Analyzing The Lambda Labs Partnership With VAST Data

By Matt Kimball, Patrick Moorhead - November 27, 2023

VAST Data scored another cloud partnership last week when GPU cloud provider Lambda Labs selected the company’s technology as the storage backbone of its offerings. This partnership is the second significant cloud win for VAST, as the company struck a similar deal with CoreWeave a couple of months back.

In previous coverage, I dug deeper into VAST’s approach to data management and how its development of a data management platform based on its DataBase and DataSpace offerings were potential game changers. In this analysis, I’ll show how these technologies bring hybrid AI to life through such partnerships.

A Little Bit About Lambda

Founded in 2012, Lambda delivers a high-performance computing platform for some of the most demanding customers in the world, including MIT, Microsoft, Raytheon, Boston Scientific, Airbus and many more.

While customers can acquire Lambda’s Echelon hardware for use in their datacenters, the company’s focus appears to be on its GPU cloud offering known as Lambda On-Demand Cloud. Think of the Lambda cloud as a virtual sandbox where any enterprise can play. Here are some possibilities:

  • If an organization has a deep learning environment and needs access to a cluster of Nvidia H100 GPUs for bursting while performing deep training, the Lambda cloud is a good fit.
  • If an organization has its own AI cluster but would like Lambda to house it and manage it, colocation services are available.
  • Finally, if an enterprise wants to build its entire deep training environment in the cloud, Lambda is happy to support that model as well.

It’s important to note that Lambda’s roots are in designing hardware and infrastructure to support the most highly performant HPC and AI workloads on the market. The servers one deploys for hosting typical virtual workloads have a different set of requirements than those trying to solve the most challenging issues being tackled in the computing world. Lambda seems to have figured this out, judging by the who’s who of players in technical computing using the company’s technology.

Forbes Daily: Get our best stories, exclusive reporting and essential analysis of the day’s news in your inbox every weekday.

Why is it important to note Lambda’s experience in hardware design? Because it gives the company a distinct edge in its GPU cloud offering. The Lambda cloud is not simply a bunch of commodity servers with some acceleration added trying to solve the AI challenge with a “more is better” approach. Instead, it connects very high-performing GPU servers at scale and delivers this computing power—along with Lambda’s hardware expertise—as a service.

So, while the company may not have been in the cloud business for years, it certainly has delivered what one would consider a real AI cloud.

There’s one other thing I would like to point out about Lambda. Because it has been so focused on absolute best performance since its inception, the company has set a very high bar for partnerships. It will not partner simply to check a box. There is no compromise in terms of quality or performance for complementary solutions.

The Partnership: VAST DataSpace In Action

Lambda has selected the VAST Data Platform as the storage and data management foundation powering its Lambda On-Demand GPU Cloud. What does this mean? Data is the key to GPU computing, whether you’re handling an HPC workload or the ever-so-fashionable generative AI. In the case of training, the faster that data can be fed into a model, the faster it can be trained.

This is where VAST comes into play. As a data management solution designed from the ground up to deliver best-in-class scale and performance, it powers some of the most demanding HPC and AI clusters on the planet. Combine this scale and performance with simplicity of operation, and you have a winning proposition that enables enterprises to easily access and manage data distributed across the environment.

In many ways, VAST is the natural partner for companies looking for the best AI performance. First, its certification as a datastore for the Nvidia DGX SuperPOD (the first enterprise NAS to achieve this, in fact) means that customers are assured of the performance and scale required for large AI projects.

Second, the company’s recent launch of its Data Platform really separates it in the marketplace. In this data-saturated era, storage is kind of a given. All of the same components are going into the same boxes and constrained by the same laws of physics. Differentiation comes from the way vendors best utilize these commodities. Further differentiation comes from how easily vendors enable their customers to exploit their capabilities fully. And this is what VAST has done so well.

Finally, we are seeing VAST’s recent innovations come to life in its cloud partnerships. DataBase allows users to perform analytics on both structured and unstructured data—it acts as a single interface to access all your data across the enterprise. And DataSpace enables enterprise customers to access their data across all three Lambda environments: the cloud, on-premises or colocated.

VAST DataSpace removes the notion of data locality and gravity.

I will be the first (and maybe only) analyst to admit that when VAST talked about its DataBase and DataSpace technologies, it wasn’t really clicking for me. I understood what was being said, but it all seemed fuzzy in terms of its real-world application. Watching how companies like Lambda are utilizing these capabilities, it all becomes crystal clear. And the VAST vision seems spot-on.

What This Means For Enterprise IT

Every enterprise is at least experimenting with AI, and each enterprise is unique in terms of AI capabilities and maturity. Regardless, a big part of the AI equation is infrastructure. The bespoke hardware and operating environment needed to perform training at scale is very costly—and difficult to deploy, optimize and manage.

Further, managing data is incredibly challenging. It requires far more than having a storage administrator order a NAS and deploy it. In this context, formulating an enterprise-wide strategy that spans the datacenter, edge and cloud is critical. And having a data management platform that supports this is fundamental to success.

The Lambda and VAST partnership helps organizations achieve on both the infrastructure and data fronts. While I’m a little tired of the “D” word (democratize) as it’s applied to AI, this partnership can help enterprise organizations deploy AI at scale with greater ease, better performance and lower cost.

Closing Thoughts

I like the Lambda-VAST partnership, just as I was a fan of the CoreWeave-VAST partnership. Real value is being delivered to customers, and both cloud providers can deliver differentiation in a market where that seems more and more difficult to do.

I have to admit I’m impressed with the folks at VAST. I’d like to say they’ve made storage cool, but they haven’t. Instead, they’ve made data management cool. And that is really what companies are trying to achieve.

Check back in the next few quarters as I’ll report on how the Lambda-VAST partnership is evolving.

+ posts

Matt Kimball is a Moor Insights & Strategy senior datacenter analyst covering servers and storage. Matt’s 25 plus years of real-world experience in high tech spans from hardware to software as a product manager, product marketer, engineer and enterprise IT practitioner.  This experience has led to a firm conviction that the success of an offering lies, of course, in a profitable, unique and targeted offering, but most importantly in the ability to position and communicate it effectively to the target audience.

+ posts

Patrick founded the firm based on his real-world world technology experiences with the understanding of what he wasn’t getting from analysts and consultants. Ten years later, Patrick is ranked #1 among technology industry analysts in terms of “power” (ARInsights)  in “press citations” (Apollo Research). Moorhead is a contributor at Forbes and frequently appears on CNBC. He is a broad-based analyst covering a wide variety of topics including the cloud, enterprise SaaS, collaboration, client computing, and semiconductors. He has 30 years of experience including 15 years of executive experience at high tech companies (NCR, AT&T, Compaq, now HP, and AMD) leading strategy, product management, product marketing, and corporate marketing, including three industry board appointments.