Lenovo leans into SAN, hypervisor multiplicity, Azure Local and AI hyperconvergence

Lenovo is introducing a new ThinkSystem SAN array product line along with three ThinkAgile hyperconverged infrastructure (HCI) systems for VMware, Nutanix, Azure Local and AI.

Stuart McRae, Executive Director and GM of Data Storage at Lenovo, said: “With disruptions in enterprise virtualization strategies and the mandate to make their data ready for the most demanding workloads, organizations are looking to modernize their legacy infrastructure with open solutions.”

He’s referring to Broadcom’s acquisition of VMWare and its subsequent licensing changes opening the doors to organizations wedded to VMware so they can look elsewhere. They may choose different hypervisors, or container-based systems, for different workloads instead of looking for a single VMware replacement. The “demanding workloads” reference includes block-based (SAN) workloads held back by disk drive-based storage as well as AI inferencing-type applications. Lenovo has systems for both areas.

As well as the ThinkSystem DS Series for SANs, there are two ThinkAgile FX Series for VMware and Nutanix hypervisor systems. The MX system provides disaggregated external Fibre Channel storage for Azure Local. Lenovo’s HX Seres is dedicated to AI workloads with support for an Nvidia Blackwell GPU. The company is also launching services to advise and help customers select from, implement and manage these systems.

The DS Series is Lenovo’s all-flash SAN storage array offering, and there are four products; the DS3200, 5200, 5200C, and 7200. The DS3200, 5200 and 7200 use TLC NAND while the capacity-optimized 5200C is fitted with QLC drives. A table lays out their main characteristics;

These are based on OEM’d NetApp ASA (All-flash SAN Array) systems. As we understand it, the ASA A20 becomes the DS3200, the ASA A30 underlies the DS5200, the ASA A50 is the base of the DS7200, and the ASA C30 is used for the DS5200C.

Lenovo is taking the low-end and mid-range ASAs but not the high-end A70, A90, and A1K models. We speculate that this is to avoid competition with the forthcoming to-be-acquired infinidat arrays, but Lenovo couldn’t comment on this; the acquisition has not yet closed.

There are two ThinkAgile FX Series products; the 1 RU FX630 V4 and larger, 2 RU, FX650 V4;

They can use either VCF (VMware Cloud Foundation) or Nutanix AOS software with its AHV hypervisor, and scale out from 3 to 32 nodes. It means that customers can use them for existing VCF implementations and then migrate to AOS if and when they choose without needing to refresh their hardware.

A further HCI offering is the MX Series which provides Fibre Channel-connected external or disaggregated block array storage for Azure Local deployments. Azure Local is the on-prem, private cloud version of the Azure public cloud, running Azure VMs and containers on bare metal server hardware provided by Microsoft partners such as Lenovo.

The MX650a V4 HCI system supports Nvidia’s RTX Pro 6000 GPU, a Blackwell system, and so can run AI inferencing-type workloads.

Lenovo’s final HCI system is the updated HX Series, using Nutanix’ AI software stack, for running AI workloads in a secure and turnkey system.

We asked McRae if Lenovo was developing AI pipeline stack software to enable its systems to feed data to AI models and agents. He said not, Lenovo wanting to provide the best storage foundation for such pipelines but not to go into the pipeline software area itself. That was for partners.

All these systems will be gratefully received by Lenovo’s channel. Lenovo is second only to Dell as a primarily x86-based PC, servers and storage business. It operates in 180 countries and, with its latest $20.45 billion quarterly earnings, more than 2x HPE’s latest $9.7 billion quarter and meaning a near $82 billion run rate, and profits of $512 million, is a formidable business. Dell by the way registered more than $1.5 billion in profits for its latest quarter.