At the VAST Forward World Tour in London, VAST Data founder and CEO Renen Hallak explains why he believes his company has built “the AI operating system” powering the next wave of training, inference and agentic applications.
Hallak describes VAST Data’s shift from serving “the organisations that were on the bleeding edge of AI” to supporting Fortune 500 firms now moving “really really fast” from sandbox tests into production. He discusses how new AI clouds such as CoreWeave, Lambda and Crusoe are rebuilding infrastructure “from the ground up” and why VAST aims to be “that fabric across the d...
Questions asked in this interview
10
0:05Can you just for our audience who might not be aware of Vast Data, very, very quickly explain to them what you guys do and where you fit into the kind of AI ecosystem?
1:28So can you talk us through that journey that you guys have been on?
2:14Would a CoreWeave or a Lambda or a Crusoe, would they be like a gateway for an enterprise to use the Vast platform if they're using their AI cloud services, or are they kind of direct customers themselves? How does that work?
3:13How is that going to change and grow moving forward?
4:03What do you think about kind of enterprise adoption as a whole moving forward?
5:05Do we need to start thinking about from a latency perspective moving closer back to population centers or creating these kind of more like inference edge nodes? How do you view that?
6:22So is there anything that we could be doing or that Vast is doing to be able to try and ensure that that uptime is approaching the kind of five nine level that you would expect from a traditional cloud service?
Hi, we're here at the Vast Forward World Tour, the London stop of the World Tour. I'm here with the man of the hour, Renen Hallak himself, CEO of Vast Data. So, Renen, thank you very much for the invite to the event. We're having a great time so far. Can you just for our audience who might not be aware of Vast Data, very, very quickly explain to them what you guys do and where you fit into the kind of AI ecosystem?
Yeah, at Vast we built the AI operating system. We're basically the software infrastructure layer for these new training and inference and fine-tuning applications. We started 10 years ago building the storage layer and then we added the database layer and the compute layer trying to build out all of the tools required for these new agentic applications.
So again when we started it was the organizations that were on the bleeding edge of AI, a lot of hedge funds and life science institutes. And then as AI became a thing it was the large model builders like xAI and a bunch of others. And obviously the new AI clouds like CoreWeave and Lambda and Crusoe are big customers of ours over the last 12 months. As AI makes its way into the enterprise, Fortune 500 companies are starting to leverage our platform in order to fuel these new agentic applications.
How have you had to change your approach as you've moved through those different layers of, I guess, comfortableness with AI if you want to put it that way? I really liked your analogy about, you know, in the 70s you had to have a computer science degree to operate a computer. Microsoft comes along, now everyone does it without thinking about it. So can you talk us through that journey that you guys have been on?
Yeah. So in the early days we were working with experts and so they were telling us what AI is and what it requires from a scale and performance perspective. And then as this evolves we want to make it widespread. And so we're focused a lot more on simplifying it, on making it secure, making it easy for enterprises and non-experts to both develop agents and use these new abilities in production.
Yeah. And where do you see the NeoClouds fitting into that overall story and that journey that you've been on? Would a CoreWeave or a Lambda or a Crusoe, would they be like a gateway for an enterprise to use the Vast platform if they're using their AI cloud services, or are they kind of direct customers themselves? How does that work?
So, a little bit of both. And so, a lot of them are using us as that software infrastructure layer in order to attract enterprises and large model builders and AI startups. The old clouds were built on old hardware, on an old stack, on an old architecture. And these new AI clouds are really rebuilding everything from the ground up. So they have a new type of power plant, new type of data center, new GPUs and fast networking and SSDs within those data centers and they need our software infrastructure layer on top of that to provide all of these AI services.
So moving forward, where do you see the bulk of Vast's growth coming from? Do you think it's like an even mix between these deals such as the one you signed with CoreWeave this week versus the kind of directly with the enterprises? How is that going to change and grow moving forward?
So, as more and more enterprises start to adopt AI and governments and basically everybody, there are different ways for them to access the hardware and software that is required for it. Some of them will build out their own infrastructure, others will go through AI clouds and our aim is to serve them wherever they may be. And in many cases it's a hybrid approach where some of their infrastructure belongs to them, other parts is in the cloud and we want to be that fabric across the data space that gives them one abstraction layer regardless of where they want to operate.
So I guess it's quite important for your business model and the business model of a lot of your partners and your customers that we do get to this point where enterprises are not just testing AI but they're using it and it's delivering value to them so they can reinvest in it and grow their use of it. How concerned are you about how that process is going? Do you think there's anything that needs to happen for us to get more enterprise adoption? Maybe there's a comparison between the US and where we are right now in the UK that you can make of the different kind of regulatory environments or what they can and can't do. What do you think about kind of enterprise adoption as a whole moving forward?
12 more exchanges in this transcript
Sign in free to read the rest of this interview. No card required.
Hallak, R. (2025, December 5). VAST Data CEO predicts enterprise AI will happen “a lot faster than we thought” | The Tech Capital [Interview transcript]. The Tech Capital. CEOInterviews.AI. https://ceointerviews.ai/interview/716123/
MLA
Renen Hallak. "VAST Data CEO predicts enterprise AI will happen “a lot faster than we thought” | The Tech Capital." The Tech Capital, 5 Dec. 2025. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/716123/.
BibTeX
@misc{hallak2025_716123,
author = {Renen Hallak},
title = {VAST Data CEO predicts enterprise AI will happen “a lot faster than we thought” | The Tech Capital},
howpublished = {Interview transcript, The Tech Capital. CEOInterviews.AI},
year = {2025},
month = {dec},
url = {https://ceointerviews.ai/interview/716123/},
note = {Speaker-attributed transcript with timestamps}
}
