Two new systems are already available today. In the long term, even more powerful hardware is expected to come onto the market.
During Lenovo Tech World, Nvidia and Lenovo announced that they would be working more closely together. The goal is to bring more advanced hybrid computing platforms to market that focus on generative AI workloads, regardless of where they are being worked on.
With the new collaboration, Nvidia CEO Jensen Huang and Lenovo CEO Yuanqing Yang want to offer a solution to the growing demand for faster systems for AI software and other “expert services”.
In addition to the two systems already available today, both parties are working on even more powerful hardware based on the flexible Nvidia MGX reference design. The hardware is available in tandem within TruScale, Lenovo’s IaaS model.
There are two systems available today
The Lenovo ThinkSystem SR675V3 combines Nvidia L40S GPUs with BlueField 3 compute units and Spectrum-X networking capabilities to enable AI workloads in any type of data center.
In addition, the Lenovo ThinkStation PX is now available, making additional performance available everywhere, including at the edge. The system can be configured with up to four Nvidia RTX 6000 GPUs.
Both systems use the Nvidia Nemo framework, where developers can find various LLM models to get started.