TL;DR:
- VMware and Nvidia announce VMware Private AI Foundation with Nvidia, a fully-integrated solution for generative AI training and deployment.
- Enterprises gain a single-stack product encompassing software, computing capacity, and tools for fine-tuning large language models.
- Solution allows private, high-performance generative AI applications on proprietary data within VMware’s hybrid cloud.
- Offering addresses data privacy, security, and control concerns while empowering enterprises to run AI workloads adjacent to their data.
- Collaboration aims to streamline gen AI app development, testing, and deployment within VMware’s cloud infrastructure.
- Nvidia NeMo framework powers the solution, enabling pre-tuning, prompt-tuning, and optimization of runtime and results.
- Nvidia AI Enterprise Systems with L40S GPUs, BlueField-3 DPUs, and ConnectX-7 SmartNICs support the foundation.
- Scalability of the solution supports workloads of up to 16 vGPUs/GPUs in a single virtual machine across nodes.
- Additional features include deep learning VMs and a vector database integrated with Postgres for streamlined app development.
- Launching in early 2024, with AI-ready systems to debut by year’s end from 20+ global OEMs, including Dell, HPE, and Lenovo.
Main AI News:
In a strategic partnership spanning over a decade, VMware and Nvidia have announced a groundbreaking advancement that is set to revolutionize the landscape of generative AI training and deployment. This visionary offering, christened the VMware Private AI Foundation with Nvidia, introduces a fully integrated solution that promises to empower enterprises with comprehensive tools to harness the potential of large language models and facilitate private and high-performance generative AI applications within VMware’s hybrid cloud infrastructure.
Addressing the ever-present challenge of managing diverse sets of data scattered across various environments, Raghu Raghuram, CEO of VMware, affirms, “Customer data is everywhere — in their data centers, at the edge, and in their clouds. Together with Nvidia, we’ll empower enterprises to run their generative AI workloads adjacent to their data with confidence while addressing their corporate data privacy, security and control concerns.”
Projected for a debut in early 2024, this groundbreaking solution is poised to offer a comprehensive suite of capabilities to usher in a new era of AI-driven innovation. As enterprises globally vie to capitalize on the immense potential of large language models, such as for developing intelligent chatbots and summarization tools, the demand for efficient solutions to handle these models efficiently has surged. McKinsey forecasts that generative AI, or “gen AI,” could potentially contribute up to $4.4 trillion annually to the global economy. However, the pursuit of these goals often leaves teams grappling with fragmented environments that compromise data security and AI performance.
The VMware Private AI Foundation with Nvidia is poised to quell these concerns, offering a unified solution that transforms VMware’s cloud infrastructure into a centralized hub. Here, enterprises can effortlessly select from various open models like Llama 2, MPT, or Falcon and streamline their development, testing, and deployment processes for gen AI applications. Paul Turner, VP of product management at VMware, elucidates, “It takes those models and provides all the power of Nvidia NeMo framework, which lets you take those models and helps you pre-tune and prompt-tune as well as optimize the runtime and results from gen AI workloads. It’s all built on VMware Cloud Foundation on our virtualized platform.”
The NeMo framework, renowned for its end-to-end cloud-native capabilities, merges customization frameworks, data curation tools, and pre-trained models to facilitate the seamless deployment of generative AI. Complementing this, VMware Cloud Foundation equips enterprises to harness their data effectively by offering a suite of software-defined services to run and manage developed applications.
Central to this offering is the unwavering commitment to data privacy. Nvidia’s robust infrastructure provides the computational prowess required, rivaling or even surpassing the capabilities of bare metal configurations in select scenarios. This is bolstered by collaborations with ecosystem OEMs, slated to unveil Nvidia AI Enterprise Systems equipped with Nvidia L40S GPUs, BlueField-3 DPUs, and ConnectX-7 SmartNICs, all tailored to support the VMware Private AI Foundation with Nvidia.
Paul Turner emphasizes the scalability of the solution, highlighting its ability to manage workloads scaling up to 16 vGPUs/GPUs within a single virtual machine across multiple nodes. This scalability ensures expedited fine-tuning and deployment of generative AI models.
The innovation doesn’t stop there. VMware is additionally incorporating distinctive features, including deep learning VMs that accelerate enterprises’ journey toward building generative AI apps. This includes a vector database integrated with Postgres and PG vector, a valuable resource for managing rapidly evolving information integral to model development.
While the work on VMware Private AI Foundation with Nvidia advances, the first AI-ready systems are slated for release by year’s end, with the full-stack suite expected to debut in early 2024. In a promising trajectory, Nvidia anticipates the availability of over 100 servers supporting VMware Private AI Foundation from a diverse lineup of global OEMs, including industry giants like Dell Technologies, Hewlett Packard Enterprise, and Lenovo. This symbiotic collaboration is poised to reshape the future of AI deployment, driving transformative value across industries.
Conclusion:
The collaboration between VMware and Nvidia presents a game-changing solution that harmonizes AI deployment complexities. By offering a holistic platform for gen AI app development, coupled with data privacy assurance and unmatched computing power, enterprises are poised to harness AI’s transformative potential. This strategic move underscores the evolving landscape of AI applications in the business realm, setting a precedent for efficient, secure, and high-performance AI adoption.
