Disclaimer: The Six Five webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.

Transcript:

Patrick Moorhead: The Six Five is on the road at Cloudera Evolve 2023 in New York City. Dan, it’s great to be back and hosting Evolve.

Daniel Newman: Yeah, it’s good to be here. We’re looking over the Hudson, beautiful skyline view, but we’re here to talk about AI?

Patrick Moorhead: That’s right.

Daniel Newman: And data. And of course, Cloudera.

Patrick Moorhead: Exactly. And it’s amazing how much AI and Generative AI has really taken over what we do as analysts, because quite frankly, enterprises need to know what to do, how to do it, where should they start? But the important thing, too, is, is like we’ve seen in all of tech, it takes a village, right? It takes partnerships, it takes tech companies banding together to be able to serve those enterprises. And that’s, interestingly enough, we’re here to talk about right now. And I’d like to introduce Priyank, how are you? Good to see you.

Priyank Patel: Thank you.

Patrick Moorhead: Welcome to The Six Five, first time on there. But we’ve had a lot of Cloudera folks on here and we’re looking forward to understanding what is going on here in the show. So thank you for coming on.

Priyank Patel: Thank you for having me on. And big fan of you guys.

Patrick Moorhead: Thank you. Appreciate that.

Daniel Newman: It’s always great when, even if it’s the first time that they’ve watched, they’re going to know what’s going on. He’s going to be right in the flow. Maybe even need to take one of our seats one day, Priyank. Start thinking about that. But listen, it’s a big moment. Perhaps this last year has brought an inflection to something that actually is very much in the heritage of Cloudera. Cloudera has been helping companies get their data in order for a long time. And before it was popular and before ChatGPT was a thing, Cloudera has been here working closely with enterprises, many of the biggest in every single industry. And in fact, I saw that slide this morning when Charles gave his presentation. It was very good. But I want to talk about Ecosystem. Today, the company made a lot of commentary and focus and its first wave of its new ecosystem partners. Talk to me about what’s this ecosystem launch? Who are the partners? How was this all decided? Give us the skinny.

Priyank Patel: Look, I’ll start with what Pat said earlier, right? It takes a village, but it actually takes an ecosystem for customers to realize value from their data. And it’s not new to us. You nailed it. We have always fostered big ecosystems around our platforms. This particular day, this is exciting for us because it is the announcement of the first wave of our AI ecosystem partners. And I think the genesis of this is it’s quite simple. Anywhere we have gone and spoken to our customers and even folks who have not used us for a while, but they’re thinking about using us, they’ve always said the data platform is the root of the AI platform. And there’s something in it that really sticks with me with that, which is whenever you start building any kind of AI strategy to make sense of ChatGPT, thinking past ChatGPT is about thinking how to leverage that technology in the enterprise or in the company and doing so you have to start with the core, which is your data core, right?

And that’s really the opportunity that we have to add value to our customers. But we realize that doing everything with a single technology is not the right strategy for almost every customer out there. And that’s where the ecosystem comes into play and the evolving and emerging AI stack. And so some of the announcements, the inaugural partners that we announced today are essentially targeted to start out with that deep integration from the product side, which is why I’m talking to you about it as well as making sure that there are multiple options that the customers have as they move fast on this journey. And fast is the keyword over there. I’ll come back to it.

Patrick Moorhead: So Priyank, both of our analyst firms believe and we have believed whether it was analytics based AI, machine learning, deep learning, now generative AI, you have to have your data in order and it’s good to see what you’re doing now in this world of generative AI as well. And in fact, the challenge is exasperated because the future of generative AI is not just doing domain specific magic tricks here. It’s across domains in areas that typically enterprises haven’t merged data, right? So your first partner out the gate, unsurprisingly, and somebody you’ve done a lot of work with, is NVIDIA. So, the question is what are you doing with them and how is that different from what you were doing with them before? Because I think I’ve seen them on your slides in previous years.

Priyank Patel: Yes. So NVIDIA has been, as you said, a longstanding partner and we have had integrations with GPUs or accelerated compute of NVIDIA for acceleration of our Spark based workloads on premise for a while, right? With the generative AI wave and the adoption that we are seeing across customers, what we are doing new now is to enable NVIDIA powered accelerated compute across data engineering, machine learning, AI training, AI inference, the entire lifecycle of the AI application is where we are powering with NVIDIA now, right? And that takes, sitting behind that headline, is the integrations that we have done with the GPUs and the newer generation of GPU servers that NVIDIA has brought to market as well. So that’s what we are doing.

Patrick Moorhead: So it’s absolutely a step up in, so going from Spark to basically what sounds like the entire AI workflow. Okay, how about this? I can’t wait to get more information on it and the video and all that stuff, but great stuff.

Daniel Newman: It’s interesting too, Priyank, because NVIDIA has become a bit in vogue for every company. And so, what I’m hearing here, though, with really full data, pipeline lifecycle hardware to software integration is that the focus here is not for Cloudera to just be another, I mean, there’s been a lot of arm waving, mostly of Jensen in front of large rooms of people from every tech company on the planet.

Patrick Moorhead: He’s keynoted, well, he showed up on, I think, seven stages at events that you and I have been to in the last nine months, right?

Daniel Newman: It was a very short period of time. And so that’s what I’m hearing. And obviously I think that’s what customers want to hear is look, it’s not arm waving, it’s not hand waving. It’s fully thought out, end to end, integrations that help companies deal with this complex data ecosystem and take advantage of all the compute power that’s out there. Now, you also mentioned, maybe a lesser known, and that’s Pinecone.

Priyank Patel: Yes.

Daniel Newman: Now if you’re a data scientist, you’re very familiar, but if you’re in the business side, you’re probably not hearing much. Talk a little bit about that release and what that’s about.

Priyank Patel: So to put into perspective, right, there are three thesis that we have around our AI ecosystem. One is on the data side, we understand that data powers the AI. Second is the models and the intelligence that’s sitting out in the large language models, whether it’s foundation models or open source. And the third one is to be able to enable it at the right price point in a hybrid form factor wherever the customer wants to deploy. And the third one is what we talked about first, which is with the NVIDIA where us integrating the entire lifecycle on NVIDIA makes it easy.

Pinecone fits into the first thesis, which is if data is the one that is going to determine the outcome and impact of the applications that we are seeing, for example, code completion, chat doc summarization or speech to text analysis and you name it, we’ve seen we have production applications running on Cloudera, then it is important to provide quick access to cloud-based vector stores because Vector Store is a building block of these applications. You can have the data stored in large scale systems like ours, but you do need to represent them as vectors before you can actually feed them into the models, or the LLMs, to make anything useful out of it, right? That is the genesis behind our Pinecone app integration, which what we did there is we took the Pinecone APIs and we published a Cloudera machine learning amp or an applied prototype that gives a blueprint for how to leverage Pinecone alongside Cloudera for any customer who wants to use it. And you can imagine a lot of our customers are obviously looking for cloud first options, which is really what Pinecone offers alongside us.

Patrick Moorhead: The amount of unstructured data, too, is where vector databases come in and add value. That could be images, those could be videos, that could be anything that’s the spoken word, like a conversation that you want to follow up and automagically create action items for a lot of popularity there. I’d heard of Pinecone previously, but not nearly as much as I am now. And plus my son came, he’d had an internship and he was using Pinecone for his generative AI work that he was doing.

Priyank Patel: It is the fast way to get started, right? Going back to fast, right? And you need to move quickly is what customers have been telling us because the space itself is evolving so fast that if you start six months late, you might be six years behind because your competition may have moved fast enough in getting a sense of how to adopt this technology for the impact that it’s generating.

Patrick Moorhead: Does vectorizing data in some way help secure the data in this new world here?

Priyank Patel: If architected right, right? When you actually store the vector representation of the unstructured data or text or the prompts, as we call it, into a vector database, you do need to make sure that you are architecting the authorization and the authentication around the vectors appropriately. So in our integration with Pinecone, for example, when you deploy the Pinecone, when you deploy our amp, which is feeding data into the Pinecone database and then leveraging that on the other side for analytics, the Cloudera side is integrated with SDX, which is our security and governance framework. And that’s really, it’s important when you build out the application to be within that framework so that you can eventually trust that the data coming out or feeding your application is the right one and thereby the impact or the results are right.

Patrick Moorhead: The caveat of, if architected correctly, is an interesting one.

Daniel Newman: It’s almost something an analyst would say.

Patrick Moorhead: Oh, exactly. Exactly. So, your third partner that you announced at the event was, somewhat unsurprisingly, AWS, right? Number one market share in IAS, out there right now, the 800 pound gorilla for IAS and Dr. Matt Wood actually published a blog outlining what the two of you’re doing together, which by the way, I think the analysts, right? When we see partners putting effort in describing what they’re doing, that’s a true partnership. And I know that AWS doesn’t throw around blogs just for everybody. So what are you doing with them and what’s new? What’s different? Because you have CDP that operates with AWS already, what is this new stage of your relationship?

Priyank Patel: This is about Cloudera and AWS AI. Dr. Matt Wood’s blog outlines our joint strategy on AI as it relates to Cloudera and the services in the AWS AI ecosystem. Of specific note from our side is we have a two-pronged strategy of how we enable AI or how we leverage AI. Number one is to build AI into Cloudera so that all our users, all our customers, can leverage the benefits of large language models, engine AI directly in the platform without ever needing to understand or know the technology. The second, which is also of interest for us because we serve enterprises, is to make Cloudera the best platform to build AI applications with. In both of those, we are partnering with Amazon AIs, particularly the bedrock serverless offering that just went GA a few weeks ago.

And we’ve been working with them for much longer prior to the service being generally available and integrated that into our platform to provide our SQL AI assistant, right? So in Cloudera we have a data warehouse service as part of our platform. With the SQL AI assistant integrated with Bedrock, now, we can essentially make that service accessible to English language prompts, so that you don’t need to know or write SQL to be able to leverage the data or ask questions of the data within the Cloudera ecosystem. So that’s the concrete integration point.

The other one is building applications, whether it’s architectures like RAG or Retrieval Augmented Generation, fine-tuning these patterns are the ones that we published CML amps on with Bedrock to make it easy for customers to leverage these as well. I hope that makes sense. This is about the AI services within the Amazon ecosystem. Of course, our partnership, as you noted, is longstanding and broader than that, this one is particularly exciting because of the opportunity that stands ahead of us.

Daniel Newman: So what I’m seeing is strong partnerships with respectively, the leaders in categories. You’ve got rapid development with Pinecone. You’ve got the absolute leader and hardware right now with NVIDIA. And then, of course, AWS has not only the largest IS deployment globally, but the most customers. Talk a little bit about how this progresses. I mean, for a company like Cloudera, is it about getting Mindshare with the rest of, I know Intel’s a partner and AMD is a partner. So is it building everything out? Is it about being with all the different cloud providers? Is it moving to more of a vocal partnership with Google and with Microsoft? And is that the approach? Is it to start with the leaders and move out or what are you thinking? What’s the next here for the Ecosystem?

Priyank Patel: We optimized for the customer’s viewpoint to get started fast for the customer. What is required for that? And it so happens that we started out going deep with NVIDIA, Amazon and Pinecone. That doesn’t mean that that ends there. That’s why it’s the first wave, right? Something that we are super excited about is the other part that we are hearing is you have the access to the broadest range of foundation models with Amazon with the integration. That’s great. There is an even larger community of models sitting out in open source communities like Hugging Face, right? And so we are hearing from customers who are already leveraging models on these communities and hubs correctly on the Cloudera platform or using the Cloudera data to fine tune those models.

And there’s something that we are super excited about, how to make that easier. There’s a set of customers who have already figured it out and they’re ahead on the journey, but there is a good chunk of customers who would need it to be well packaged, well integrated before they move. And we are definitely here to solve that. And that’s another example of where the ecosystem moves next, speaking specifically on the AI side.

Daniel Newman: Priyank, I have to say that you guys hit it. Hugging Face, by the way, Pat was another kind of legend of 2023. It’s been everywhere and anywhere out in front because it really was one of those that democratized the open source, large language model. And it sounds like you’ve made some very smart early decisions in Ecosystem and, of course, we’ll be paying attention and providing you some input on maybe what the next wave should be, because that’s what we do.

Patrick Moorhead: I’d love to hear, next year the progress and new partners you’ve added.

Daniel Newman: Absolutely. So, thanks so much for joining us after all this time being a viewer. We loved having you on the show.

Priyank Patel: Thank you, Dan. Thanks, Pat.

Daniel Newman: Thank you.

Priyank Patel: Appreciate it.

Daniel Newman: All right everybody, you heard it here. We’re at Cloudera Evolve 2023 in New York City. But for Patrick and myself, we’ve got to say goodbye. Subscribe and watch all the other episodes and all the other coverage here. We’ll see y’all soon.