radios with Ericsson-made silicon with neural network accelerators. Again, everything will be on show at MWC. The new portfolio is structured around “three pillars”: RAN hardware (“10 AI-ready radios”), RAN software, and antennas (“five high-performing” ones).
It lists the improvements with each portfolio pillar, summarised as follows: expanded higher-power performance breakthroughs (RAN hardware); better beamforming, positioning, coverage prediction (as above), plus latency prioritisation (RAN software); and better spectrum utilization and site design with interleaved and passive units (antennas).
Mårten Lerner, head of networks strategy at Ericsson, said: “We are embedding AI solutions across our full portfolio, introducing AI RAN software solutions that deliver revolutionary improvements in spectral efficiency. We are now taking the final step to full AI enablement across our portfolio by introducing neural network accelerators in our leading Massive MIMO portfolio.”
8 | Enterprise 5G integrated into Windows 11
Another NIB, just because… well, it’s a laptop: Ericsson has a joint development with Microsoft to integrate advanced 5G directly into Windows 11 so enterprises “stand to benefit from secure, policy-driven laptop connectivity, simplifying how IT departments manage, protect and scale mobile PC fleets” – as Ericsson writes. Seven operators – including T-Mobile in the US, Telenor in Sweden, Singtel in Singapore, Softbank in Japan – have committed to “early launch” deals, it said, with broad availability from the second quarter of 2026
The joint solution combines Microsoft Intune device management with Ericsson Enterprise 5G Connect (formerly Ericsson Enterprise Virtual Cellular Network; EVCN). Embedding AI-based 5G management into Windows 11 means IT teams can automate how devices connect to and switch between networks, and apply in-house enterprise policies on top, said Ericsson. “This joint solution has the potential to transform enterprise IT and employee ways of working around the world,” said Ericsson.
9 | Agentic / generative AI rApp as-a-service on AWS
Further to the rApps work with Swisscom above, Ericsson has launched a new agentic rApp as-a-service solution for fast and flexible deployment of rApps on AWS. The solution introduces agentic AI capabilities for network optimization; it also adds a natural language generative AI interface. It uses an open service management and orchestration (SMO) architecture, and is presented as a “catalyst to achieve Level 4 network autonomy”.
The addition of generative AI means “communication service providers [can] effectively talk with the network, translating natural language into instructions for the agentic-AI powered system”, it said. Carriers can access the app on AWS’s application marketplace. Vivo Brazil, working with Ericsson and AWS, is the first to experience the real-world impact and the new capabilities, reckons Ericsson.
Ericsson said its AI solutions for network optimization currently handle more than 100 million AI inferences daily across 11 million cells serving more than two billion subscribers. James Crawshaw, practice leader at Omdia, commented “While many in the market now claim to provide rApps, only a small number have demonstrated successful, production-level deployments with ORAN-compliant interfaces. And no other vendor offers a comparable rApp-as-a-Service solution. These two aspects create a real distinction, setting this new offering apart.”
