Teradata Corporation (TDC) Earnings Call Transcript & Summary

Thanks, everyone. I'm Derrick Wood, senior analyst covering enterprise software. We've Steve McMillan, CEO of Teradata. Thanks, Steve, for coming.

Derrick, it's great to be here.

And a reminder, voting starts next week. We appreciate your support. With that, let's dive in. I mean we're coming off of Salesforce's earnings last night, and there's been a number of companies that have talked about softer conditions in Q1. So I'd love to kind of get your take on what you've seen in the macro, what kind of buying behavior you've seen really over the last year and how that's evolved?

Yes. I think certainly for Q1 and Q2, I think we've seen some spending restrictions in the environment. I think it's going to free up as we move into Q3, Q4 as we move into the second half, just at a super macro level. The great thing is from a Teradata perspective, that CIO spend, if you look at the categorization of CIO spend, AI, data and analytics has overtaken cybersecurity now in terms of priority for spend and spend analysis. I think what's going to happen as we move into the second half of the year, you're going to see organizations moving from kind of proof of concepts around some of the AI solutions that they've had and to real production-oriented solutions. So that's going to be an exciting catalyst and tailwind, I think, for the entire data and analytics marketplace. And then if you look at it from a geography perspective, we're seeing some real success in what we call our international. So that's EMEA and APJ. The teams there are doing a great job. I think organizations there are kind of a little bit behind the U.S. in terms of their data analytics and AI strategies. And so we see that spend environment continuing. But as we move into the second half of the year, I think it's going to free up in terms of the opportunity that's in front of us.

Okay. You brought us right into generative AI, maybe I'll go there before I touch on some other subjects. So -- what -- I think you did have a new product that -- or a new solution that came out -- was announced in the last couple of weeks. Could you give us a sense of what the strategy is there?

Yes. So we've always had a great capability in the Teradata platform for AI and ML functions. And in fact, in August 2021, we announced ClearScape Analytics, which was essentially taken AI and ML functions and embedding them right into our database engine so that you can run complex data and analytics at really super high enterprise scale performance levels. Interestingly, just 2 months before all of the interest around ChatGPT come out, if you think that's just November 2022 when that happened. So it's interest in terms of looking at how that marketplace has evolved. Just from a generative AI perspective, we've a very open and connected viewpoint from open AI. So we've kind of bring your own model. So bring your own language model. You can run it inside the Teradata environment, and you can actually trust the data that you're feeding into that language model because it's all the highly pristine data that you run your company on a day-to-day basis. So it reduces the hallucinations from a language model perspective, gives organizations a real ability to have a trusted AI. In fact, we did a survey recently that will be up on our website, asking C level execs, what they thought about AI and what their key challenges were from an AI perspective. And trust really came out as one of the key areas, trust and security. So making sure that your data isn't like made available to really the general populous in terms of these generative AI models that are out there. Just from a business perspective, we see it as a real tailwind for us. So as you said, we launched a product called AI Unlimited. It was featured at the Microsoft Developer Conference. It was great to have the Teradata logo up behind Satya, as he was talking about Fabric and the Microsoft Fabric. And having Teradata been able to integrate right into the Microsoft Fabric, utilize all of the advanced services of Azure from an AI perspective has been a real catalyst for us.

Okay. So it sounds like you don't have any plans to build your own large language models. You want to be switch or lend and enable companies to bring open source, something that's been commercially built, open AI, whatever.

That's exactly right. I think the language model battle hasn't been -- hasn't even been fought yet in terms of who's going to win. So open AI, certainly a fantastic capability. But everybody is developing their own language model. A lot of them are, we make a large language model, so they're ingesting tons of data from lots of different sources. We actually have a viewpoint that the value from generative AI will actually come more from medium-sized language models or small language models that are trained in particular solution areas, be it customer experience or supply chain optimization because the cost of running these large language models can be considerable. And one thing that from a Teradata perspective, we're very interested in financial governance. So how do our customers control the cost of running their data analytics and now AI environment. And that's something that we can bring to the table to help our customers put in place a really trusted GenAI solution. I think the other thing is if we actually look at where organizations getting business value from AI today. It's not as much in generative AI. Generative AI is a nice way to present that to your users or your customers or consumers. But really predictive AI, decorative AI, some of the more traditional functions around that as in machine learning is really driving a lot of business value today. And that's where we see a lot of investment from our customers.

Okay. Does Teradata support unstructured data? Obviously, structured SQL is a big core part and transactional data, but yes, I mean, how -- because that's part of the new opportunity with generative AI.

Exactly. I was talking to one of our customers, a bank in Canada, and they were saying like they've 10x the data in object store than they do in their enterprise data warehouse. So as we've looked at how do we support object store, how do we support streaming data, how do we support things like video data directly from the platform and native object store. We've got customers now developing solutions that do just that, be it robots on the production line, streaming data into a native object store in AWS S3 storage or a restaurant chain taking video data, looking at the queuing system, the queues that they've in the restaurants and how to process that data. And so these are all capabilities that are built right into the Teradata platform today. Some really advanced analytics that our customers are just starting to use, and we see it as a great expansion opportunity for us as we move forward.

The other new product development, I think, is around open table formats. I think you guys announced that last quarter with support for Iceberg and Delta from Databricks or no, I think -- well, yes, they built Delta. But those are both open source. So what's the strategy behind that? How are you unlocking additional value? Why did you start supporting those open source table format?

Yes. I think open table format, it gives a real opportunity for organizations to use cheaper storage, object store storage, but then develop a kind of lakehouse construct around that. So structuring that object store data, utilizing open table format gives a low-cost capability in the market to really have that structured data in place. We saw it as an opportunity to increase our time essentially in terms of being able to access that open table format data. And we went into it with a very open mindset in terms of both supporting Iceberg and Delta. And we didn't -- the way our query engine works, we can support multiple different table formats as we move through time. And it's getting a lot of interest. In fact, it's picked up more and more over the last 6 months in terms of organizations, thinking about how to modernize their data environment, utilize an open table format. And the way that we talk to our customers about it is, look, you're always going to have to have that what we'd see, gold standard of performance from a storage and a structured data perspective and maybe your enterprise data warehouse that you use to close your books every quarter. Then you could have a silver tier and the silver tier for us is essentially Teradata manage an object store, utilizing our file system that's our kind of second tier of performance capability. And then the third tier is Teradata querying open table format on native object store. And so our strategy there has been, how do we make that OTF or native object store as performant as possible. If you think about the one thing that Teradata is known for is our enterprise price performance. So how do we get that open table format data towards, not on par with our gold standard of storage, but how do we improve that performance so that, that data can become useful for organizations.

Yes. That's a helpful landscape of the canvas that you're going after. Is open table -- support for open table, is that fully generally available now? Or what's the time line for that?

Yes. It's fully available in our VantageCloud Lake product. And we're going to continuously develop that capability to improve the performance, bringing some of our secret sauce to the table in terms of improving that performance of open table format.

Will that have any deflationary impact to storage revenue for you guys? If customers are saying, "Hey, you know what, I'm going to take some of this data that I've stored on Teradata, and I'm going to push it out to Cloud object store. Do you....

Yes. I think what we see there is that our customers are always going to have to have a lot of investment in what we'd consider that gold tier. So Teradata managing a Teradata file system on block storage. And our customers run absolutely mission-critical enterprise scale workloads and solutions inside the Teradata platform, where that kind of silver and bronze level performance levels don't meet their business requirements. And so the interesting thing is even if you -- and we see this -- even if you take the customer master record and you put it into an open table format, in order to close your books every month, you still need that customer master record and the gold standard of performance to close your books. And so we see making available these other tiers of storage and data constructs as something that's going to expand our opportunity inside our customers.

I wanted to shift gears and just talk about, obviously, a big part of your model is getting customers over to the cloud and you see better expansion rates, what you do. Where are we in that journey? Does AI accelerate the interest in migrating to cloud and kind of what percentage of your base do you think will get to cloud over the next few years?

Yes. So I think we've done a great job. We've gone from essentially a very nascent cloud business to over $0.5 billion at the end of last year of cloud ARR. So 4 years developing that $0.5 billion of cloud ARR has been a tremendous achievement. I think we do see the interest in AI and really generally the desire to utilize cloud native services. So if it's Azure ML or open AI services, we see that as more of a catalyst to move some of the workloads into the cloud. If we look at our general customer base, you can think about it, we've got kind of $1.5 billion of ARR. We've moved about $500 million of that into the cloud. 70% of the customers that run in the cloud with us also still have an on-prem capability and our ability to execute hybrid cloud environments, I believe, is second to none. And so we've still got a long road in terms of the opportunity to migrate customers to the cloud with us. And obviously, when they get to the cloud, they really grow with us. But we always see that due to regulation or data sovereignty or just the security standards that our customers want to implement that we'll have some customers that will always remain a 100% on-prem . So then it comes into play in terms of our partnerships with organizations like Dell in terms of running on their converged infrastructure in people's data centers. That enables us to focus and become more and more of a software business as we move through time. Software and cloud is clearly our strategy as we move forward.

Do you -- you guys do have an advantage of being able to offer hybrid over some of the pure-play cloud vendors? Is there an appetite that's changed to actually kind of move more workloads or keep more workloads on-premise for cost reasons?

I think our customers have seen 2 things. One, there's -- if you look at some of the banks and financial services environments inside the U.S., there's a question around regulation. What regulatory compliance are they going to have for the data that moves into the cloud. And we see that a lot in Europe as well. We also see data sovereignty requirements like making organizations want to stay in the cloud. And also, I think really super large enterprises. We focus on the G1000, the G10000 now. They'll always have enough scale to essentially have data center environments that they can operate at scale very effectively and efficiently. And so I think the reason that organizations are now using the cloud is to get access to these advanced services, to get access to things like the machine learning or AI services that the CSPs are making available. But our focus now as well is how do we make those services available on-prem. So working with organizations like Dell, how do we start to enable companies hit all the advantages of running in a CSP environment, given that to them on-prem in terms of running a language model on-prem and say, the Teradata environment. And so that hybrid world for us is a key catalyst for us.

Okay. Okay. Can we just touch base on kind of recent execution, you had a couple of large customer losses that had been in the works for a while, and then there's been some tougher sales cycles where deals are pushed. First of all, on the large customers, how do you feel about retention of your remaining top 20, top 50, whatever you want to call it? Let's start there.

Yes. So I think if you -- as we said, as we look through our earnings and as we plotted out the 2024, we did see -- we knew about some erosions that were going to happen in terms of our customer base. These have been in the works really for 3 or 4 years before we had the cloud offering that we do today. And indeed so, one of those customers, in particular, they wouldn't accept an evaluation of any product that hadn't been GA for 12 months. And so we knew about these major transactions that were going on and we factored that into our guidance for the year. I think as we look at our customer base and to your point, all of our top customers, but not only that, all of the customer base, we've a really good idea of what's going on inside our customers and we see great opportunity to drive more and more value for that customer base over time as we continue to invest and innovate in the platform that we've got, as we demonstrate how we can migrate them to the cloud for certain workloads, take advantage of those advanced services and advanced workloads. We don't see any change in terms of our customer environment or the marketplace environment, and it gives us confidence as we look out into the second half of the year in terms of our execution against the guidance that we put out there.