VMware Explore5 min read

Private AI Gains Momentum at VMware Explore Barcelona

Chris Wolf

Today at VMware Explore Barcelona, we unveiled several new and expanded partnerships that further demonstrate VMware’s Private AI momentum. Private AI is an architectural approach that aims to balance the business gains from AI with the practical privacy and compliance needs of the organization.

Why Private AI

Since announcing our Private AI news at VMware Explore Las Vegas, there has been significant enthusiasm around our AI strategy and its value from both customers and partners. Customers are excited by the flexibility we provide through our growing open ecosystem. They highly value our ability to offer centralized management and operations for both AI and non-AI workloads, as well as our ability to enable them to virtualize and share their in-demand GPU resources – both of which can lead to a reduced total cost of ownership (TCO).

While most organizations plan to use combinations of both public and private AI service stacks, there are several use cases for running AI compute and models adjacent to where enterprise data is created, processed, or consumed. Common emerging use cases include:

  • Code generation – Improving developer productivity while preserving privacy and control of an organization’s source code through solutions such as our SafeCoder offering with Hugging Face.
  • Contact centers resolution – Improving time-to-resolution for customer support while safeguarding private internal support documentation.
  • IT operations automation – Providing richer automation and insights into internal IT processes while safeguarding management and operational metadata.
  • Advanced information retrieval – Streamlining document search, summation, and content creation while preserving an organization’s unique intellectual property as well as its unique voice in how it communicates with customers and partners.
Partner Momentum

On the partner front, VMware’s approach to Private AI is creating many opportunities for our partners to differentiate and add value. Our focus on Private AI infrastructure has made us extremely attractive to AI independent software vendors (ISVs) because we are offering them a differentiated platform to reach customers while not directly competing against their core business. Systems integrators are lining up to help accelerate AI use case adoptions on VMware Cloud Foundation to meet customer demand and address top use cases. At VMware Explore Barcelona, we announced that two additional SI partners – IBM Consulting and Kyndryl – are working with VMware to accelerate Private AI infrastructure and services growth to help organizations quickly realize value through Private AI.

VMware’s AI strategy is centered on differentiated technologies, simplicity, low-cost, and choice that allows organizations to invest in a common AI platform that can quickly onboard new services – wherever they do business – as business or industry needs change. Our growing partner ecosystem is critical to our success, and we are excited about the new and expanded partnerships announced in Barcelona.

Intel Partnership

Our decades long partnership with Intel is now expanding to Private AI. Today, we shared that VMware and Intel are designing a VMware Private AI reference architecture that will enable customers to build and deploy Private AI models and reduce TCO by harnessing the Intel AI software kit, processors, and hardware accelerators with VMware Cloud Foundation.

This VMware Private AI reference architecture with Intel AI will include the following on VMware Cloud Foundation:

  • Intel Max Series GPUs - delivering up to 128 Xe-HPC cores, the new foundational architecture targeted at the most demanding AI workloads.
  • 4th Gen Xeon CPUs with Intel® Advanced Matrix Extensions (Intel® AMX) - delivering significant out-of-box performance improvements spanning multiple frameworks, end-to-end data science tools, and a broad ecosystem of smart solutions.
  • Intel AI Software Kit – packaged end-to-end open-source software and optional licensing components to accelerate building multi-node scaling and deploying AI.

At Explore Barcelona, we will be demonstrating multiple generative AI services running on CPUs while showing impressive accuracy and performance. These demos are likely to get many to rethink how they can put their CPU capacity to work for their AI needs, especially in inferencing use cases. For more information, see the VMware Private AI with Intel solution brief.

IBM watsonx Partnership

VMware is partnering with IBM to enable enterprises to access IBM watsonx in private, on-premises environments, as well as hybrid cloud for the secure training and fine-tuning of their models with the watsonx platform. This full-stack architecture built on VMware Cloud Foundation and Red Hat OpenShift will enable organizations to deploy watsonx AI functionality for MLOps, data management, and governance. Additionally, organizations will be able to access IBM-selected open-source models from Hugging Face, as well as other third-party models and a family of IBM-trained foundation models to support GenAI use cases. For more information, read the VMware Private AI and IBM watsonx announcement here.

VMware to Help Lead Unified Acceleration (UXL) Foundation

VMware has joined ARM, Fujitsu, Google Cloud, imagination, Intel, Qualcomm and Samsung to help lead the Unified Acceleration (UXL) Foundation as a steering member. UXL Foundation is a cross-industry group committed to delivering an open standard accelerator programming model that simplifies the development of performant, cross-platform applications. UXL Foundation is building on the oneAPI specification and harnessing Industry-based open standards like SYCL. The UXL foundation will define a vendor-neutral way to develop software for heterogeneous architectures by embracing neutrality and independence. UXL’s ambitions align to VMware’s decades-long track record of streamlining interoperability in both hardware and software.

Data Services Partnerships

VMware is expanding its data services offerings beyond our existing multi-cloud Tanzu Data Solutions to include Data Services Manager (DSM) in VMware Cloud Foundation. Expanding our on-premises data services ecosystem is essential to enable organizations to have more flexibility in how they build AI services to meet business needs. Up to this point, DSM has supported VMware Tanzu SQL Postgres and MySQL and is now expanding to offer a unified control plane for managing VMware and partner data services.

Today, we announced the addition of new partner services that integrate with DSM to provide new levels of flexibility to your on-premises environments. First, we are partnering with MinIO to bring their highly performant and resilient object store to VMware Cloud Foundation. Additionally, we have expanded our partnership with Google Cloud to bring AlloyDB Omni’s superior performance, scalability, and cost-efficiency to VMware Cloud Foundation. When it runs on vSAN, Google has seen a greater than 2x performance boost for transactional workloads and up to a 100x boost for analytical workloads compared to standard Postgres.1

High-performance databases and object storage are key data services to power AI/ML workloads, and these new partnerships will help you to further accelerate your Private AI journeys.

Our Collective AI Journey

In addition to the partnerships announced in Barcelona, we continue to work closely with NVIDIA – following the unveiling of our joint offering at VMware Explore Las Vegas – to bring VMware Private AI Foundation with NVIDIA to market in the coming months. A selection of customers are already working with VMware and NVIDIA as design partners running this new offering in their environments and helping to shape the solution’s future.

A common question I hear from customers and partners regarding Private AI is, “This is great! How can I get started?” At VMware Explore Barcelona, we have released Starter Packs with Private AI proofs-of-concept that are aligned to popular use cases, with both Coding Assistant and Retrieval Augmented Generation (RAG) use cases available today.

Are you ready to rethink your approach to AI? If your organization desires data privacy and control, flexibility, choice, and lower costs, VMware would like to talk to you. Join us on our collective AI journey: https://www.vmware.com/artificial-intelligence.html

Sources

  1. Internal Google Cloud Testing, March 2023