Vodafone and Ericsson have successfully demonstrated network-native AI for real-time noise cancellation and language translation during mobile calls. The trial utilized Ericsson’s IP Multimedia Subsystem and Vodafone’s application server to process AI services directly within the network without requiring specialized hardware or apps.
This demonstration signals a shift toward programmable networks where high-compute AI tasks like translation occur at the infrastructure level rather than on the device. For the streaming and communications ecosystem, this validates that low-latency AI processing can be handled natively, potentially reducing the hardware requirements for end-user devices while increasing the value of carrier-grade IMS architectures. As telecom providers seek new revenue streams, these intelligent network services offer a path to differentiate standard voice and data products against over-the-top applications. Watch for whether Vodafone moves this from lab conditions to a commercial rollout across its European footprint to gauge consumer demand for network-integrated AI.
Ericsson has been building out its network-native AI portfolio well beyond the Vodafone voice translation demo. In March 2026, Nokia and Ericsson announced a landmark collaboration to advance intelligent automation across Open RAN and cloud RAN networks, with Ericsson joining Nokia's SMO Marketplace and Nokia becoming a member of Ericsson's rApp Ecosystem. The mutual ecosystem membership centers on the R1 interface through which rApps interact with the Service Management and Orchestration layer, giving communication service providers greater choice as they pursue Level 4 autonomous network targets.
The competitive positioning between Ericsson and Nokia on network AI has become increasingly explicit. Analysts noted that Nokia and Ericsson fundamentally disagree on how and where to deploy AI in the mobile network, with Ericsson reinforcing its custom ASIC strategy and energy-efficient Layer 1 acceleration while Nokia aligns its baseband roadmap with Nvidia platforms. Ericsson's approach places AI across multiple network locations from radio to baseband to RAN aggregation to core, whereas Nokia positions RAN as a distributed compute layer. This architectural divergence directly shapes how each vendor approaches use cases like the Vodafone IMS-based voice translation demo.
Ericsson's Intelligent Automation Platform has grown substantially as the foundation for these network AI services. The company reported that its rApp ecosystem now includes around 105 to 106 members with more than 100 rApps, and several CSPs are developing and running their own rApps in production. Ericsson is also evolving the platform to natively support agent-based capabilities by exposing MCPs, knowledge, tools, and planners that allow AI agents to integrate with the platform for optimizing both RAN and core networks. This agent-ready architecture provides the operational substrate on which services like real-time voice translation and noise cancellation can be orchestrated at scale.
Ericsson and Vodafone have successfully demonstrated real-time AI voice translation and noise cancellation processed directly within mobile network infrastructure. By utilizing Ericsson’s IP Multimedia Subsystem and Vodafone application servers, the trial proves that high-compute AI tasks can be handled natively, potentially reducing reliance on specialized device hardware for advanced communication features.
The translation and noise cancellation are processed natively within the mobile network using Ericsson’s IP Multimedia Subsystem and Vodafone application servers, rather than relying on third-party apps or specialized phone hardware.
It signals a shift toward programmable networks where high-compute AI tasks occur at the infrastructure level. This approach could reduce hardware requirements for end-user devices and allow telecom providers to differentiate their voice services.
It serves as the foundation for network AI services, supporting over 100 rApps and evolving to include agent-based capabilities that help orchestrate services like real-time voice translation at scale.
Ericsson focuses on custom ASIC strategies and energy-efficient Layer 1 acceleration across multiple network locations, whereas Nokia aligns its baseband roadmap with Nvidia platforms and positions RAN as a distributed compute layer.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source
