1Finity Proposed Monetizing Idle Network Compute

The Fujitsu company is encouraging network operators to lease underutilized radio access compute for enterprise AI tasks.

Updated on Oct. 5, 2026 in Artificial Intelligence

Isometric editorial illustration of a modular edge computing hardware enclosure featuring vertical cooling fins, colored in muted teal and oxblood.
Fujitsu's 1Finity is promoting a new model for network operators to monetize idle radio access computing capacity for enterprise AI processing. AI Illustration. Upload story photo >

Live Poll

Do you trust telecommunications companies to provide reliable enterprise AI services alongside their network connectivity?

1Finity has outlined a strategy for network operators to generate revenue by selling excess AI-RAN compute capacity to enterprise customers. The initiative utilizes policy-driven orchestration to ensure enterprise AI workloads do not compromise core radio access network performance.

Why it matters

Network operators currently struggle to justify costly infrastructure investments based on connectivity performance alone. This model aims to turn underutilized edge compute headroom into a new revenue stream, challenging the dominance of traditional hyperscalers.

The system utilizes Armada software for resource management and monitors specific RAN key performance indicators. A demo successfully identified shoplifting by chaining CCTV feeds, Open RAN-compliant radio hardware, and 3 NVIDIA GPUs.

The players

1Finity

A Fujitsu company that specializes in network infrastructure solutions and AI-RAN integration.

Aible

A technology partner that provides standard AI application agents for use within enterprise environments.

Armada

A software provider that develops resource management tools designed to monitor performance metrics within radio access networks.

NVIDIA

A leading technology company that manufactures the graphics processing units used to power the demonstrated AI compute workloads.

The details

Operators are encouraged to deploy AI-RAN services through self-service models, allowing enterprises to customize applications without direct operator intervention. By leveraging existing last-mile connectivity and distributed edge sites, companies hope to compete effectively with cloud hyperscalers for AI processing tasks.

Timeline

  1. October 5, 2026: The initiative was formally presented at the Intelligent RAN Forum.

The Tech Race

This strategy signals a shift from traditional telecommunications toward AI-native public networks that directly challenge the compute dominance of cloud hyperscalers. It leverages unique advantages in spectrum and distributed edge sites to position operators as central hubs in the AI ecosystem.

Enterprise customers may soon access powerful AI processing capabilities through their existing network providers rather than relying solely on centralized cloud vendors. This could lead to more localized and responsive AI applications that leverage the proximity of network edge sites to the end user.

The takeaway

Operators possess a strategic advantage in physical edge presence that could decentralize AI workloads away from a few massive cloud providers. Businesses should monitor whether these pilot demonstrations evolve into standardized, easily accessible service offerings.

Further reading

For more background on the integration of computing and connectivity, explore our Artificial Intelligence section.

Live Poll

Do you trust telecommunications companies to provide reliable enterprise AI services alongside their network connectivity?