Cloud computing orchestration startup Avesha Inc. has partnered with the Linux software giant SUSE SE on an enterprise-grade artificial intelligence infrastructure blueprint that’s designed to simplify and optimize AI workload deployments.

Announced today, the AI infrastructure blueprint combines SUSE’s specialized AI development software with Avesha’s elastic graphics processing units, enabling GPU resources to be allocated to AI workloads dynamically, so they can be optimized for both performance and costs, the companies said.

Using the blueprint, enterprises will be able to deploy, manage and monitor their AI workloads on a production-grade stack, paving the way for “self-service AI” for any team that needs it.

Avesha has created a technology platform that’s used by companies to optimize traditional application workloads across any cloud environment, including Amazon Web Services, Microsoft Azure, Google Cloud and Oracle Cloud Infrastructure. It supports multicloud and edge locations, and uses AI to manage and scale cloud applications efficiently, ensuring top performance without bloated costs.

Users can oversee their applications via Avesha’s Smart Network Application Platform, which creates a virtual overlay of the infrastructure for each application, providing detail into its network, the virtual machines that power it, and other services. With the Elastic GPU service, users have the option to quickly access and dynamically allocate GPUs through a no-code interface to enhance AI workloads.

Meanwhile, Germany-based SUSE is the creator of the popular SUSE Linux operating system, which is considered to be one of the most flexible and reliable platforms of its kind, used by enterprises to run cloud instances, industrial equipment and applications, and various other systems. SUSE also develops other products, such as the Rancher platform for managing software containers that host the components of modern applications, and the NeuVector cybersecurity service.

The Avesha + SUSE AI blueprint is being made available to all customers today, and targets industries including healthcare, financial services, manufacturing, government and telecommunications, where robust governance for GPU-based AI workloads is critical. Customers can use the SUSE AI platform to design and build AI applications by integrating frameworks such as Ollama, MLFlow and PyTorch. It’s built on Rancher Prime, which supports GPU-aware Kubernetes management and boasts tools for observability and compliance, the companies said.

Blueprints such as this one can be useful for enterprises, because they provide a structured framework that simplifies the process of piecing together AI infrastructure that’s reliable, scalable, optimized for performance and cost efficient. As such, teams can focus more on building and deploying their AI projects, without worrying about the overheads. They’re becoming popular with infrastructure providers, with other examples being Nvidia Corp.’s and Nutanix Inc.’s blueprint for advanced AI agents, and AWS’ three-layered gen AI stack.

Avesha Chief Executive Raj Nair said the blueprint simplifies GPU orchestration, which is one of the most complex aspects of AI infrastructure. This means companies can eliminate underutilized GPU resources in order to save on costs. Other benefits include project- and team-level isolation and zero-trust container runtime protection.

“SUSE AI gives enterprises the choice to use the right tools to innovate with confidence,” said Abhinav Puri, SUSE’s vice president and general manager of Portfolio Solutions. “Our collaboration with Avesha brings security, scalability and simplicity, making enterprise-grade AI infrastructure accessible to any team.”

Photo: SUSE SE