The Six Five Pod | EP 281: Inside the Week in Tech - Oracle, Salesforce, TSMC & the AI Cloud Race

On this episode of The Six Five Pod, hosts Patrick Moorhead and Daniel Newman discuss the tech news stories that made headlines this week. The handpicked topics for this week are:

1. Market Turbulence and Geopolitical Tensions: Discussion of recent market volatility triggered by comments from world leaders. The hosts analyze the impact of US-China tensions on rare earth elements and tech industries.

‍

2. Tech Industry Developments: Oracle's Financial Analyst Day 2025 and AI-focused announcements. Salesforce's Dreamforce event and the future of enterprise software.

‍

3. Semiconductor Industry Insights: TSMC's strong performance and its implications for the AI and chip markets. Plus, an analysis of ASML's earnings and its role in the semiconductor supply chain.

‍

4. AI Infrastructure and Cloud Competition: The hosts debate Oracle's potential to challenge major cloud providers in the AI era, spurring a discussion of AI-driven growth in cloud and infrastructure markets.

‍

5. Economic and Market Trends: Examination of market reactions to geopolitical events and major tech announcements. Pat and Dan speculate on the sustainability of AI-driven growth and potential bubbles.

‍

6. Industry Events and Announcements: Coverage of multiple tech events, including the Dell Analyst Summit, Oracle World, and Salesforce’s Dreamforce. Hosts unpack the key announcements and their potential impact on the tech landscape.

‍

7. Future of Enterprise Software and AI: Predictions on the consolidation of AI platforms and agents in enterprise software and a discussion on the challenges and opportunities in integrating AI into business processes.

‍

For a deeper dive into each topic, please click on the links above. Be sure to subscribe to The Six Five Pod so you never miss an episode.

‍

Or listen to the audio here:

Disclaimer: The Six Five Pod 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.

‍

Daniel Newman: Episode 280, Pat, we're only about 125 episodes away from finally catching up with my bench press.

‍

Patrick Moorhead: There we go.

‍

Daniel Newman: It's not bad, but, well, you know, I mean, we do. We go small here. We try to stay. Keep the numbers low wherever we can. Viewers, engagement, quality of the content, and of course, the bench press, always very, very low. What do they say, Pat? If you can't do 225, you don't have a dad anyway. That's all right. That's all right. You'll get there. It's. It will. It's an injury, buddy. It's Friday morning, and where are you? In Dubai. Looks like you're sitting in front of the Burj. Dubai out there in the background. Where are you today?

‍

Patrick Moorhead: Yeah, so I'm definitely not in a bunker. And maybe this is a fake background. New AI generated on our platform. Or maybe I'm sitting in my living room with the view that I get, but I'm not home a lot. Daniel, you and I are both on the road. I'd love to just sit and. And look at this view, but I'm not. I'm not. I'm traveling. I'm getting stuff done, making things happen, and just trying to knock out work so I can enjoy some F1 here in Austin.

‍

Daniel Newman: Yeah. By the way, I said 280. You know it's 281, right?

‍

Patrick Moorhead: Oh, what's that?

‍

Daniel Newman: It's actually 281. Okay, so we're going on progressive overload here. So we are. We are pushing 281. God. You know, someday we're gonna get on this thing, and like, I was young when we started this thing. I am not young anymore. That's how long we've been doing this crap.

‍

Patrick Moorhead: I was really fat and a lot more fun when we started this podcast, I think.

‍

Daniel Newman: Well, you know, I. I've. What is it? Oompha was my new thing. Make Pat fat again. Make Pat fun again. 

‍

Patrick Moorhead: Yes. You know, I'm sorry, Daniel, but unless something tragic happens, I'm not going back. I do prefer. You know, I love getting up at 4:35, knocking out a workout, you know, just. Just getting on it. I saw your tweet about Las Vegas. And you know what, you can see that early in the morning I kind of did the same walking out and you see who's sitting at the bar, what's going on at the bar. It's hard to find a cup of coffee. But then you go into the gym, there's literally nobody there. You know, I mean it's hard not to feel like a superior being as in, in the memes, you know, all the memes.

‍

Daniel Newman: I'm walking down the hall, you know, you know, I'm going to the 10th floor of a nation and it's like a 24 hour gym. It's not great, but I'm walking down the hall and at least two with badges. So they've been out all night.

‍

Patrick Moorhead: I saw that too. Yes, yes.

‍

Daniel Newman: And they just looked wrecked and they both were going back with hookers and it was like, you know, and it's just like, wow. You know, like by the way, I also went to bed at like 8, so, just to be, you know, so people are like, well, you need to sleep. I'm like, I did sleep. I went to bed at exactly the same time. I would have gone to bed at midnight. I went to the event, I went to dinner, I went back to the room and I put myself in a coma and I went to sleep. I got six and a half hours, felt good, got a good workout in, jabbed all the peptides by 3:45, I was directionally feeling great. And by the way, I was up in the air by like 6am so I got the workout in. I got, you know, quickly I got dressed, I got a shower, got to the airport and got to San Fran because I had to get to Dreamforce. Talk about an absolute nightmare in San Francisco during Dreamforce. 

‍

Patrick Moorhead: Yeah, I mean, Daniel, we're not giving awards out for analyst CEOs or analysts who went to the most shows in one week. But I'm pretty sure that you would get the award. Real quick, rattle off bullet points, all the events you attended this week.

‍

Daniel Newman: Dell Analyst Summit, a very quick stop by NetApp Insight, Oracle World, Oracle AI World or Cloud World or Open World or whatever it is today. And then Dreamforce. And then today I'm going to actually go to, you know, a little, it's not really, I'm going to go to F1, you know, kind of, it's not really an event for tech, but it's an Event for me, actually, I was home yesterday. It was nice for, for a little while, to get a little bit of work done. I always like being in the command center, you know, in the chair. Yeah. Oh, yeah, it's, it's really nice. I'm thinking about trying to not ever travel again if I can. Look, so that would be great, but figure that out.

‍

Patrick Moorhead: If you unlock that, let me know how you do it.

‍

Daniel Newman: Well, I just want to work out and take Peptides, so, that's kind of the plan of what I want to do long term. But look, people like to hear our nonsense, but they really come here for the tech. So I will, I'll do the whole pivot thing now. We have a great show today. It is episode 281. We did, actually. We both went to most of the same events, even though we did it in slightly different orders. We were on the ground together at Dell. You made it to Oracle after me for Financial Day. I was in San Francisco for Salesforce Financial Day. I was at Oracle for the main day in the keynotes. But again, another week, more announcements. There were announcements made, maybe some that weren't probably more in the queue. We had an interesting week from the, you know, economy standpoint. We'll hit that later in Bulls and bears. You know, AMD's been on fire. We're gonna hit on that a little bit. You know, we will talk a little bit more. Surprisingly, we might talk about open AI. I don't know. We'll find a way to fit them into the story. Apple comes up this week. You know, Dreamforce came up this week. And then we're gonna, we're gonna do our simulated argument once again, where I basically acquiesce to having less experience than Pat. And he takes the cake every week. I'm gonna just stay humble here, see how that goes. And then we'll hit the Bulls and Bears, the part that I like a lot. You know, we're gonna talk about markets and why every morning we wake up to my new favorite, non. I have a new word that's better than nonsense, but I'm not gonna say it on the show because I don't want to get a beep, an explicit mark on our podcast just yet. But so much schmuckery going on on the, on the, in the, in the markets right now. Stay vigilant, people. Anyways, all right, Pat, look, we got a lot to hit, though, so I am just going to kick us straight over. Let's go to the decode. Pat. I've been just basically quoting you because you just put out so much great content on the financial day. You know what I love about financial days by the way, is I love that we go to events and we have to sit for three days and sometimes it's all worth it as industry analysts. But let's be candid, like in two hours. These financial days tend to give everything that we want. I love how fast and furious they are. And so I didn't get to Oracles. But you did. But you have given us your overall Oracle AI world impressions. 

‍

Patrick Moorhead: So Danny, I'd love for the financial analyst day to come first and then we could do the drill down at the, at the analyst, at the industry analyst days. But yeah, so I was at the event, I came in the tail end of the analyst event, spent a half day there and then went into this event and it was well worth it. And, and let me tell you why the meme is going in and, and this was from the information article by the way. I'm going to be doing a video with them on OpenAI on this day when this airs that said that Oracle or OCI had 15% gross margins on this big deal. And Clay basically got up there and he's the co CEO who runs infrastructure, did a really good job talking about segmentation and, and across four different areas like enterprise, cloud, distributed cloud, cloud natives and AI infrastructure. And as you might expect, as I think any sane person would expect, the more value that Oracle adds aside from capital, the higher the margin. So that's exactly what popped out. Right, so enterprise right at 65 to 80%, that's, that's amazing. Distributed cloud, which is more like a sovereign 40 to 60% cloud. Natives, which is kind of like non AI born in the cloud companies up to 60% and then. Drumroll. AI infrastructure was not 15%, it was 30 to 40%. So Clay did a good job going through that and I think it put the memes to rest. The other thing he did is he put up literally a six year schedule showing the economics of a gigawatt. Right. We've been throwing that around as a unit of AI infrastructure and essentially for Oracle it's $60 billion at a gross margin of around 35% and that's blended across multiple GPUs. And, and so you know, if you weren't convinced about the 30 to 40, the big, the big, the big boys are 30, 35%. The other thing I want to add is apps. Okay. You know we used to come to Oracle Open World and then it became Oracle Cloud World and now it's Oracle AI World. I feel like a lot of the oxygen was dedicated to OCI, which I totally get because that quite frankly is where the stratospheric growth is coming from. And even the Oracle apps folks, even Steven Miranda came on, it was very clear about the structural advantage that his company has sitting on top of OCI and how they deliver not only the applications but generative AI frameworks but most importantly agents. The final comment here was quite frankly that they came out and they brought out a 2030 estimate that was very high. I think a lot of the folks in the audience already had spreadsheeted it and even before they gave the new forecast Oracle was up seven points. And I think that was based upon what Clay had said. But essentially 31 CAGR on the revenue side and 28 CAGR on the EPS side. I love the flow through, right? You know, this horrible infrastructure business, you know, you know they still quote unquote horrible infrastructure business. 28% CAGR on EPS is absolutely amazing. Now the asymptotic nature of this happens in 2028 they go from 85 billion to 130. But between this year and, and 28 they've literally, they've more than doubled their, their revenue which is, is, is eye watering. And if you look at you know, companies like IBM at 60, 60 billion today with I think a 5 to 8% growth, it makes you know, Oracle just looks like an absolute darling. And they have to execute, right? They just, they have to execute. There cannot be a collapse in the AI demand which we've debated but wow, what a juggernaut.

‍

Daniel Newman: Yeah, there is no collapse coming anytime soon. Sorry people, I'll let the skeptics stay skeptical. I don't know why they're so against making money, but whatever, let them feel that way. Look, Oracle, you know, put a mic drop moment. It was a mic drop moment for Clay. I was very skeptical of the reporting. I got called out a little bit because I'm not supposed to be skeptical of the recording of that kind of reporting. But I, it was very hard because like I said it was, it was so asynchronous that the way they were kind of doing the, it's like, it's kind of the way financials work. If you kind of work back far enough or hard enough, you can find a way to make something look bad. The bottom line here is, if there are some big capex commits, there are some big investments being made up front. There are some limitations. Certainly power energy is going to be one of them. But the margins look good and, you know, you know, Oracle's obviously also doing a whole lot to make sure it has capacity, it has scale, it's building out these Stargate data centers. It's also partnering across the ecosystem on this. On the silicon side we saw this week, Pat, we didn't mention this, but AMD had another moment with Oracle that's worth noting. I mean, what do they have, 50,000? Yeah, 50,000 GPU commits starting in 2026. That's a really huge win for them. And it's separate from the agreement with OpenAI. Some people wondered if that was all tied together.

‍

Patrick Moorhead: Great point, great point, great point, Daniel.

‍

Daniel Newman: Yeah, so that was cool, but they actually announced partnerships kind of across the board. You know, they had a zeta scale cluster with Nvidia that they announced. You know, one of the things that I really thought was cool was their AI data platform announcement. This pad is kind of what you and I have been talking about, just endlessly proverbially connecting the back end and the front end in a seamless manner. The future is all about letting enterprise data and generative, generative AI speak. It's got to be in a trusted execution environment that's secure, governed and compliant. And once someone cracks that code and does it in this sort of OpenAI like environment, that is going to be the future of enterprise software is going to disrupt anything in the market, because that's what people want. People want context, whether they want to talk to their application, whether they want to be able to type in their query. They don't want static reports, they don't want to have five people on staff to be able to get them a visualization of their sales funnel. You know, they want something that works the way ChatGPT works. Or perplexity, you pick your poison because they all work a lot better. And I think Oracle's trying to break those silos and it has some structural advantages. Between its massive data center footprint that it's building and its other structurals, it has some of the most important enterprise data on the planet that it actually runs most, you know, half the world's businesses, you know, them and SAP, and so I think they have the opportunity to build it. But yeah, Pat, it was just, you know, go back, it was just a mic drop moment, in my opinion. It was like everybody out there after those announcements was trying to, you know, derate Oracle and Clay just basically said, we've got the power, we've got the commitment we've got, the access to the infrastructure, we can make money from this. And now the numbers look incredibly good. AI has completely created a new way that the entire AI industry, the entire cloud industry needs to be rated. It's going to be very.

‍

Patrick Moorhead: Daniel, just final comment on the data. For a decade, cloud companies were trying to rip and replace Oracle. You don't call your database redshift unless you're going straight after Oracle. Here we are in 2025 and you have, I'll call them database appliances called Exadata, literally Oracle's hardware sitting in all the top hyperscaler clouds. And that was kind of a. Oracle one moment. But also in this whole on prem versus public cloud, Oracle enabling its database to be in the public cloud, I think does facilitate the movement of many workloads to the public cloud. And we shouldn't overlook that. I mean, let's look back in five years, but this is really important. And now that Oracle has this new, I'll call it an AI infused database with the ability to keep your private data secure but also take advantage of public data, I think could be a big deal for the company and its customers.

‍

Daniel Newman: Absolutely. All right, there's a lot more to cover, Pat. And we've spent 20 minutes and we're only on one topic. Is that good? That's not very good. So Open AI news this week and other shocking news. Have you seen that meme with Bart Simpson? The one that's like. Say it Bart, say it Bart. It's like we are with Open AI.

‍

Patrick Moorhead: Yeah, Ryan. Ryan had set that one.

‍

Daniel Newman: I saw it some somewhere but I think it's sort of making the rounds.

‍

Patrick Moorhead: I wish I'd have thought of it first.

‍

Daniel Newman: Say it, Bart. We're doing a deal with OpenAI. So OpenAI had a couple pieces of news. I'm gonna bring up the first one, the Broadcom deal, which this one was really interesting because there was this thing on earnings where they had that 10 billion dollar new customer and everybody assumed that was OpenAI.

‍

Patrick Moorhead: Right.

‍

Daniel Newman: They found a piece of data out of OpenAI's data that suggested OpenAI. It raised 10 billion and they thought that was the 10 billion for it looks like that wasn't what it was. So it looks like a couple of interesting things emerged. Broadcom's got another 10 billion dollar customer buying accelerators. We don't know who it is. Some thoughts are maybe anthropic, some thoughts are maybe Apple, but it's someone else buying their XPU’s. Super interesting. The other part though is that OpenAI did in fact announce a plan to now partner with OpenAI so that Pat, you know, you've got the Nvidia deal, you've got the AMD deal and let's take all of the wind out of those two deals and announce another deal. This again goes back to my relentless Twitter nonsense of saying all hands on deck. There is no, it's not zero sum, there is no one winner here. If you can build it, they will buy it. Literally right now. If you can produce a gigawatt of energy, there's a buyer. If you can produce an XPU, there's a buyer for it. If you can produce a cpu, if you can produce a server, if you can produce a networking appliance, there is a buyer for everything out there. It is an absolute arms race right now. This is just another proof point. This isn't to suggest that all those deals aren't awesome for those companies and don't create a great forward looking set of metrics. It also doesn't suggest for a minute that there shouldn't be concern about who besides OpenAI is actually consuming all this stuff. But the TLDR is, it's a 10 gigawatts pattern. We've done some math on this, right? And I mean on a per gigawatt basis was it like 30 billion? 30, 40. It's some really 20 to 30 billion or something on a per gigawatt of infrastructure. So for OpenAI to do this with Broadcom, it's another massive commitment. OpenAI is not going to get any stock in Broadcom. This is not any sort of reverse deal. It appears they are going to buy it and anyone that knows their CEO Hock Tan knows how shrewd he is. So when those bills come due, Sam better have the money to pay them. So but I don't know Pat, I thought this was, it's 10 gigawatts of custom chips. And you know the, the stat that Reuters gave was that was equivalent to the power needs of more than 8 million households or five times the electricity of the Hoover Dam. So there is a big commitment to custom chips. Sam Altman is not picking winners. He's. Well actually I could say he's not picking a winner. He's picking winners in every category and he's completely changed the economics of AMD. AMD. Basically we estimate 3 to 4% more market share single handedly because of Open AI's deal. And he's rerating Broadcom, which now I think you saw TSMC came out with their earnings and they're saying accelerator growth for the CAG, for the TAM, CAGAR for 5 years. Pat, What? Do you see what they said?

‍

Patrick Moorhead: Yes.

‍

Daniel Newman: Mid-40s.

‍

Patrick Moorhead: Yeah.

‍

Daniel Newman: By the way, this was in the 20 to 30 range when we put ours out. And people were telling me I was high. Like, they were saying it. It was. It was a double entendre. They were saying the number was high, but they also thought that maybe I was smoking. I'm in Texas, folks. I can't do that here. Sorry. Not legal. But anyways, Pat. 45. So anyways, the deals roll on.

‍

Patrick Moorhead: Yeah. So I guess this means that the next deal will be cut with Intel.

‍

Patrick Moorhead: If it's everything you can ship, you know, Jaguar Shores and this new accelerator, this fanless accelerator. It's an island. I just always forget the name. I guess that's next week. We did see that.

‍

Daniel Newman: I sort of tweeted a smirky face on Monday after the last Monday deal. I did say, does this mean Intel gets a deal next?

‍

Patrick Moorhead: You know, I can't. I can't keep up with all your stock tweets, man. It's to the moon rocket ship Hodl. I'm waiting for that.

‍

Daniel Newman: You just don't get it because you're a boomer. I'm a millennial. This is how we speak to each other, dude.

‍

Patrick Moorhead: Raging Bulls. Altavista. It was basically the same thing that we see here. 

‍

Daniel Newman: Wait, wait. Are you saying it's a bubble? Get on the record.

‍

Patrick Moorhead: Of course it's a bubble. It's just not. May not burst.

‍

Daniel Newman: Well, no, then it's not a bubble.

‍

Patrick Moorhead: It has to burst.

‍

Daniel Newman: It can't be a bubble if it doesn't burst. If it doesn't burst it's a growth. It's. It's growth.

Patrick Moorhead: I didn't read the rule book, dude.

‍

Daniel Newman: So bubble. A bubble. You know, an infinite bubble. I mean, till when?

‍

Patrick Moorhead: I don't know.

‍

Daniel Newman: I'm gonna say this online today. I'm gonna say Pat calls the infinite bubble. It's not. It's a bubble, but it's never gonna burst. I think people love you. You could put a rocket ship after that. That tweet will go viral.

‍

Patrick Moorhead: Maybe I'll add a HODL too.

‍

Daniel Newman: Come on, man. Do cool. Let's, let's, let's. Let's quickly touch. By the way, Pat, you know, we saw some new partnerships with OpenAI this week. Salesforce, Walmart, by the way. So what we're seeing, though, Is, you know, I say this with love, but can't beat them. Join them. Like, the idea that people, you know, we just talked about, like, an open AI for enterprise, like a ChatGPT for enterprise. Like, you're starting to see, like. Like Salesforce. Mark Benioff was like, hey, let's make Agent Force available in chat GPT. Yeah. Like, why not incorporate sales data or Walmart saying, let people shop inside ChatGpt? Pat, this is going to be the future. We're going to see this inside these successful models. It's just. It's happening.

‍

Patrick Moorhead: Yeah. A couple comments here that I think. One thing that I think people are missing on Broadcom is that Broadcom's not just delivering chips, they're delivering the entire rack. And I had talked to some people in the industry, and they're partnering with ODMs and rack providers, so they are literally delivering a rack very similar to what AMD and Nvidia are doing. And one thing I love about the connector, the Salesforce connector with OpenAI is the simplicity. I'm using the heck out of these connectors from OpenAI right now, and it's literally changing my workflow and simplifying primarily right now, my planning of what I do. So good move. This doesn't mean that Salesforce isn't completely tied into their way of doing things, but I think this is just a pragmatic move.

‍

Daniel Newman: Yeah. Smart stuff. Smart stuff. So, Pat, have you been watching movies on your Vision Pro lately?

‍

Patrick Moorhead: You know, I don't even know where it is, Daniel.

‍

Daniel Newman: That's not fun.

‍

Patrick Moorhead: No, I know, I know. And there's just not a lot of movies even available. You know, apparently the Lakers are going to be in full, you know, full Vision Pro support. You can look around, look at the stands, kind of look at people probably move around there. I think that's pretty cool that the content absolutely blows. By the way, this new version of the headset is heavier, which is one of the biggest obstacles to people wanting to use this. And I know people are saying, hey, it's a. You know, it's all built into the headset, but, man, this thing is heavy. I am glad that I'm doing my neck exercises, Daniel, as part of my training regimen, to be able to wear this. No, I literally have no idea where it is. I think it's in the storage closet. It would be interesting to turn it in, if nothing else, to get the. The M5 is here. But no, listen, I think the big deal here is the M5 processor for MacBook Pro, iPad Pro and the Vision Pro. Apple is absolutely fired on processors. Make no mistake. They are showing people that with endless resources and being first in line and making the capex at tsmc, you can make some absolutely amazing things happen. And the other thing that they have to their advantage is they're a full stack right all the way from the chips to metal bending and everything in between. And that has an inherent advantage, particularly if you can deliver something on time and the rest of the industry can't band together and invest more. So hats off to Apple on that. I do think that their positioning is definitely muddy which says, hey, I've got this new, you know, M5 chip. But hey, you know what's even better? The M4 Pro and the M4 Max. But you know what the most highest performance chip is the M3 Ultra Ultra. So I think that is going to make a difference to their sales? I would say no. It is getting ugly. Is it a signal that Apple is, is potentially slowing down? I would say maybe. Final comment on here: the biggest thing that the M5 did is improved peak GPU computers. And I see a lot of people, a lot of developers use Macs. My son who is an intern at Nvidia prefers a Mac to do his programming on. So it's going to take a lot to unseat Apple, I think even from the developer crowd.

‍

Daniel Newman: Show me those neck exercises real quick. What are you doing? Can you do a profile real quick? Just so. That would be so ugly.

‍

Patrick Moorhead: That would be worse than that.

‍

Daniel Newman: Show me the neck up that you're in. Does Barton have you with us?

‍

Patrick Moorhead: Yeah, my physical therapist actually.

‍

Daniel Newman: It's like this is it. Yeah. Okay.

‍

Patrick Moorhead: Oh no, I do. No, I do what's called a strong neck or I get in here and it's actually a thing: a cone I put around my head with that's weighted.

‍

Daniel Newman: Can you put a video up next time?

‍

Patrick Moorhead: I don't know if that would be very flattering for me.

‍

Daniel Newman: It'd be funny for me though. That's what it's. That's what I'm asking.

‍

Patrick Moorhead: Of course you are.

‍

Daniel Newman: You really want to see it, do it. The profile shot would be great though.

‍

Patrick Moorhead: It's not always about you.

‍

Daniel Newman: Are you not entertained? I'm entertained sometimes. It's not about anyone but my own entertainment for myself. So let's talk a little bit about Dreamforce. You know this all in parallel, right? So many things going on at one time. Big week for Salesforce, Big week for Agent Force. You know I said about the SaaS market Super interesting, but the business performances haven't been bad. But the market basically just says we don't believe it. We just don't believe it. You know, I saw something unrelated about Adobe. Like Adobe, you know, in three years has seen its earnings per share go up 67% but the stock's trading lower than it traded at the lowest point in 22 during the drawdown. So the market is just kind of saying we don't believe in SaaS. And so that was the mission for Salesforce this week at Dreamforce. It was at both their financial day but also in their analyst summit was to really sort of how do we counter this whole sass is dead narrative? And it's no longer being sort of kept hush. They're addressing it, they're saying it out loud. They're, they're talking about their deals. You saw they made the deal with OpenAI, we just kind of talked about that. But you know, CEO Mark Benioff was really adamant. You know, he actually had Michael Dell there telling him, talking about their story. He's telling stories about companies like Pandora and FedEx and WebsoNova, you know, talking about 12,000 agent force deals that the company's doing. You know, it's one of those really interesting things, Pat, and you've heard me talk about this a lot, like there is going to be some sort of software consolidation like the industrial complex for software. Like people are not going to want to pay for agents in every application. At some point they're going to have to pick an agentic platform. It's going to be one of the mission critical ones, whether it's at the infrastructure layer, the platform layer, the application layer. But they're probably not going to buy agents from 20 companies. They're going to want one agent orchestrator that's going to work across all the data. That's what we talked about, the Oracle AI platform. You know, an agent force is attempting to do the same thing. You know, the structural advantage I see for Salesforce is just the entrenched customer. They are the, they are the interface to the customer right now. And you know, having that in early agentic use cases, which are a lot of them are customer experience, their chat, their interactions, their sales. It definitely gives them an opportunity. But I also do think, you know, the rise of the Google Gemini for business enterprise, the rise of OpenAI, basically being a unit app wanting to control and do everything for people is going to put some pressure. So the job of Salesforce was to prove it will Grow and succeed. You know, they did put out a new model, PAT, of about 9.6% or so growth, 60 billion in revenue by the end of the decade, while also having doubled their margin over the last few years. So that was really positive, a thing that came out of their, their analyst day, the financial part of it. But they also, you know, the event was all about agents though, all about Agent Force, Agent Force360 basically giving customers that enterprise grade agent, AI agent app that works. And they're very focused on the kind of human in the loop. And what's the story, what's the differentiation and where are they really leaning? They're leaning on the fact that you can't do this stuff in a trusted environment on any of these open platforms. Meaning that you need a platform like Salesforce to deal with keeping data safe, keeping it proprietary, keeping the right data from being shared with the wrong platform or the wrong data collection and the security of it all. So, you know, really at this point, the apps companies are all in on this trusted execution layer and that's what has to happen in order for these companies to maintain and stay, you know, on a growth trajectory. Is 9.5 to 10% growth annually enough, Pat, to excite the market? It definitely saw a nice boom, but it's not a nice immediate because I think the market was starting to wonder if there's growth at all. But having said that, I think over the long term, when you have businesses, you know, OpenAI's revenue is expected to grow from 10 to 150 billion in the same period of time that Salesforce thinks it can grow from 40 to 60. It's an interesting juxtaposition these companies are in, Pat. I think that they did a good job. I think they made some good announcements. I think they're showing some success with customers. But I do think there's some real proof points that need to be made at this inflection for Salesforce to continue to grow, but also change that sort of narrative that's just really pinning down not just them, but this whole space.

‍

Patrick Moorhead: Yeah. The entire software industry is in an awkward position right now. And it's interesting, Daniel, every successive conversation that I have with a CIO or a CXO, they're getting more bold in telling me what they want to do with software. Right. I literally talked to one this week that said that literally the plan of record is in 10 years to have no enterprise software. And you know, we're putting together the steps in order, in order to do this and it's a long journey. And then I think the way this happens is there's this decision point between time to AI and I'll use Salesforce as an example. Do I pay the extra money to get the generative AI magic on top or do I build that myself with let's say an agent from Microsoft or Google? And then it goes a step further which says do I just freeze my implementation and look at this thing like a database. I keep paying the fees on it, but the lowest, bare minimum. And then the next step is to come up with a data lake that has the data in there and you're literally just querying it. You know some companies that they're just at different stages here, but I'm seeing companies get, get, get bolder. I do believe that if you want time to AI, time to market, going SaaS is a good solution. And the reality is, enterprises aren't going to just standardize on one agentic orchestrator. I just don't see that as a, as a, as a reality. Like what has any large enterprise standardized on like anything, right?

‍

Daniel Newman: Is it 20?

‍

Patrick Moorhead: I mean, no, no, no, no, no. It's like, it's like two or three. It's like a cloud.

‍

Daniel Newman: It's going to be like a multi cloud.

‍

Patrick Moorhead: How that, that, that's exactly, that, that's, that's exactly right. And there is a scenario that says you have a front office, you have a front office orchestrator and you have a back office orchestrator and the rigidity of the back office tends to be a lot higher. I mean, my gosh, even looking at an enterprise's data mesh, it's like they're all different. And even if you had the magic product that could pull it all together, it would still be different. One topic that I did want to bring up that we haven't talked about yet regarding Salesforce. Salesforce got into the ITSM market, right? Like to compete directly with ServiceNow and both companies are impinging. ServiceNow brought out a CRM product at their recent big tent event and we're going to see a lot of this poaching, we're going to see a lot of this. But I do think ultimately we will see some consolidation in this market. We will see some acquisitions, we will see some smaller SaaS companies just completely go out of business and go bankrupt and get scooped up and integrated here on the big companies. Their valuations have to go down significantly. I do think it's a stretch if Salesforce thinks it can be the orchestrator for all enterprise workflows. Right. Like for, ERP and SCM and finance. I just, I just don't, I don't see that. It's not because technologically they don't have it. It's just about the way that people look at Salesforce. Right. You go to their show and they've got teddy bears and rainbows and green grass. And I think a lot of those back office people look at it and they're like, this is not a serious company. Right. I need a serious company to do this, like an IBM or, or you know, somebody like this to be able to do that.

‍

Daniel Newman: Yeah, it'll be very interesting to kind of see how this all plays out. I do believe there's going to be a lot of consolidation, but I don't, I don't think either it's going to be one. And even if it was to get to that, it's not going to happen instantaneously. It's going to be a step function. There's also likely going to be a hardware infrastructure orchestration layer, for instance. You'll see a certain orchestration of agents that are actually managing all your infrastructure that could be very different from the ones that are going to be managing all your business applications. We'll see where this all lands in the end, but it is a fast changing pad and it makes for good TV. So let's hit the flip. It's the fun part where we like to go after it. You know, you argue one side, I argue another side. You get it right. I basically put my tail between my legs and we walked into the sunset together. But let's talk about something we already talked about today. You know, Oracle showed some numbers, big numbers that all of a sudden don't really have them. As the fourth or even the third cloud, there's an argument made that they could become number one Pat. So, you know, can Oracle really challenge aws, Azure, Google in the AI cloud era? Let's see who says yes.

‍

Patrick Moorhead: My gosh, it's you, dude.

‍

Daniel Newman: You are the, you are the one that has to somehow convince the world and defeat your arch enemy Daniel Newman in this debate that Oracle will overtake the big three clouds.

‍

Patrick Moorhead: Yeah, I mean this is going to be, this is going to be really easy because they can absolutely get to number number three. I think number one and number two, I don't, I don't see that happening. Although interestingly enough, if you look at AWS at 240 billion dollars based on a 15% CAGR from 2024 puts them at around 240 billion dollars in valuation, sorry, term total revenue. And you look at Oracle, we'll be at 225 so it could hit that. But I don't think AWS CAGR is going to be 15% that people have been, have been modeling for them. Google on the other hand, very much right. If you assume a 25 to 30% CAGR from the current, I think it's like a $75 billion run rate. Oracle gets ahead of Google by 2029. I didn't do as much research as I should have coming in here, but to me it's about the proof points like how did this happen? So first of all, Oracle got in late to the cloud. Version one of OCI was not good. It was horrible. They brought in clay and a complete new RE architecture of the entire platform. They started off with bare metal which the other larger cloud providers had to back into. The other proof point I think is that if you look at now every one of their competitors is big in data, but that's Oracle's bread and butter. And if you look at the ability to pull that data in and the moves that they've brought in and the expanding amount of object and file data for the AI age, I think that is another proof point that while I think again Microsoft and the other hyperscalers have data products but this is large, gigantic amounts of data. So yeah, to me it's pretty simple. Just looking at the projected growth rates for the cloud companies, at least my math says that Oracle will come in at, at number three.

‍

Daniel Newman: Well, that isn't all that compelling to me. The real question was can they beat them all? And so I appreciate you making a strong argument to agree with me in advance. I win. Mic dropped. Seriously though, look, Oracle's had its run. It's made some great announcements, but it's one on being the cost leader. It's, it's coming at it from a standpoint of being efficient. What they don't have yet is all the bells and whistles and all the tools and all the software and all the developers. And so what I think they are going to be is one of the AI clouds, but they're not going to be the only AI cloud. And let's just be candid, there's a lot of deals that haven't been announced yet. So what we're doing right now is we're playing whack a mole with the deal making. So Oracle made its deal. Everybody knows how much OpenAI it's getting. But is this a suggested OpenAI isn't going to go make other deals? Is this suggest anthropic, which by the way is pacing next year to have almost the same revenue as OpenAI and is closely knitted to AWS isn't going to create some huge acceleration within AWS's revenue, which really hasn't seen any AI acceleration because they were so far behind on infra, but they have that huge structural advantage of the customer base that they already have. And when all of a sudden that customer base starts turning on AI, all those enterprises start turning on AI, I think their number is going to accelerate faster. So the problem here is just like we talked earlier about the tsmc, new numbers rerating AI growth, we hear about new energy numbers rerating AI growth, we hear about new fab and foundry numbers, wafer numbers, new memory constraint numbers. Every day there's a new number, there's a new deal, there's a new expansion. Well, Oracle got theirs early and therefore their number and projections are more clear and the future is more clear to them. Other companies and other ones of these cloud providers haven't even announced their deals yet. When their deals come through, they're going to get the same sort of lightning in a bottle. There's going to be new ratings. And if Oracle doesn't have some more deals to come behind this, Oracle is going to quickly fall right back to where it started at number four. Not a bad thing being number four. Not a bad thing at 35% margins. Not a bad thing having 4.6 gigawatts here in the United States and a whole bunch of aws, I'm sorry, a whole bunch of Nvidia, AMD and Broadcom chips. But unfortunately, Pat, all these other cloud providers probably still have a few aces up their sleeves, which when they come out, people are going to realize that Oracle got the first wave of growth, but they won't get the last.

‍

Patrick Moorhead: All right, you win. I'm gonna go dark. There I go. I'm in the light again. That was good.

‍

Daniel Newman: That was fun.

‍

Patrick Moorhead: Well, what was fun is I really looked at the question after we were done and I had to argue that they would overtake Azure and AWS us.

‍

Daniel Newman: Yeah, as I said, you kind of left me, you kind of gave me the win before I even started. But it's still a fun debate. They, you know, they, they, they, they're having a Good run. Let's just put it that way.

‍

Patrick Moorhead: Yeah. The one thing they are going to have to figure out and I, they have, you know, when I, when I talk to Oracle, they just completely shrug this off is where's your first party Silicon, folks? Where is it? You're going to be at a cost disadvantage in the long run. Azure really hasn't done anything with its custom silicon meaningfully yet. They're playing. This is hard and they've got to invest more years into it, but I just don't see how they'll be cost competitive on the compute side moving forward.

‍

Daniel Newman: It's interesting, Pat, my predictions are starting. Did you hear about the Army Broadcom partnership? They were going to chip. Who's that for?

Patrick Moorhead: OpenAI.

‍

Daniel Newman: For OpenAI, right. But what I was saying is my other prediction was that there would be an ARM broadcast collaboration for a Oracle specific OpenAI chip that would be fabbed out by Intel. So we're getting closer. The OpenAI arm thing is moving forward, you know, and sorry, the OpenAI ARM Broadcom thing. Oracle's obviously Stargate, so there's a lot to be said about that potentially being where they'd house a lot of this stuff. I don't know. I think Oracle will do a chip. I've heard the rumblings that they will. I think it will happen in Pat, you know, at the speed that Oracle's moved versus the speed that Microsoft's moved, Oracle's not that far behind. What year? Maybe two.

‍

Patrick Moorhead: Oh, from a custom chip?

‍

Daniel Newman: Yeah, behind Microsoft.

‍

Patrick Moorhead: Oh yeah.

‍

Daniel Newman: Microsoft basically scrapped and had to restart.

‍

Patrick Moorhead: I mean they, they did, they did. And the one thing that, that, that needs to be mentioned is the alliance with Ampere. I'm convinced that most of the SaaS apps are running on top of Ampere and not x86. Somebody pulled me aside at the financial analyst day and literally told me, hey, did you know that Oracle doesn't buy any Intel? They're all AMD anyways. I don't know if that's fact. I need to double click on that. Moving forward. I'd find it hard to imagine that they don't have intel when they're serving up some, some enterprise applications. But, we will see.

‍

Daniel Newman: We will see. All right, buddy, let's get to the last part. This was an interesting week. Turbulence all around in the markets. Let's do a little Bulls of bears, man. Fridays are, are really something. Last Friday was capitulation. Friday there was a Trump rant. President Trump went on a rampage. We're in a new world. We don't pick up the phone and call our political partners and adversaries and talk. We just tweet, apparently, or truth our posts and then they get reshared on Twitter. But there's got to be a Trump algo out there that, like when he shares something, the algo instantaneously creates buying, selling pressure. Because last Friday after he said more or less that Xi Jinping was, you know, holding back rare earths, asking for special licensing and starting and escalating a trade war, that started a selling motion. And then he came back with a second post that said something like 100 tariffs on November 1st. And it just sent the entire market cataclysmic. It was like the world had ended. You saw things falling so fast that you would have thought that we were back into April 1st time ever doing this Liberation Day type moment. Pat, we know we have an issue with rare earths. We know we've overtly created an over dependence on China for the processing of these. It's not necessarily getting them all there. It's the processing of these and then it's the ability to get them back. And having access to these rare earths is critical to industries like EVs, it's critical to chips and semiconductors. It's critical to many, many things that we, you know, that we need. Pat. And so interestingly enough though, only a handful of, I think it was Saturday morning, President Trump started the de-escalation process. Sunday night, the markets ripped Monday. Tuesday, everything went up. Wednesday, we heard there were some issues with credit. Maybe this thing with Xi Jinping isn't over. Markets went back down. I mean, Pat, what is all this schmuckery that is going on out there with us and China? Are we back in it or is this just like we are just waiting for headlines and completely erratic at this point?

‍

Patrick Moorhead: You're so funny. Like this entire topic. But yeah, let's go. So last Friday I got up on CNBC. Sorry, yeah, market with, with John Ford on, on overtime and you know, Trump hadn't done his olive branch yet. And I literally said this is all nonsense. You know, check the last six or seven threats that that happened. I, I don't see that this is going to be an issue at all. And you're right. And then Trump came back and noticed that. So net. I just think this is complete nonsense. And to try to trade based off of this and do the whipsaw, I don't know. I guess if you're a trader, you have to trade and, and do things like this. But in the end, I just think this is going to be all nonsense and we will arrive at something. Note the, the Rare Earth proxy, which is, hey, if you have a subsidiary or something like that, was done at the same time that we clamped down on chip commitment, chip equipment commitments to those who are on the entity list. So if you had a JV or something like that inside of China, you were a Chinese company or there was an affiliation, so, so we cut that off. So, you know, this is a tit for tat battle. A couple things, you know, also hitting as well. Even though I believe about 50% of what comes out of the WTO, there was a forecast that said, hey, we're gonna take a 7% global GDP hit for 2026. So, yeah, I don't know where this thing's going to end. I think it'll be a negotiation. If you look at Trump and the way that he works, I mean, he's a negotiator and I think at the end of the day, we will get there. The only the, the, I would call the barriers on the left and the right, the guardrails are going to be all about defense. Right. You know, China is, you know, the US's biggest nemesis at this point. I think vice versa. And that's the one thing that could keep what I would call a rational deal from moving forward.

Daniel Newman: Yeah. So sorry, I took that whole thing. I'm so excited. So excited. Yeah, like, you know, we are definitely. There are a lot of whipsaws out there. I think if you haven't learned by now, though, it's just emotional chess being played. And every escalation comes with de-escalation. Every de escalation comes with re escalation. And by the way, it's a great way to kind of, you know, help your friends get bitcoin at low prices. I don't know if that's true. I'm just kidding. Having a little fun, Pat. We can't get so serious all the time, but let's talk about a serious bullish call. Tsmc, man. TSMC absolutely crushed it. And this is a bellwether, Pat. Like, it really is an indicator of everything that's about to happen. It's a pretty good indicator for a few quarters ahead. Their guides, really great indicators of the, like, the year ahead. And, you know, when you look at all the calls or like, I don't know how you feel. I think I've now answered the AI bubble question on TV like no less than 12 times in two to three weeks. You look at the stuff that TSMC is saying about this space and it's like maybe there's a bubble. I like your point, an infinite bubble. I mean maybe there are, and I've said this publicly, like there are definitely drag-ons in the market like companies that are getting incredible runs that are like you're not even an AI company, you're not even doing AI. I don't know why you're up like 8,000% right now. But down, down the stream Pat, from raw materials to anything for chip making through anything that basically builds a server to construction to data centers, like this is not a short thing, this is not going to stop in 2026. Every indicator and these numbers from TSMC for instance indicates, you know, over the next five years we're going to see category growth and expansion at an incredible rate, higher than almost every forecast that's been put into the market. You know, when I said our 583 billion dollar number is starting to look increasingly like between pricing power that these companies are going to have, basically everything that can be made will be bought and the actual implementation of all this stuff, we could see the accelerator numbers with memory all in and network and go to a trillion dollars just with cost power, the cost power that they're going to have. You're seeing memory constraints, pat. You remember 2018, that period of time for memory and how constrained it is to get. They're thinking it's going to get worse. Which means that if you don't have a long term agreement right now for HBM and you don't have enough HBM for what you intend to produce, you're not gonna get it. Which that's going to give Micron and S. It's going to be creating incredible pricing power, incredible growth. And of course TSMC holds all the cards. I mean is there a company that holds more cards? I mean I guess you could argue asml, but that holds more. I think we'll talk about that in a minute. But like, they're expanding capacity here, they're going to build out more in, in the US here in Arizona still doesn't solve our problem with US IP but the overall performance pad of TSMC is a beat on every, on every line. It's a beat.

‍

Patrick Moorhead: Yeah. So hey, I'm going to challenge you a bit here. TSMC did not issue a forward guidance on CapEx and the whisper number is for a whopping 15% growth in 2026. And let's say best case 50, 50 billion. So you know, that's 20%, that's 20% growth. So if it's the bellwether and we're looking at 20% growth, I mean, I guess that's good. They did talk about land that they were buying. To me that's probably the only factual predictor that we have, that they would look at, at, at expanding. So I hear you, Daniel.

‍

Daniel Newman: Fair. It's a fair point that their revenue guide and their estimated guide, the 45 accelerator guidelines.

Patrick Moorhead: Oh, right. That's based on capex they've already committed to.

‍

Daniel Newman: So we're talking about, that's through 2030. So what are you talking about? That their, the lack of Capex going to slow that down after 2030?

‍

Patrick Moorhead: No, I'm just saying that there's no commitment for, for Capex. So should we believe that revenue number? I mean, I guess we should. I guess we should.

‍

Daniel Newman: I mean, obviously a guide's a guide and they could always be wrong, but TSMC is not typically a company that is emboldened with its guidance. It's not a, typically a company that tends to run out in front of its capabilities in terms of the way it tells its story. Look, there are, there are holes to be poked at some point, Pat, like the obvious ones like I mentioned with memory, like you can't get a memory, you can't get enough. If you can't get enough wafers, if you can't get energy, like if you can't actually build the amount of energy we're going to require to build all these data centers, like these are real problems. Like that is a problem. But in the supply demand type curve that's probably going to drive pricing up, revenues up and it's going to create, it's not like their factories or fabs are going to stop. They're just going to do everything they can. So then you can start basically enumerating that it's everything they're able to create with all the pricing power. So their growth may not come from volume, it will come from having more pricing. Because we know that every chip Nvidia makes, every chip AMD makes, at least for those two right now and pretty much Broadcom on the AI side are being committed. So unless that infinite glitch of money suddenly runs out and nobody's willing to fund these things, it seems that there's a path to a really bright future. And yes, I'm incredibly optimistic. What I'm most concerned about, which is funny because you're not on the other side of this, is when AI is doing all the work that humans do, where are the consumers that create 70 plus percent of the economy? That's the problem that I'm actually trying to figure out. If we build all this AI, how do people get money to buy stuff? So we have our, we both have our bearish themes. It's just this part that seems to be pretty.

Patrick Moorhead: Daniel, just to be crystal clear, I don't, you know what? I consider my comments on TSMC aren't bearish. What I'm just getting to is the facts. The facts are they didn't increase their capex and their forecast I believe was, was for, for that 30% growth, was not for 2030. But I have to go back to Tagger.

‍

Daniel Newman: It's a five year average.

‍

Patrick Moorhead: I'm gonna have to go back and I have the quote here.

‍

Daniel Newman: I think it was five years. There could be upside to its expectation of a mid 40% CAGR for AI accelerators through 2029. Okay.

‍

Patrick Moorhead: I was looking at their overall revenue.

‍

Daniel Newman: Yes, I'm specifically talking about AI accelerators.

‍

Patrick Moorhead: Okay. Yeah. It's interesting if you look at the chip industry and in totality the benchmark is applied materials because they operate in a lot more of these categories. But anyways, I think we've drained, I think we've drained this topic for everything it's worth. ASML brought out earnings as well. It's exactly what you thought it would be. Good discipline on margins. They echoed what TSMC said. I mean obviously they're the least lithography darling for EUV and NAV going into the future. They're echoing that now. That's balanced by the pullback in China which should be a surprise to absolutely no one. Daniel, anything else interesting about ASML that you saw?

‍

Daniel Newman: No, but I think to your point, they are a great one to watch for that five plus year horizon. You know, are, are the foundries buying the capital equipment that's going to be required to, to. Because ASML is the one stop for the, for the leading edge. So keeping an eye on ASML's performance and their order flow is always a good way to indicate what's going to happen in the future. These things can't be stood up quickly. They can't be stood up quickly. So Pat is a good show. I feel like, I feel like it's just another week and you know, next week, you know, I don't know what the hell we're gonna wake up to on Monday morning, but it doesn't seem to ever stop. So I hope you have fun at F1. We won't be together, but we're both going to make an appearance, which will be great. I'm also going to try to actually stay home a little bit this weekend because frankly, next week's another big week of travel. Followed by another big week of travel. Followed by another big week of travel. It's going to be pretty relentless now through Thanksgiving at least, or a little bit after Thanksgiving.

‍

Patrick Moorhead: So, yeah, it's week four for me. I've got four more. Four more weeks to go.

‍

Daniel Newman: Stay strong. Don't get injured.

‍

Patrick Moorhead: Well, I already did that.

‍

Daniel Newman: Don't get any more injured.. And you have a good weekend. And all of you out there, all of you that are watching this show, you probably caught this after our weekend. So you'll have to check in next week to see how much fun we had and what the heck happened this week, what announcements and what deals are going to come down the pipeline, how big this bubble has gotten. And is the Infinite Glitch ever going to run out of money? I wish I had the Infinite Glitch. Pat and subscribe. Be part of our community. We appreciate all of you so much for tuning in. This was episode 281 for this week, though, saying goodbye. See you all later.

‍