Episode Transcript
[00:00:00] Good morning everyone. Today I'm going to take a few minutes and talk about Oracle. And Oracle is a company I don't talk about very much, but I probably should not from the HR technology standpoint alone. But as a software and infrastructure company, they are setting the pace and creating in a sense, a new type of business that not only sells enterprise software, but sells compute capacity as well. So in some sense they're a hybrid between a workday type company that sells software and Google or Microsoft that sells technology infrastructure. And the bet that Oracle is placing, which I believe is a pretty good bet, is that over the next decade most enterprise applications and corporate applications and individual applications will be developed by AI models. And that application software that we used to build by hand will, will be dynamically generated by AI. And therefore you will not be necessarily buying software. You will be buying compute capacity in the form of either tokens or megawatts of energy. Now that sounds really bizarre because it's very different from licensing software by the number of users and the price per user.
[00:01:18] But that's where Oracle is going. And the analysts were very skeptical because their stock has actually dropped a lot over the last year. But yesterday they announced spectacularly positive earnings on both the infrastructure side and the software side. And I am really impressed. So let me talk a little bit about Oracle. Now I happen to know Oracle pretty well historically, not as much lately, because I used to work for one of their biggest competitors, sybase, in the 1990s and early 2000s. And this is a software company that has been always known for technology prowess, aggressive sales and marketing, and very strong focus on go to market. The company was founded by Larry Ellison and another guy who may not be around anymore. And it was originally the company that took the mainframe databases that we used at IBM and moved them to open platforms. They were. Larry was brilliant enough to figure out that if he could take the database software and put it on Unix, companies could build all sorts of new systems on top of it and free them from the grips of IBM. IBM was the by far dominant computing player in the 70s and the 80s. And you had to buy IBM mainframes, you had to use IBM proprietary databases. There was a SQL version of the IBM database called DB2 that Oracle just broke that monopoly and gave birth to a new market of open systems, which was then a very competitive market. I worked for Sybase, there was Informix, there was Ingres, there were a bunch of them that were unlocking the potential for you to build software on open platforms. And that was a Very interesting time. Not that much different from today because the Oracle Sybase is Informix. Database war was a war for infrastructure, just like we have the war for LLM infrastructure. Anyway, fast forward over the next 30 years, Oracle built a lot of corporate applications, sold the database very aggressively, outsold Sybase, even though Sybase had better technology. For a while, Sybase ended up getting sold to SAP. The other companies got sold to other providers and built a massive infrastructure of customers and trained Oracle database administrators and application developers, built development tools, and then they built applications. And the original Oracle Business suite was a fairly, maybe I wouldn't say elementary, but rather basic financial system. And they built more and more and more. Then they acquired PeopleSoft, which most of you know about. It was a very contentious, hostile takeover, which set in place the later birth of workday. But they acquired a bunch of CRM software to compete with Salesforce, got into financials, got into, bought a whole bunch of analytics tools to build Oracle analytics tools. So very successful company, not afraid to go through acquisitions, not afraid to restructure, not afraid to lay people off. I think one of the most interesting acquisitions they made last year or the year before was to acquire Cerner, which is the number two provider of hospital healthcare clinical systems, which is a massive, massive market. So they're very, very smart business people. They understand the software industry. And as the AI infrastructure world has evolved, about a year or two ago, Larry Ellison stood up in front of the Oracle community and I listened to this speech and I was quite impressed. And he basically said, we're going to build AI compute capabilities into the database. And the benefit of that potentially is that if you can vectorize data in your database, then your corporate data, which is sitting around in legacy systems or CRM systems or financial systems, is immediately available to an AI model. And so that's where they're going. And in the middle of that, they built a cloud infrastructure to support Oracle cloud applications. So they got very good at building data centers and running computer servers. By the way, in the early days they acquired Sun Microsystems. Sun Microsystems was one of the, if not the leader in open system chips and open system operating systems back in the days of Unix. And so they have had for years a lot of expertise in hardware, in operating systems and of course the databases around them. So this is a very unique company. You know, I would say Microsoft has this kind of expertise because Microsoft plays in all these different spaces. But most enterprise software companies don't invest this much in infrastructure or core operating system level stuff anyway. So that's moving along. And then they do this big deal with OpenAI to build Stargate, and they announced, with President Trump, 100 billion investment in a giant data center that apparently may have gone sideways. But what it really shows is the company's really leaning into this idea that software will be developed by AI. And interestingly enough, yesterday was the day of the Oracle earnings announcement, which was really, really good. I mean, the software revenue went up 13%.
[00:06:27] The infrastructure revenue went up something like 70%, but the infrastructure revenue is now equal to the software revenue. So they make as much money selling compute capacity to companies like Lucid Motors or Lockheed Martin as they do selling HCM software to JP Morgan or another bank. And they do both very aggressively and very well. And one of the things that Chris Leone, who runs the HCM product, just, and he and I were just having a little debate on LinkedIn, is they've also built an AI studio through the acquisition of a bunch of smaller companies that allow people to build AI agents in the Oracle infrastructure within Oracle applications. So they're way into this AI reinvention of HR software and enterprise software in a big way. Now, workday is not sitting down, and I'm going to talk more about workday later, but we'll talk about them and SAP later. So, anyway, so you look at the numbers and you see a company that is transitioning from selling licensed software to selling tokens to selling megawatts. And interestingly enough, when you go through the earnings announcement of Oracle, they talk about the fact that they delivered 400 megawatts of capacity, because one of the ways to measure compute capacity is the amount of energy that it takes to build that compute. Now, you can read my article for more details. And, you know, I think this is a really, really fascinating business story and kind of bullish time for Oracle, despite the fact that their stock dropped a day or two earlier. Jensen Huang, the CEO of Nvidia, wrote an article that wasn't very long, but it was really, I think, profound, essentially saying that in the stack of technology and software that we have purchased for our companies all these years, the very top of the stack has always been the application. You know, the bottom of the stack's the hardware, and then the middle of the stack is the middleware and the database software, and then there's security software and user interface software, and then on top is applications. And most financial analysts and buyers of software have always felt that it's the application layer that drives most of the value. I mean, it's. It's I'm not saying it drives more value, but it drives a lot of the value. And then underneath it, these other markets for the other layers have competitive, you know, markets of their own. Well, the point that Jensen was making, which I'm going to talk a lot about next week at Unleash, is that in the new world, a high percentage of the application software is generated dynamically by AI. And I don't mean just generated for a software engineer, I mean generated dynamically. When you go to an AI tool and you ask it a question, or maybe you give it a really complex prompt, you're essentially telling it to write a piece of software to answer your question. You may not see the software and it may not look like software, but it is dynamically assembling information for you.
[00:09:20] Now, that would normally have been done by a tool that somebody built or a code that you wrote. So what's happening now is the dynamic intelligence of AI that we sort of think about as an assistant or an agent is generating application code. And when we talk about our blueprint of agents and super agents, these application fragments or application pieces, you could call them Claude skills, there's different names for them, but they're basically little agents. And what Oracle did with the Oracle Agent Studio, and I will link you to that, is they built a tool set for companies to more easily build these small agents and connect them together into different patterns. They call them patterns, different types of agents, workflow agents, analytics agents, and other types of agents.
[00:10:11] So Oracle is basically saying to their customers, we're going to keep selling you the applications we've built for you. We're going to keep building stuff for you in all these different parts of business. And we're also going to let you use RStudio to add to the agents, change them, implement them and build around them. And we're going to give you the infrastructure to run this stuff. So this is a new kind of company.
[00:10:34] I mean, SAP, to my knowledge, has not built a tool set like this. Workday is assembling this through acquisitions, but they have not quite launched anything nearly as sophisticated yet. And Oracle's revenues are sort of an order of magnitude bigger than these other companies. Now, SAP is very big because they have many, many applications. Workday smaller. Salesforce is in sort of a different space, but has a similar problem. But what you see here is that the fundamental nature of an Oracle is different from the fundamental nature of a typical software company. In other words, AI is not a feature of Oracle. It is not a module of Oracle that you Talk to. It's built into their infrastructure, it's built into their strategy, it's built into their business model. Now for those of you that are sort of big business analyst follower types, it's a big bet. And I think the stock market has had a hard time absorbing this and understanding it because the pure play AI vendor vendors like OpenAI, Anthropic, Nvidia in the chip business, Micron in the data storage business are trading at extraordinarily high multiples because they're down in the lower levels of the stack where there's a massive scramble to get technology. Oracle's playing in the whole stack. So Oracle's relative valuation on a price to earnings ratio is lower. But you know, if you look at the value proposition of what businesses need, businesses don't want piece parts, businesses want solutions.
[00:12:04] I don't know about you guys. I don't want to buy a database and then buy a piece of hardware and buy this and buy that. I mean if I'm Lockheed Martin, of course I have to deal with a lot of, you know, technology based contracts where I do that, but most of us don't want to do that. We want a total solution. So the Oracle strategy, if you listen to their conference call, is very sound and targeted towards the business buyer. They claimed to have 2,000 new customers go live in the last quarter, 2,000. Now that also includes smaller customers that use NetSuite and some of their lower end products. But that's a lot. And they cited about 10 or 15 client wins over SAP and Workday. So that's sort of not huge, but that's significant. And it was a very, very positive, bullish announcement. And usually you can hear some amount of skepticism from the financial analysts who listened to these calls. I didn't really hear any. And then at the end of the call, Larry Ellison spent 10 minutes kind of describing the future of where Oracle is going. And I guess what I would say that I've learned as a, you know, business guy and an analyst for many, many years is, you know, you really kind of don't bet against people like Larry Ellison. He obviously is also involved in Paramount and the media industry. They also own 15% of TikTok. He's a very successful businessman, a very shrewd competitor, a very savvy, experienced leader.
[00:13:29] And Oracle as a company that, you know, many people questioned over the years, where they were going in some sense is a, is a beacon or a leader for others. Now I'm sure there's going to be a lot of criticism from their competitors. You know, this business software market is extremely competitive. There's a lot of fist fighting back and forth between the leaders, but it basically comes down to us and what we want to buy and what we want to use. So I'm going to do a little more of kind of podcast. I want to talk about Microsoft fairly soon and all the stuff they're doing with work iq and just to give you a high level perspective on what's going on in the technology market. And next week, I'll be at the Unleash conference. If any of you want to come talk to me, I'll be giving a talk there. And we're going to be releasing the new version of Galileo next week, and you'll hear all about that. So that's kind of my quick update for today, and look forward to talking to you guys again soon.
