Extreme Networks is previewing a technology that will let enterprise customers run AI applications locally using a mix of on-premises compute and spare network equipment processing capacity.
With Extreme Exchange zones, enterprises will be able to pool compute resources, including spare capacity on Extreme access points and switches already deployed at branch offices and other edge sites, to run AI applications and plug-ins, the vendor says.
Extreme Exchange zones are designed to address customer challenges around data sovereignty and control by enabling them to run AI applications on premises and in the cloud, wherever their data originates and workloads run, with full control over data residency and processing, according to Nabil Bukhari, CTO and president of AI platforms at Extreme.
“We’ve spent decades building networks that reach every corner of the enterprise, and those networks hold a lot of idle compute. Exchange zones will put that compute to work, letting AI run locally on infrastructure customers already own. That delivers sovereignty, local inferencing, and better economics than sending everything to the cloud. Customers control their data, inference, and cost inside a boundary they define, with a record of every exception,” Bukhari said in a statement.
A zone makes use of data from Extreme’s Universal Compute Platform—or a qualifying customer-owned server—with spare compute capacity from deployed access points and switches. Extreme says Exchange zones will pool these resources to run AI applications and plug-ins closer to where data is generated. The zones also draw on data from other Extreme network components, such as its Platform ONE management and security package and Extreme AI intelligence core.
Extreme’s systems generate operational data that is valuable beyond network management, from security and business operations to logistics, the company stated. “Exchange zones will bring that data on premises, unlocking use cases that cloud-only architectures can’t yet deliver. Models run locally by default, and cloud inference is available only when customer policy allows it and is logged whenever it occurs,” Extreme stated.
According to Extreme, cloud infrastructure is the right home for many AI workloads, but not all: “Some applications can’t tolerate the delay of a round trip to the cloud, some generate more data than it’s economical to move, and regulated industries and distributed operations often need tighter control over where AI runs. Exchange zones will address this gap by running Exchange applications and their models inside a boundary the customer controls.”
Exchange zones build on Extreme Exchange, the vendor’s AI marketplace where it offers a catalog of AI tools, agents and applications. Exchange is designed to give enterprise customers a way to discover, deploy, and create AI agents, microapps, and workflows rather than developing such components on their own, the vendor says.
Because the platform is in preview, the company did not detail which Extreme hardware is involved in Exchange zones or how it determines available processing capacity.