Lenovo has launched a new GPU as a Service (GPUaaS) offering, positioning itself in a market that anticipates an increase in on-premises AI workloads. Alongside this, the company has enhanced its as-a-service portfolio with features aimed at AI, including automated management software and expanded liquid cooling services.
The latest innovations from Lenovo target AI workloads on-site. The newly introduced GPUaaS is an additional feature of the TruScale subscription-based infrastructure model, which allows customers to access GPUs on demand. Furthermore, Lenovo is incorporating automation, predictive analytics, and generative AI to enhance data protection in its management-as-a-service solution, XClarity One. Additionally, Lenovo has unveiled its Power and Cooling Services, which provide guidance on hardware integration.
As companies refine their AI applications, many are expected to implement these solutions within their own data centers. Steven Dickens, an analyst at Futurum Group, emphasizes this trend, stating, “Many organizations will prefer private AI use cases to keep their data and intellectual property off public clouds.” He believes Lenovo is well-positioned to support this movement, given its comprehensive focus on infrastructure, computing, and energy efficiency.
GPUaaS Overview
Lenovo’s TruScale GPUaaS features Nvidia GPUs, including the Nvidia H100 and Nvidia L40S, and offers usage-based resources via Lenovo Intelligent Computer Orchestration software for AI model development and training. This solution enables better management of GPU workloads and associated costs. Both Dell and HPE have long provided similar services, initially aimed at high-performance computing (HPC) customers but now shifting towards generative AI. For instance, Dell expanded its collaboration with Nvidia in March through its AI Factory, while HPE introduced HPE Private Cloud AI in June, integrating Nvidia GPUs with its hardware for private cloud solutions.
Lenovo, however, is distinguished by a more tailored approach to its GPUaaS offering, according to Matt Kimball, an analyst at Moor Insights & Strategy. He notes that Dell and HPE offer broader infrastructure solutions whereas Lenovo provides a more precise and detailed service. He believes heavily regulated industries, including healthcare and finance, will likely use TruScale GPUaaS to minimize costs by allowing multiple business units access to shared GPU resources.
Integrating AI in Management
Lenovo is not just focused on AI workloads; it is also integrating AI into its management offerings. The XClarity One platform now features predictive analytics to identify potential infrastructure failures before they occur, thus reducing downtime. Moreover, Lenovo is embedding generative AI into certain security functions to support continuous risk assessment and prevention. Kimball notes that the trend of incorporating AI for efficiency is not new, citing HPE’s acquisition of Morpheus, a hybrid cloud management platform known for its automation features.
In an environment where resources are limited, companies face increasing complexity, making automation a valuable asset for customers.
Liquid Cooling Innovations
AI and GPU usage can significantly increase energy consumption since GPUs require considerable power. Lenovo’s Neptune liquid cooling technology, already in its sixth iteration, has been supporting energy efficiency efforts for over a decade. The recent launch of their Power and Cooling Services enhances thermal management and energy needs, offering integration support for liquid cooling and data-driven insights for optimizing data center power requirements.
Dickens points out that liquid cooling, once confined to niche HPC and edge applications, is becoming more mainstream due to the demands of high-performance GPUs. While other vendors aim to catch up, Lenovo maintains a strong lead in this field, with a wealth of experience from its Neptune series. “With its sixth generation, Lenovo has valuable insights from previous iterations that benefit clients,” Dickens adds.
Adam Armstrong is a news writer for TechTarget, specializing in file and block storage hardware and private clouds, with previous experience at StorageReview.