Mavenir partners with Neysa to sell AI services on GPU-backed cloud

Mavenir, which says its software is deployed with more than 300 operators in over 120 countries, announced a partnership with Neysa on 22 September to run its AI Integrated Platform on Neysa’s GPU-backed cloud.

Mavenir adds billing to Neysa’s GPU cloud

Neysa supplies the GPU capacity, cloud stack and deployment environment. Mavenir says its AI Integrated Platform supplies model orchestration, agent workflows, security, policy controls and metering by token. The companies are aiming to let providers charge for AI services delivered over that infrastructure.

For neocloud providers, Mavenir says the arrangement is designed to package GPU capacity into services for enterprise customers. Its billing tools would track token consumption, while Mavenir supplies policy controls for the offering.

Bejoy Pankajakshan, Mavenir’s Chief Technology & Strategy Officer, said the arrangement would give neocloud providers a way to sell AI as a service rather than infrastructure alone. Neysa’s Chief Revenue Officer, Xavier Kurian, said Mavenir’s choice of Neysa reflected demand from enterprises for governed AI infrastructure with usage metering and cost visibility.

Neysa is a Mumbai-based AI cloud company backed by Blackstone, Business Standard reported. The partnership also targets enterprises seeking to run AI on infrastructure they control. Mavenir says the combined platform is designed to put model management, security and usage tracking on top of that infrastructure without customers assembling those components separately.

Operators get a Neysa deployment option

Mavenir’s existing AI Integrated Platform is intended to let mobile operators host models, run agents and charge for token consumption on their own infrastructure, according to the company. The Neysa partnership adds a GPU-backed cloud option alongside deployments on equipment an operator already owns.

Mavenir says the combined system can be deployed on-premises or across hybrid environments. That gives operators and enterprise customers a choice of deployment settings for the same orchestration and metering tools. The proposed uses differ by customer: operators would offer AI services over their networks, while neoclouds would package their GPU capacity for sale.

The operator pitch draws on Mavenir’s existing telecom business. The company says its software is deployed with operators across more than 120 countries. Neysa’s role is to provide cloud infrastructure and GPU capacity for customers choosing that route, while Mavenir supplies the software used to manage and bill for AI workloads.

Indian telecom operators are also developing cloud and data centre infrastructure for enterprise customers, Business Standard reported. That places the partnership alongside operators’ own efforts to supply computing capacity, while giving Mavenir a cloud deployment route for its AI software.

Mavenir tests its products on Neysa

Mavenir plans to develop, test and scale products on Neysa’s infrastructure. They include AI Service Assurance for automated network operations, AI Security Agents for fraud and threat detection, and AI Voice Services for uses including real-time translation and virtual agents for small businesses.

The company says it also runs its own AI workloads on Neysa’s cloud. Those deployments give Mavenir a use for the GPU infrastructure beyond offering the combined platform to operators, enterprises and neocloud providers.

Mavenir says its token optimiser and model-routing tools have reduced its spending on frontier models in its own deployments. The tools send each request to the model Mavenir’s routing system considers capable of handling it most cost-effectively. Pankajakshan said Mavenir could pass those savings on to customers using Neysa.

Mavenir says its internal work on Neysa’s cloud also includes AI-assisted software development and intelligent model selection.