Linux Foundation puts AI workloads and autonomous networks on one Tokyo agenda
Arpit Joshipura's ONE Summit Japan program pairs AI-run network operations with the infrastructure needed to carry AI workloads, from edge systems to data centers.
By Ryan Merket · Published
Primary source: Linux Foundation Newsroom
Why it matters
The program puts AI's networking demands and AI-based network automation in the same room, but an agenda is an industry priority map, not proof that autonomous networks or AI-RAN are ready for production.
The Linux Foundation's networking group has put AI-driven network operations and the infrastructure that carries AI workloads on the same agenda for ONE Summit Japan 2026, scheduled for December 10th and 11th in Tokyo. The initial program spans agentic AI, autonomous networks, open-source radio access networks, 6G, edge computing and data-center systems.
For Arpit Joshipura, the Linux Foundation's general manager for Networking, Edge and IoT, the lineup reflects a long-running effort to get companies and contributors to build shared networking technology. Joshipura joined the Foundation in 2016 after holding strategy, product, marketing, engineering and standards roles at Prevoty, Dell/Force10, Ericsson/Redback, ONI/CIENA and BNR/Nortel. The Linux Foundation says his career has covered more than 25 years of networking work across carriers, enterprises and cloud infrastructure.
The distinction in this year's agenda is that AI appears in two roles: as a demanding workload networks must support, and as software intended to help operate those networks. In its October 1st announcement, LF Networking lists agentic and intent-based networking, multi-agent collaboration and autonomous operations alongside edge AI infrastructure, digital twins and physical AI.
That pairing is the most useful way to read the program. Data-center and enterprise sessions are expected to address AI network operating systems, SONiC and DPUs and IPUs, while telecom-oriented topics include AI-RAN, GPU and NPU acceleration, cloud-native operations and next-generation wireless. The announcement also names network APIs, including CAMARA, and approaches to exposing and monetizing network capabilities.
A program built around two sides of the same infrastructure
AI training and inference place demands on networking capacity and architecture. At the same time, operators are exploring software that can automate network configuration, monitoring and response. The summit's initial topics bring both concerns together, from AI infrastructure in data centers to autonomous operations across telecom networks.
The program is community-selected, according to the Linux Foundation. That makes it a snapshot of what contributors and industry participants want to discuss, rather than evidence that the listed technologies have reached broad commercial deployment. The announcement does not tie its themes to named operator commitments, customer deployments or performance results. Nor does it establish that agentic network operations are ready to act autonomously in production.
The project names provide a more grounded view of the work being assembled. Nephio and OpenAN appear among the telecom operations topics; SONiC, an open network operating system, appears under AI data-center networking; and OCUDU is listed with open RAN and next-generation wireless. RuntimeWire reported last month on GNU Radio joining OCUDU, adding a software-defined radio project to the open-source RAN initiative. That earlier development gives the Tokyo agenda a concrete connection to ongoing work, though it does not demonstrate that a production network will result.
Joshipura described the program as a shift from agentic AI and autonomous operations through 6G, open RAN and network APIs. His view comes from a career spent across telecom and enterprise networking, including executive roles at both startups and established vendors. In 2016, when he joined the Linux Foundation, he described open source as a way for vendor and user communities to shape networking technology together. The 2026 agenda applies that same shared-development approach to AI-era networking.
A conference agenda, not a deployment report
ONE Summit will run alongside Open Source Summit Japan. The event page positions it for communications, cloud and telecom providers, as well as enterprises that depend on networking across core, access and edge infrastructure. Its stated program includes sessions, a solutions showcase and project mini-summits across the two days.
The scope is broad: agentic systems, MCP, 6G, non-terrestrial networks, AI-RAN, private networks, AI data-center fabrics and API monetization all sit under one conference. That breadth can help participants connect projects that are often discussed separately. It also makes the actual sessions important: an agenda category alone cannot show whether a topic will produce deployable software, an operator commitment or a commercial business model.
The Linux Foundation has announced an initial session lineup, with the event still more than two months away. For Joshipura and LF Networking, the near-term task is convening contributors and operators around common technical problems. The longer test is whether the shared work discussed in Tokyo can translate into interoperable systems that network operators will deploy and maintain.