Kyndryl Partners With Nvidia To Up Its GenAI Game

‘Nvidia is the market leader today. They have an entire ecosystem that they bring to the table from the GPU to the infrastructure stack and to the application frameworks that will help to accelerate it. We are coupling our Kyndryl services with the Nvidia hardware and software to help clients with solutions independent of the industry sector they operate in,’ says Venkat Jagana, Kyndryl’s vice president and distinguished engineer.

Kyndryl Monday unveiled its first partnership with Nvidia in a move aimed at accelerating AI adoption for clients.

With the new partnership, Kyndryl is combining its infrastructure and IT expertise with Nvidia's hardware and software capabilities to expand its AI and GenAI capabilities beyond the cloud to on-premises, edge and hybrid infrastructures, said Venkat Jagana, vice president and distinguished engineer for AI platforms for New York-based Kyndryl.

Kyndryl currently offers AI and GenAI-related services via its Kyndryl Bridge platform for integrating and orchestrating complex IT infrastructures and turning applications and operations into as-a-service capabilities.

[Related: Kyndryl CEO: Customers Know ‘Their AI Is Only Going To Be As Good As Their Data’]

Kyndryl has over 30 years of managing complex infrastructures and mission-critical workloads and has built a partner ecosystem on top of that experience, Jagana told CRN.

“We will further bring in capabilities to provide end-to-end AI life-cycle services with Nvidia,” he said. “Nvidia is the market leader today. They have an entire ecosystem that they bring to the table from the GPU to the infrastructure stack and to the application frameworks that will help to accelerate it. We are coupling our Kyndryl services with the Nvidia hardware and software to help clients with solutions independent of the industry sector they operate in.”

Despite Kyndryl’s history with AI, the company has not had a direct relationship with Nvidia, Jagana said.

Before the Nvidia relationship, Kyndryl worked with hardware partners such as Dell Technologies, Lenovo, Hewlett Packard Enterprise and Cisco Systems to bring private cloud Nvidia-based solutions to clients, Jagana said. The company going forward will provide related managed services that take advantage of Nvidia’s DGX clusters, he said.

Kyndryl has provided cloud-native services around AI requiring GPUs in partnership with the top hyperscalers, including Amazon Web Services, Microsoft Azure, Google Cloud, Oracle and others, he said.

“Of course, those services all run in one way or another on GPUs for building models and things like that, along with Google TPUs, Nvidia DPUs, etc.,” he said. “But with this Nvidia partnership, we are taking Kyndryl services into the private AI space. It also covers hybrid and multi-cloud and [is] not necessarily restricted to the cloud or on-prem. Nvidia not only has the hardware, but they also have built an entire software ecosystem. We are working with Nvidia to provide the service capabilities around the Nvidia software ecosystem to our clients by bringing their hardware and software plus our services across on-prem, private cloud, hyperscaler and hybrid scenarios, including the edge. So our services will cover from the edge to the cloud to the data center.”

Unlike other capabilities available with Nvidia GPUs, the company’s software stack is highly optimized for its GPUs, Jagana said.

“By coupling with that optimized software stack, whether it is the NeMo large language model framework or NIM microservices, all of these things will help to further amplify and accelerate the whole AI journey for our clients and optimize and accelerate it further,” he said.

The new Nvidia relationship complements Kyndryl’s competitive advantages, Jagana said.

“Because of our Kyndyl Bridge platform, we are already providing day-one and day-two support for over 1,200 clients,” he said. “We can use the Nvidia technology to accelerate the GenAI capabilities in the Kyndryl Bridge platform to support day-two type of services from incident resolution and other areas with a faster response time. That IP is a differentiating factor in the marketplace.”