Mavenir and Neysa Partnership links GPUs with AI

AI infrastructure is moving beyond the question of how much GPU capacity an organisation can buy. The bigger challenge is turning that capacity into AI services that can be deployed, governed and tracked. The Mavenir and Neysa partnership is aimed at that gap by combining AI orchestration with GPU infrastructure and a customer-ready AI cloud.

The partnership brings Mavenir’s AI orchestration, agent and token-metering capabilities together with Neysa’s AI-native GPU infrastructure and deployment environment. The result is a stack designed for enterprises that want to run AI on infrastructure they control.

AI-native infrastructure gets a software layer

Neysa provides the GPU capacity, AI cloud stack and deployment environment, while Mavenir adds orchestration, agent workflows, security, policy control and token-level metering.

That combination changes the role of the GPU layer. Instead of asking IT teams to assemble multiple components, the platform is designed to package computing capacity into governed and billable AI services.

For neocloud providers, this creates another use case. Raw GPU capacity can be turned into ready-to-sell AI services without building the complete platform layer themselves.

Sovereign AI cloud brings control closer

The partnership also targets enterprises that want more control over where AI workloads run. The platform can be deployed on-premises or across hybrid environments, while Neysa provides the GPU-backed environment.

This approach puts the focus on sovereign AI cloud capabilities, with model orchestration, security, usage tracking and policy control sitting alongside the underlying infrastructure.

For mobile operators, the platform also extends Mavenir’s existing AI strategy by providing another GPU-backed route for hosting models, running agents and monetising token consumption.

Mavenir tests its own AI stack first

Mavenir plans to develop, test and scale AI products on Neysa’s infrastructure, including AI Service Assurance, AI Security Agents and AI Voice Services.

The company is also using the same environment for internal AI initiatives such as AI-assisted software development and intelligent model selection. This creates a “run it on itself first” approach before extending the stack to operators, enterprises and neocloud providers.

Cost is another part of the equation. Mavenir’s token optimiser and model-routing tools are designed to direct requests towards models that can handle them at a lower cost, with the company saying its own deployments have reduced frontier model spend.

Read More: 

Googlebook Intel Core Ultra Series 3 launch brings AI laptops

Microsoft Data Centre expansion signals a regional shift

Commvault SHIFT 2026 to focus on AI, cyber recovery and resilience

Lalit Choudhary Takes Charge as ISODA President