At VMware Explore 2024, Broadcom and NVIDIA have unveiled a series of new capabilities for their joint platform, VMware Private AI Foundation with NVIDIA, which will be officially launched in May 2024. This generative artificial intelligence solution is designed to provide a private and secure infrastructure for enterprises, focusing on privacy, flexibility, performance, and security.
New Features with VMware Cloud Foundation 5.2.1
The release of VMware Cloud Foundation (VCF) 5.2.1, scheduled for later this year, will include several new capabilities that will enhance the user experience of VMware Private AI Foundation with NVIDIA:
– Model Store: This functionality will allow MLOps teams and data scientists to securely curate and provide Large Language Models (LLMs) with integrated access control. The Model Store will improve governance and security in the environment, ensuring the privacy of enterprise data and IP.
– Guided Deployment: To simplify the deployment process of Gen AI, this new capability optimizes the creation of workload domains and the deployment of additional components, accelerating implementation time and reducing administrative efforts.
Capabilities of NVIDIA AI Enterprise
– NVIDIA NIM Agent Blueprints: These reference workflows enable companies to build their own generative AI solutions. They include necessary tools for developing custom applications, such as workflows for customer service, drug discovery, and PDF data extraction.
– NVIDIA NIM: A set of microservices designed for secure and reliable deployment of high-performance AI models. The NIM microservices support a wide range of AI models and easily integrate into enterprise applications with simple commands.
– NVIDIA NIM Operator: Facilitates the deployment and management of generative AI pipelines by automating deployment, scaling, and inference management, reducing latency and improving automatic scaling performance.
Future Capabilities
Broadcom has also announced additional capabilities for future versions of VMware Private AI Foundation with NVIDIA:
– vGPU Profile Visibility: Enables administrators to see all vGPUs created through an interface in vCenter, eliminating the need for manual tracking and improving operational efficiency.
– GPU Reservations: This new feature will allow administrators to reserve resources for vGPUs in advance, improving capacity planning and performance.
– Data Indexing and Retrieval Service: Facilitating the preparation of private data for generative AI, this feature will allow indexing and vectorizing private data sources, improving the quality of Gen AI results.
– AI Agent Builder Service: Will help developers and data scientists build and deploy custom AI agents using LLMs and data from the indexing and retrieval service.
Expansion of the Ecosystem
Broadcom is also expanding the ecosystem of VMware Private AI Foundation with NVIDIA by adding new providers and partners, including:
– Codeium: Offers assistance in code generation and debugging using AI, improving development efficiency.
– HCLTech: Provides customized Gen AI solutions to accelerate Gen AI adoption with a tailored pricing model.
– Tabnine: Custom AI tools for software development, maintaining privacy and control.
– WWT: Technology solutions provider supporting companies in the implementation and operation of AI applications.
The collaboration between Broadcom and NVIDIA on the VMware Private AI Foundation platform represents a significant advancement in generative AI infrastructure for enterprises, offering new tools and capabilities to enhance efficiency and security in data and AI application management.