
VMware Private AI Foundation with NVIDIA is a joint initiative aimed at lowering the starting costs of AI for enterprises. The platform is designed to deliver a predictable cost model along with greater privacy and physical control of the infrastructure for users.
Launched into general availability this summer, the solution is already gaining significant traction across VMware’s customer base in all major verticals, many of whom are leveraging it to drive use cases like cost and capacity planning, and tool sprawl management, said Jake Augustine, GTM lead for VMware Private AI at the AI Field Day event in September.
“Our focus has been around enabling AI workloads across the enterprise and simplifying the way those are deployed throughout the lines of business.”
The AI workload footprint starts light, but grows organically over time, leaving many organizations struggling to scale adoption. The biggest value proposition of Broadcom and NVIDIA’s joint solution is optionality to choose infrastructure components, AI services and even AI models as business needs evolve and emerge.
Built on top of VMware Cloud Foundation (VCF), a core product on the VMware portfolio, Private AI is an add-on SKU that is available on an additive core-based license. Private AI Foundation with NVIDIA utilizes NVIDIA AI Enterprise software suite above the platform, and NVIDIA GPUs down below, which currently include H100, A100 and L40s. The underlying hardware comes from VMware’s OEM partners, namely Dell, HPE, Hitachi, Lenovo and Supermicro.
Emphasizing the need for private AI solutions as AI becomes more mainstream in organization, Augustine said, “There’s a huge dominance of large language models and now we’re seeing the entrance of smaller language models, or micro language models in some instances.”
Data scientists and lines of businesses want choice to be able to take something that’s very very new into the market, prove it out internally and deliver it as quickly as possible to the enterprise for everybody to consume.”
The optionality that Private AI with NVIDIA offers gives customers an extra shot of agility and drives faster time to value. “The ability for a data scientist, without having to strip all the way to the bare metal, to have the agility on top of the platform gives them that choice.”
Augustine shared customer stories from multiple niches where this acceleration has helped companies get GPU cluster drivers ready in just hours and build out and deliver AI technologies internally within days.
“Time to value and how quickly you can start utilizing and distributing the GPUs as an AI resource within the data center is probably one of the better talk tracks in the feedback we get from customers.”
According to Futurum Intelligence data, VMware by Broadcom is currently among the top ten AIOPs vendors that new AI companies have on their radar. The combination of VMware and NVIDIA has clicked with a lot of customers primarily because of the broad familiarity of the VMware technology across the enterprise world, said Augustine.
The Computer Weekly reported in September that the platform has gained strong traction in the Asia-Pacific region, after the US which has the highest customer deployment.
Augustine highlighted one of the solution’s biggest selling points which is improved visibility of the infrastructure. Leveraging that, administrators can not only oversee utilization of resources, but also plan and allocate them optimally to reduce waste.
“This is no longer just a siloed project where GPUs are tied to single applications or single teams,” he said.
In a blogpost, Chris Wolf, VMware’s global head of AI wrote that Private AI customers that are in the mature stages of AI deployment have been able to up rake up 3 to 5 times more cost savings compared to public cloud AI services.
This is a key reason, Augustine said, that VMware Private AI Foundation with NVIDIA has resonated so widely with customers. Both token-based billing of public cloud and the fixed-cost model of on-premises tend to escalate cost as applications start to scale, making it harder for organizations to continue on their AI journey. Being in control of the infrastructure allows for long-term budgeting without sacrificing innovation.
“This is a very clear moment in time in the AI ecosystem. Our customers are very intent on proving out their use cases and the investment they’re making to justify the cost to move forward on some of these AI initiatives. How quickly they can accelerate and bring that value back to the business is top of mind for most leaders.”
Check out VMware’s demo and deep-dives of VMware Private AI Foundation with NVIDIA from AI Field Day at TechFieldDay.com.