- First CoreWeave and now Google Cloud have inked big partnership deals with startup VAST Data
- VAST offers something called the AI Operating System, which helps organize and orchestrate data so it can easily feed AI models
- Others are taking a similar approach, but VAST's big name partnerships could give it a leg up as it eyes enterprise deployments
VAST Data is on a roll. Just a week after inking a $1.17 billion commercial deal with neocloud leader CoreWeave, the company expanded an existing partnership with Google Cloud to make it easier to run enterprise AI in hybrid cloud environments. And a few months back, news broke that the company was working on a fresh funding round valuing the company at a whopping $30 billion.
But who exactly is VAST Data and why does everyone suddenly want a piece of them?
To answer the latter question first: VAST Data is, in a nutshell, building the data infrastructure that will feed the AI factories of the future. As Moor Insights and Strategy VP and Principal Analyst Matt Kimball told Fierce, that’s a task that’s “Easy to write down. Hard to implement.”
Kimball explained that VAST’s AI Operating System “connects to all of your data sources, readies the data and feeds your pipelines. Think of it like Linux - it unifies all of your data and database semantics under a layer for consumption by AI.”
Other companies like Weka, DDN and NetApp are working on similar concepts. But Kimball noted VAST has partnerships with key players that make it stand out. Chief among these is its collaboration with Nvidia (which is also one of its investors) to integrate its technology in the chipmaker’s GPU ingest path. As Kimball noted, that has, of course, drawn the interest of cloud giants like Google and CoreWeave. But there’s more to it than that.
“I think most look at VAST as only suitable for the largest of the large AI players,” Kimball said. “However, it is quietly developing an architecture that I believe will translate into a very consumable enterprise play as we see the inference wave hit the market in full force.”
About that architecture
But VAST Data didn’t necessarily start as an AI-focused company. Founded in 2016, the company originally targeted cloud storage, and from there it moved into a second phase focused on integrating database capabilities into its platform. Now, of course, it’s focused on the AI OS in its third phase.
It’s actually this phased approach that set VAST up perfectly to address the AI data pipeline issue in a unique way. Why? Because it had already built and integrated all the pieces of the puzzle.
“Most systems are building on lots of disparate pieces of software to address these workflows,” VAST Data Field CTO Kyle Lamb recently told Fierce at Nvidia’s GTC event. “You have a separate database, you have a separate storage system, you have a separate computing system and those three interoperate but there’s no central planning of who’s managing what.”
He continued: “By taking a data-focused approach on this system, we’re able to know the data itself, know the security of the data and then extrapolate based off that for those other pieces. And it’s all orchestrated within a single platform that allows you to have actual control over who’s doing what and when and where.”
When making its case to enterprise customers, it certainly won’t hurt that VAST has an established presence in major AI clouds.
AWS, Google Cloud, Microsoft Azure, CoreWeave, Crusoe, Lambda and Core42 are all listed among VAST’s cloud partners, though it seems to have the deepest relationships with CoreWeave and Google.
To date, VAST has raised around $380 million from a range of investors, including Nvidia. Its last funding round was in 2023, but as noted above the company appears to be angling for another go.