Softbank Corp and Ericsson struck a deal to jointly explore the potential of shared hardware infrastructure for AI and RAN applications with the goal of opening new operator use cases from increased network efficiency.

The vendor explained the research will be around use of common network and compute infrastructure, which enables AI engines and RAN applications to run on the same hardware in a converged architecture.

As part of the research Ericsson and Softbank will conduct “techno-economic analyses, develop prototypes and run lab demos to optimise RAN and AI convergence at the edge” alongside assessing hardware partitioning.

The duo identified three key areas of focus: optimising AI and RAN convergence on the edge using Centralised RAN (C-RAN), including assessment of the pros and cons of the architecture; management of hardware and workload sharing between applications; and engineering demos.

Ericsson EVP and head of business area networks Fredrik Jejdling indicated the move was intended to “lead to new solutions that empower communications service providers to build more open, efficient and versatile networks”.

Both Ericsson and Softbank are part of the AI-RAN Alliance launched earlier this year, a working group which also counts giants including Nokia, Microsoft and T-Mobile US as members.