Nutanix Updates Gen AI System, Partners with Dell

NUTANIX NEXT — Nutanix on Tuesday upgraded its GPT-in-a-Box generative artificial intelligence (AI) system, making the announcement at its annual user conference in Barcelona. Channel Futures is there.

Nutanix also took steps to make it easier for hyperconverged infrastructure (HCI) customers to switch from VMware, including signing a partnership with VMware’s former parent company Dell.

The moves highlight Nutanix’s two big pushes at Next — to take on virtualization giant VMware and increase its value for generative AI adoption.

GPT-in-a-Box 2.0 follows the initial version of the product Nutanix rolled out last year. GPT-in-a-Box 1.0 included the Nutanix Cloud Infrastructure platform, Nutanix File Storage and Object Storage, AHV, Nvidia GPUs and tools to run machine learning with Kubernetes.

Version 2.0 adds support for Nvidia NIM microservices for using large language models (LLMs) with Nvidia GPUs. Nutanix also added foundation model management, API endpoint creation, end-user access key management, a built-in graphical user interface, role-based access control, auditability and dark site support, among other benefits. Nutanix also now allows customers to use GPT-in-a-Box to run LLMs from Hugging Face’s library.

The new AI partner program will expand Nutanix’s ecosystem of AI partners to help build and run generative AI apps on the Nutanix Cloud Platform and GPT-in-a-Box. AI use cases supported by Nutanix include chatbots, document search and analysis, and code generation.

Related:Nutanix Takes Aim at VMware, Gen AI at Next Conference

Nutanix is adding a Launch Pad starter kit program giving channel partners a preconfigured software SKU to help them quote prices faster.

Furthermore, Nutanix is making news on the hyperconverged infrastructure (HCI) front, with its Dell deal and support for hardware used with VMware Ready Node deployments. Ready Nodes are servers sold by vendors such as Cisco, Dell, HPE and Lenovo that run VMware hypervisors and vSAN. Nutanix’s Ready Node support aims to help VMware customers run Nutanix on the same hardware as VMware.

The Nutanix-Dell partnership includes two offerings under joint development scheduled for early access availability in late 2024. Dell will sell an integrated turnkey HCI appliance running Nutanix Cloud Platform on Dell PowerEdge servers, and a Nutanix Cloud Platform for Dell PowerFlex includes the Nutanix AHV hypervisor for compute with independent PowerFlex storage. Dell and Nutanix will collaborate on engineering, go-to-market support, and services.

The Nutanix-Dell deal is another shot at VMware, which is used in the Dell VxRail HCI system on PowerEdge servers. Dell was an early Nutanix partner, but the relationship cooled after Dell bought VMware as part of its $67 billion EMC acquisition in 2015.

“We have been partners with Dell since 2014, early on in the history of Nutanix,” Nutanix CEO Rajiv Ramaswami said, suggesting the name “NxRail” for the new Dell turnkey appliance.

Nutanix also broadened its Cisco partnership, allowing its software stack to run on Cisco X Series UCS servers. That expands the Cisco Compute Hyperconverged with Nutanix platform launched last year.

Lee Caswell, Nutanix SVP of products and solutions market, said the launches focus on three areas. “Customers are thinking how they can adopt Nutanix quickly, how they can modernize to be future ready – particularly for containers and generative AI – and how they can think about making infrastructure do more with the same staffing,” he said.

Good Timing for GPT-in-a-Box 2.0?

Steve McDowell, founder and chief analyst at NAND Research, said Nutanix was a bit ahead of the market when it launched GPT-in-a-Box 1.0. He said he expects the company will gain more traction in the market with the upgrade.

“I think they rolled out version one of GPT-in-a-Box a little early, because this is the year when enterprises are starting to look at generative AI,” he said. “Generative AI is so complex and has so many moving pieces, and nobody understands how this all works. So the more bunding you can do, the better. With Nvidia NIM, that’s a containerized approach. That’s going to take away a lot of pain for Nutanix customers. It’s like a shortcut.”

Thomas Cornely,  Nutanix SVP of product management, agreed that GPT-in-a-Box is a shortcut to generative AI.

Nutanix's Thomas Cornely

“The whole idea is, how do we help customers deploy those AI models on premises and basically get an on-prem AI inference endpoint? You can do it yourself, which is very hard. So how can we give you a software stack that builds on infrastructure basically out of the box? Or, in this case, 'in the box,'” Cornely said.

Nutanix Chasing VMware Converts

Another theme at Next is Nutanix’s attempt to make it easier for VMware customers to switch. That applies to customers who would be new to Nutanix, or those using Nutanix HCI with VMware virtualization.

Cornely said it's easy to convert VMware ESX and vSphere virtualization to the Nutanix AHV hypervisor, and from vSAN to Nutanix HCI. But for customers running three-tier architectures with separate compute and storage infrastructure, “there is more work to be done. We have our own functionality that is bundled with our core platform. If customers are just using vSphere, it’s a very simple transition [to Nutanix AHV]. And hyperconvergence on VMware to hyper-convergence on Nutanix is a simple conversion," said Cornely.

Other Nutanix Next news includes: