Lenovo launched a GPUaaS providing, getting into a market that is getting ready for patrons to carry AI workloads on premises. It additionally added different elements of its as-a-service providing with AI in thoughts, together with automation options in its administration software program and expanded providers for liquid cooling.
Lenovo is rolling out three developments geared toward AI workloads on premises. GPU as a service is an added characteristic to the corporate’s TruScale, a subscription-based infrastructure mannequin, giving clients entry to GPUs on demand. It is also including automation, predictive analytics and generative AI for knowledge safety to XClarity One, its management-as-a-service providing for Lenovo’s infrastructure. As well as, Lenovo unveiled its Energy and Cooling Companies, together with {hardware} integration advisement.
As AI workloads and use instances proceed to get refined, many shoppers are going to look to deploy them in their very own knowledge facilities, in response to Steven Dickens, an analyst at Futurum Group.
“Lots of people are going to make use of non-public AI use instances the place they do not need all their knowledge, their IP, going to the general public cloud,” he stated.
Lenovo is offering the means for this with no gaps in what they’re doing, given its concentrate on infrastructure, compute and vitality consumption, Dickens stated.
GPUaaS
Lenovo TruScale GPUaaS is a brand new characteristic for Lenovo. The service makes use of Nvidia GPUs, together with Nvidia H100 and Nvidia L40S, and provides metered sources by Lenovo Clever Pc Orchestration, which is software program for AI mannequin growth and coaching. With the software program, GPU workloads and prices could be managed, Lenovo stated.
Dell and HPE each have supplied comparable providers for years. Initially, Dell and HPE targeted on high-performance computing (HPC) clients, however that has modified with the onset of generative AI. In March, Dell expanded its partnership with Nvidia and launched its AI Manufacturing facility, a mixture of infrastructure, associate ecosystem {and professional} providers geared toward serving to clients with AI workloads. In June, HPE launched HPE Personal Cloud AI, which additionally companions with Nvidia to mix GPUs with HPE {hardware} and software program into a personal cloud for patrons.
Lots of people are going to make use of non-public AI use instances the place they do not need all their knowledge, their IP, going to the general public cloud. Steven DickensAnalyst, Futurum Group
However Lenovo provides a stage of specificity past the Dell and HPE choices, in response to Matt Kimball, an analyst at Moor Insights & Technique.
“Lenovo is extra surgical of their strategy of their GPU-as-a-service providing,” he stated.
For AI and GPU utilization, Dell sells infrastructure and associate constructing blocks that clients must assemble, whereas HPE sells preconfigured choices in numerous sizes. Nevertheless, Lenovo is providing a extra fine-grained strategy, Kimball stated.
Closely regulated industries akin to healthcare, monetary providers and insurance coverage will use AI fashions on premises and can use TruScale GPUaaS to decrease prices, Dickens stated.
“As a substitute of paying by the sip, Lenovo allows 5 – 6 enterprise items to make use of the GPUs,” he stated.
Including AI to administration
Lenovo is not simply fascinated by AI use; it is also utilizing AI in its XClarity One management-as-a-service providing. It added predictive analytics to establish infrastructure points earlier than they fail in an effort to scale back downtime. Lenovo can be embedding generative AI into a few of its safety capabilities to conduct steady studying for threat prevention, in response to the seller.
Including AI to software program for effectivity is not new within the enterprise infrastructure market, Kimball stated. For instance, HPE ramped up including AI and automation by its acquisition of Morpheus, a hybrid cloud administration platform with automation capabilities.
“They’re driving a extra self-driving, self-healing, greater-protected infrastructure by AI,” Kimball stated.
With fewer folks and decrease budgets, complexity turns into the enemy, whereas automation turns into a profit to clients, he stated.
Liquid cooling
AI and GPU utilization additionally means elevated vitality consumption, as GPUs take a number of vitality to run. Lenovo’s Neptune liquid cooling know-how has been in use for over a decade and is presently in its sixth technology. This week, Lenovo launched its Energy and Cooling Companies program that helps vitality and thermal wants by the usage of Neptune. The brand new providers assist in {hardware} and facility integration for liquid cooling, in addition to data-driven insights to higher decide the information heart’s energy wants.
Liquid cooling has moved from extra area of interest HPC and edge use instances to the mainstream now that power-hungry GPUs are within the combine, in response to Dickens. Different distributors need to catch up, however Lenovo has a big lead in growth.
“In its sixth technology, there’s most likely good battle scars from the earlier 5 generations [of Neptune liquid cooling] which can be beneficial to shoppers,” Dickens stated.
Adam Armstrong is a TechTarget Editorial information author overlaying file and block storage {hardware} and personal clouds. He beforehand labored at StorageReview.
