At the VMware Explore 2023 event, VMware announced that it will once again work with NVIDIA to launch the new VMware Private AI Foundation With NVIDIA service to help enterprises more quickly introduce automatically generated artificial intelligence technology application resources while ensuring the privacy, security and controllability of application content data.
With the recent trend of automated artificial intelligence applications, more and more companies are beginning to observe the development potential of this technology while also assessing the associated risks. While artificial intelligence applications can bring many benefits, data privacy, security, and application controllability have also become issues of concern to companies. This is especially true when applying artificial intelligence technology to financial, medical, retail, telecommunications, media, and other services. The security of customer information and other data must be given special attention.
VMware and NVIDIA have partnered to launch the VMware Private AI Foundation With NVIDIA service, which integrates NeMo, an artificial intelligence model from NVIDIA AI Enterprise's technical resources, allowing enterprises to build automatically generated artificial intelligence at any endpoint. They can also use customized model frameworks to quickly build artificial intelligence application services in a more secure manner.
This service is also built on VMware Cloud Foundation, making it easier for enterprises to deploy self-generated AI applications in a cloud-native manner while ensuring data privacy and access security. In addition to NVIDIA NeMo, this service also offers the option to utilize Meta's open-source AI model, Llama 2, giving enterprises greater flexibility in their deployment options.
As for the in-depth cooperation with NVIDIA, VMware can also accelerate the efficiency of automatically generated artificial intelligence computing in specific service deployment environments through virtualized GPU resources. At the same time, by maximizing the use of virtualized CPU, GPU and DPU computing resources, it can help reduce service construction costs. Paired with VMware vSAN's fast storage architecture, it can also improve service data access performance. It can even directly bypass the CPU computing process and directly transfer data to the GPU for completion through NVIDIA's GPUDirect RDMA technology.
This collaboration further integrates VMware's vSphere virtualization platform with NVIDIA NVSwitch technology to improve the performance of multi-GPU accelerated computing. It can even access artificial intelligence models built by the open source community through NVIDIA AI Workbench. For example, the Llama 2 open source model hosted by Hugging Face can be used to build automatically generated artificial intelligence application content in the VMware virtualization environment.
In addition, to make it easier for enterprises to deploy various AI technology applications, VMware has also simplified the threshold for enterprises to create and build AI application services by pre-installing vSphere Deep Learning VM image resources with many AI frameworks and performance optimization database content.
VMware expects to launch the VMware Private AI Foundation with NVIDIA service in early 2024, and Dell, HPE, and Lenovo will be the first to provide server systems compatible with this service and equipped with NVIDIA L40S GPUs, BlueField-3 DPUs, and NVIDIA ConnectX-7 smart network cards.
