On this episode of The Six Five Webcast, hosts Patrick Moorhead and Daniel Newman discuss the tech news stories that made headlines this week. The handpicked topics for this week are:
- Qualcomm & Google Announce RISC-V Wearable Platform
- Lenovo 2023 Analyst Event
- Oracle Netsuite SuiteWorld 2023
- US Bans China Export of More NVIDIA AI Cards
- AWS Chip Lab Tour
- SAP Q3 2023 Earnings
For a deeper dive into each topic, please click on the links above. Be sure to subscribe to The Six Five Webcast so you never miss an episode.
Watch the episode here:
Listen to the episode on your favorite streaming platform:
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.
Daniel Newman: Hey everyone. Welcome back to another episode of The Six Five podcast. It is 7:00 AM and if you’re looking at your watch and you’re wondering why we are on early, well, it’s Friday, it’s October 20th, and it’s the US Grand Prix and Patrick’s flash in the splashing because he’s got the paddock passes and well, Pat’s just a luckier person than I am and what can I say? But someday, if I work really hard and I continue to put in the effort and do great Six Five podcasts, Pat will get to go to another US Grand Prix. I’ve waited so long to do that.
Patrick Moorhead: That’s so good. No, Dan, work hard, work your butt off and you can maybe buy me a Lambo.
Daniel Newman: Someday, maybe. Can you see it? Twinning, Lambos, left and right side by side. Lambos, Porsches. I don’t know. We’ll have to pick something awesome. And someday maybe we put in a shift we could get there. But listen, we’re really committed to this show, and I’m telling you everybody over the next few weeks, it’s going to be a bit of a whirlwind, but we will do everything in our power to each and every Friday, or maybe it’ll be a Thursday, maybe it’ll be a Monday, but to bring you our weekly show because there’s so much going on in tech, we’re staring down the barrel of another massive wave of earnings. We only have a little bit of that today, but we also have had a lot of really good technology news that we need to break down and analyze here on the Six Five. And we’ve got some good stories this week.
So we’re going to be talking about a little bit of Qualcomm, Google, and a RISC-V agreement. We’re going to talk about the Lenovo analyst event that we spent two and a half days at. We’re going to talk about Oracle’s SuiteWorld. The US is after NVIDIA and certain chips again and seeing that they don’t go to China. Pat and I had the chance to tour the Annapurna labs at AWS here in Austin, and we’re going to talk a little bit about that. Pretty cool. And then we’re going to end talking a little bit of earnings. It’s the start. Yep, I’ve said it. We sometimes don’t love it, but we got to talk about it because the ground truth of the technology industry.
Pat, good morning. I’m going to do a quick disclaimer. We’re going to get the show on the road because somebody’s got to work at the Grand Prix today. Anyway, all right, so if you haven’t listened to this show before, I’m going to ask you why and I’m going to suggest you hit that subscribe button because it’s awesome. But in all serious, this show is all about breaking down and analyzing the biggest news and events of the week in the tech industry. We’ll do as little news as possible. We’re going to tell you why or so what? Not what and how, like so many other podcasts do. And just a reminder, this show is for information and entertainment purposes only. So while we do talk about publicly traded companies, please do not take anything we say as investment advice. Pat, Google and Qualcomm are going all in on RISC-V. Is it the end of Arm?
Patrick Moorhead: Wow, you kind of jumped right in there, didn’t you?
Daniel Newman: Sorry, I was feeling it this morning.
Patrick Moorhead: No, no, never be sorry dude. Just be you. No, here’s the deal. So I do need just to explain kind of what’s going on. So Qualcomm made an announcement with a supporting quote from Google, essentially saying they’re going to build a wearable ecosystem based on RISC-V. There were no divisibility dates or anything other than the announcement. I don’t think that means it’s vaporware. In fact, it’s very much real. Qualcomm has been part of the RISC-V consortium for a while. And back in September they cranked out a pretty big press release on their continued commitment. And quite frankly, again, almost everybody has RISC-V in some way, shape or form AMD has controllers and NVIDIA does on their graphics cards. I believe that Qualcomm has some on their modems or their RF. I’m not exactly sure which one, but essentially think if you don’t know what RISC-V is, each piece of software speaks different language.
Some speak x86, some speak Arm, and this new one called RISC-V. In the past there were architectures called MIPS. You have a different instruction set architecture for the power platform at IBM as well as their Z mainframe as well. So a lot of different instruction sets. But I think the big question I had is, and the question I would get is, Hey, what does this mean? The Qualcomm wear ecosystem isn’t huge. Apple dominates in wearables and there are some low end vendors who supply a lot of the Chinese. So what does it mean and why should we care? Well, there is a distinct potential that if Google and Qualcomm successfully pull this off in the market, that this could lead to a RISC-V smartphone. Okay, and we’re looking at a 1.4 billion unit market that today is a really 100% dominated by Arm and it is a heavy turn.
But let’s look at some previous heavy turns. Peck, Intel used to do smartphones and they had an award-winning, award-winning smartphones that they powered. And what they did is they did instruction set translation in the cloud when you would actually download the application and you’re like, well, how does it work like that? And why doesn’t it work like that in Apple? Well, the oldest modern smartphone app is probably five to six years old that runs on Android or iOS and they use IDEs and they use very high level abstracted tools maybe with a section of games. So even Intel was running Android applications that were initially written for Arm. They were auto translated in the Android app store. So another example is Apple, and what they did going from Intel to Arm, and one of the keys there was backward compatibility and the ability for that chip to run x86 instructions until it wasn’t pretty.
But when you’re bringing that much performance, probably a 40% performance per watt boost over x86, you can get away with maybe 10% overhead of doing that. The final example I’ll use is Arm in the data center, where you’ll have applications. It could be 20 years old, it took Arm a little over 10 years to move that you had to do the OS, you had to do malware, you had to do application application layer. But in about 10 years now we have AWS doing a full up graviton for their entire data center. We have Ampero One running the entire NetSuite layer for Oracle. So yeah, this is possible. Now, how difficult is this going to be doing smartphones? It would be more difficult than a wearable, less difficult for a PC in the data center, probably take about five years. That’s just an educated guess to get this right. And the long poll is always games where they are doing some deep instruction level work.
Daniel Newman: So first of all, am I in?
Patrick Moorhead: You’re in, baby.
Daniel Newman: I’m in. Second of all, you have to wonder where it’s at in terms of the things that aren’t being announced. And so we know that it’s been several months now since Arm brought its lawsuit to Qualcomm. And we know that with that, Qualcomm not only is arguably the world’s largest law firm, kind of joking, but kind of serious, successfully defended an onslaught of attacks from basically every FTC in the world and Apple all at one time. And with that, I think Qualcomm certainly is going to be thinking very significantly about hedging its strategy long-term. And of course Pat, I think you made some really good arguments about where that will start and directionally where that might go. It’s going to start with simpler devices where risk architecture is less risk potentially to be able to develop around and over time likely see it expound itself into other areas, expand itself into other areas.
As you suggested, Pat, I think like you said, I think there’s lots of companies exploring risk right now and the possibilities and as Arm continues to grow, change the way it licenses, change the way it charges to make sure that it’s meeting the demands of its shareholders that could open up doors for open source that could open up doors for RISC-V. Remember, Arm is a risk architecture. So it started there and of course found its way to disrupt Intel and others taking significant market, not the majority, but significant. And over time you’ll expect new architectures to find their way into the market like this. So where does it head from here, Pat? I think there’s probably work being done in those more advanced devices that you mentioned around risk, but I do think it’s early days. I think your estimate of five years is probably pretty close into the ballpark of where it might be.
But I do think it’s coming. I do think I’d say on the record, I think these companies are going to hedge, they’re not going to let themselves get into a situation where they can be controlled by a licensing agreement if they don’t have to be. And we’ve seen it happen to every company that has substantial licensing that over time they want to have more control of vertical integration, they want to have more control over margin and pricing. And if RISC offers that opportunity, RISC-V offers that opportunity. I could see it going more in that direction. So you hit the news itself. I’m just kind of philosophizing on what’s going to happen with RISC-V, but I think you made some really good assessments there and those additional comments are mine. So all right, we’ve got to keep moving here. Let’s talk about the Lenovo analyst event.
Pat, you and I spent the better part of what two and a half days in a beautiful, green Raleigh Durham, RTP at Lenovo’s headquarters, and we were joined by really their entire global executive team from overseas. We had the honor of hearing from their CEO who goes by YY, literally the two letters, YY, that’s how he presents and that’s how everyone calls him when we talk to him. And he was really generous Pat with his time. Not only did he spend some hours with a small group of analysts, but he also took us for pretty remarkable dinner in Raleigh. Entire menu that was created just for this particular event. And Pat, I don’t know about you, but I’m still full.
Patrick Moorhead: No, I am. I mean I made the mistake of thinking we were done. We were halfway through and I had eaten so much Peking duck that I couldn’t even eat any of that lobster.
Daniel Newman: Oh, but it was so good. It was so good. Yeah, I did the exact same thing, the fifth, sixth, seventh, and eighth courses. I was like, “Oh my gosh, I shouldn’t have eaten so much of the first four”, but really great dinner. But that’s a small fraction of what we experienced during these few days. Look, Lenovo is a very diverse company that covers a immense amount of ground. It’s everything from the PCs and devices, which is what it’s probably most well known for. And whether you talk to YY president of international markets, Matt Zelinsky, head of infrastructure, Kirk Skaugen, and they would all tell you that when you had talked to people, they’re going to say they’re known for the device, the ThinkPads. That’s what Lenovo’s known for. But over the last few years, it has seen its revenue distribution change substantially above 40% of its revenue now coming from its infrastructure and services group.
And while I really focused on a, that the company’s core PC business is likely turning the corner. You heard him say the market had bottomed. And I think we’ve been waiting some time to start to hear that the chip CEOs and the OEM CEOs say that with some level of confidence. And he was able to say with some level of confidence that he thinks that we’ve arrived and that the market recovery is likely going to begin. Now if you listen to Jerome Powell’s fed speak this week, you would hear he still thinks inflation is high. He’s turning the screws on interest rates. And that worries me a little bit. The market took an absolute dive on the comments and with higher rates you would see even lower growth for a lot of companies. But I’m going to stay positive here. The company’s infrastructure business has had about 12 quarters in which it’s run at a really great clip under Kirk Skaugen’s leadership strengthened, storage strengthened server.
He did an absolute victory lap of wins that the company had had both partnership wins, ecosystem wins, and customer wins. The company also really leaned into its One Lenovo and its SSG ISG IDG strategy. Only about 25% of the company’s customers buy more than just the PC, the IDG business. So while the revenues distribution is getting more even, they still have a huge customer swap that are not purchasing across their portfolio. Now, some are consumers and the consumers will never buy infrastructure, but getting to a bigger set of their enterprise customers with their services and infrastructure is going to be a huge opportunity for the company to grow. And then finally, I’ll say this, Pat, I feel like AI was brought up at least a couple of times and by a couple, I mean a couple of thousand times it feels like that was the focus.
And look, there’s this AI continuum right now and every company’s vying for what is our place in the AI world. And companies like Lenovo and HPC have a lot of provenance and pedigree. They have a lot of history selling GPUs for data center, and these are kind of the AI links. But from a service standpoint, building out a vertically integrated solution for retail where you build edge capabilities with AI for computer vision and machine learning using OEM partnerships, cameras and integration and software. I thought that was probably one of the interesting ways that they’re really getting to AI is through vertically integrating, going out to market with solutions and offering it across the portfolio, saying we can put infrastructure services and even devices together to offer a fully integrated AI solution. That’s kind of the One Lenovo that I envision.
Look, I could go on for hours because Pat, we had like 25 hours of content, but these are kind of the big things that were on my mind. I’d like to see more and more tie together of the whole portfolio. And of course I want to see that software data layer addressed a little bit more by Lenovo if they want to be serious about their AI strategy. But when it comes to pickaxes and shovels, the company has a good story brewing.
Patrick Moorhead: Yeah, I need to watch what I say because it was a blend of public information, NDA information and a bar good information. So I’m going to probably say less than more. In fact, even the slides that the company said we could share had Lenovo confidential on it. So it was a bit confusing.
Daniel Newman: I don’t think I said anything I couldn’t did. I?
Patrick Moorhead: Don’t know. We’ll see, I guess later this afternoon.
Daniel Newman: I don’t remember. I noted things that had the, okay, go ahead.
Patrick Moorhead: I don’t think you did. I’m not saying that for that. I’m just looking through my notes wondering what I can say and what I can’t. So on a macro view, this is the first time I felt there was truly a One Lenovo and I looked at different things for sources of truth and even basic stuff like do the slides look the same and are they supportive of each other between IDG ISG and SSG? And the fact is that they were, and it was very well articulated for the first time, I believe that there’s One Lenovo. And I think that’s powerful, right? And it’s powerful. And there’s a couple of ways you can get synergy. You can get synergy across the sales and marketing plane, you can get synergy across the product plane and then you can get synergy across the solution plane. And I saw collaboration across all three, which was interesting.
And a lot of those technologies came together with AI. How was AI being leveraged in services, the data center and the edge and in devices. So that’s really good and that’s really powerful. And I don’t think right now it comes at the individual products being weaker because that’s always the trade-off. When you’re going for a synergy play, are you still best in breed in the markets that you’re in? IDG is just an absolute runaway success. The leadership that Kirk has provided really is-
Daniel Newman: You mean ISG, right? Just to be clear.
Patrick Moorhead: Sorry, ISG. Yeah, the data center and the data center edge group and the leadership there. I mean, just look at the numbers. I mean 5x, 10x getting back to profits now, it’s not that profitable of a business, but I believe that the company is still in scale mode because it serves the hyper-scalers to get a basis of volume that it could get good commodity prices. But man, I saw some logos that were not Joe’s Gas Station like you would’ve seen five years ago. These were Fortune 500 healthcare companies, these were major retailers out there like McDonald’s. So they’re really making it happen. And then when I look at the devices group, you actually have smartphones, consumer devices and PCs. PCs are still rocking it. Lenovo is still number one and ThinkPad is still number one in commercial, which is pretty amazing. And they really laid out what they wanted to do and leverage AI that I thought was very clever.
I can’t go into, they’re not just aping what Intel and Qualcomm are doing at AMD, they’re coming up with some pretty novel stuff themselves, which you will see soon. And then on the services side, and by the way, this is another reason why the device’s numbers and the data center. Hey, you’ve got background noise that I keep hearing. Oh, where are you going buddy? Come on. Oh, okay. Just mute yourself if you got that. But it’s because the services profits and revenue are taken in SSG, right? And you don’t see any other, you don’t see that at Dell, you don’t see that at HPE. It’s an absolute separate division. So the hardware margins of the hardware margins, which is pretty crazy to me. But SSG is on the move too. When it first formed, I was super skeptical because quite frankly, Lenovo five years ago was a hardware company.
And going from a hardware company to a services company is really hard. The company made acquisitions, the company completely reorganized, and Ken Wong, I think has done a fine job up to this point making inroads into that market. They focus on three specific solutions that they’re prioritizing right now, hybrid, cloud AI and workforce solutions. And those are just low hanging fruit where the company can leverage its premises in hardware and then get really good at the as a service offering. So I would say my view of the company took a tick up if nothing else on the One Lenovo, I agree with you on the data center AI side, it’s something that’s missing from their story. How are these companies going to connect ERP data, SEM data, CX data? What is Lenovo doing to help that? They’re very focused on, I would say in-depart type of solutions, which is how can I do better with the CX data?
How can I do better with the ERP data as it relates to generative AI? And to their credit, they’ve taken all the way to the vertical side of it, partnering with, I would say some of the smaller startups that are killing it in FinTech, in retail, different verticals like that, warehousing, transportation. So yeah, my view definitely went up. I think my last comment is the company needs to re-look at the way it spends its marketing dollars. If 43% of the revenue is non PC, are they actually spending 43% of their marketing dollar on SSG and on ISG? Doesn’t seem like it, but I don’t have the spreadsheet in front of me, but it doesn’t feel like it.
Daniel Newman: All right. And I think that for both you and I’s sake, we believe that if you get the momentum in one particular part of the business, like ISG, you want to lean into that and give our friend Flynn some room to grow, we would say. So let’s run into what was going on this week, Pat on the other side of the country, we weren’t able to make it, but you did a great interview conversation and analysis piece with NetSuite CEO, Evan Goldberg, I read it on Forbes, very impressive. Talk a little bit about what went on this week at NetSuite world, Pat.
Patrick Moorhead: Yeah, it was great to get with Evan the day before the show announced, and I would’ve liked to have been there, but my dance card was filled twice. So a little bit of background, if you don’t know about NetSuite, they are the leading medium company provider for enterprise business suite, right? They do what SAP does. They do what Salesforce does, but they do it in a single suite. Hey Matt, buddy, how you doing? How are you?
Matt Newman: Good.
Patrick Moorhead: Good. I love that headband.
Matt Newman: I’m going to run at least a 37 laps at school today.
Daniel Newman: All right, go get them buddy. Thanks for saying bye to me. I’ll see you. I’ll wave from the race.
Patrick Moorhead: Oh, that’s so cute. That was Dan’s son.
Daniel Newman: Hey everybody, thanks for letting Matthew come join our pod for just a minute, everybody. I’ll be there with AWS and Ferrari.
Patrick Moorhead: So good.
Daniel Newman: Thanks buddy.
Patrick Moorhead: Yeah, NetSuite does a business suite and in fact, more unicorns use NetSuite than any other package out there. But anyways, what did they announce at SuiteWorld? Imagine this. I mean, I’d have to think for hours to think of what they could possibly announce for SuiteWorld after the last nine months. No, just kidding. Essentially they announced generative AI across the entire suite, okay? Whether it’s ERP, whether it’s SEM, whether it’s CX, it’s across everything. And one of the examples that is just simple, and one of the benefits of having a suite like this is NetSuite text enhance. Think of text enhance as maybe you’ve used the generative AI tools inside of Gmail where it says, “Hey, what do you want to write?” And then you just type what you want to write, who you want to write it for, and it cranks it out. Well, this entire feature is across finance, accounting, HR, supply chain, sales, marketing, customer support functions.
And one of the things that I think people sometimes miss is the value, for instance, this one feature text enhancer works the same across your entire platform. You don’t have to have separate training for everybody. It’s like one training for pretty much everybody. And you figure out how to learn that. And in smaller companies, you have a lot of people who are doing not cross-training, but cross-working, your AP people do some ar, your AR people do some AP, your revenue generators do some marketing. So I think there’s an incredible amount of value in this. Other examples they talked about is creating special narratives for collection letters. Dan, you and I can relate to that. I mean, we have these giant Fortune 100 tech companies who love to not pay us and leverage, oh, you’re still muted.
Daniel Newman: They pay really slowly, really slowly. The Bank of Pat, as I like to call it, the Bank of Daniel.
Patrick Moorhead: They’re financing their trillion dollar companies on the backs of companies like ours. And with this you can write a personalized collection letter and think about the power of that. If you have the authorization that you get access to accounts payable or potentially a CX engagement, it will write this thing for you. And this is one of the simplest examples of how generative AI can bring value to a company. And Dan, I could go on and on and on. There were another major announcement maybe on EPM that you can cover. But yeah, NetSuite, generative AI across the suite, one data plane, no changes to security, no data integration, nothing.
Daniel Newman: Yeah, you did cover a lot of ground. And Pat sometimes the best indicator is understanding whether or not a solution can really work for business and our business is using and adopting it. And NetSuite’s growth rate speaks a lot for the market. And it also is probably the undisputed number one application for what I would kind of call your fast growth companies. Fast growth companies that are going public, PE owned, kind of that large inner interim growth stage to becoming a large fortune type company where Fusion ends up being the platform. Oracle is sort of renowned for its ability to close its own books in 11 days. Think about that. Companies doing tens of billions of dollars. I know companies that do $10,000 in sales a quarter and can’t close their books in 11 days. It’s pretty remarkable. But you hit it on the head Pat, AI was the theme.
And yes, there was an interesting enterprise performance management, which is really all about bringing together a number of the company’s capabilities, sorry, a number of the business capabilities, budgeting, forecasting, planning account reconciliation, putting all those things together and creating better management, better visibility, better dashboards. But laying in AI both from a underpinnings of the way the software interacts with its users and then out front being able to do text creation for what you mentioned in collections or sales notes through CRM. That kind of inclusion of AI is what the customers are looking for. Look, you and I’ll talk about a chip lab tour, we’re going to talk about AI chips going to China in a little bit here, but most enterprises, businesses large and small, when you think about AI, what they’re really looking for is how AI, how is AI embedded in my user experience when I’m in the application and I want to understand my 10 best sales opportunities, can AI help me understand it?
Can it help me write a good follow-up email? Can it predict the likelihood that that deal is going to close? Can it tell me the ones that I should not bother with? Can it tell me which customer might pay faster and which one probably won’t? I mean these are the kinds of things that businesses are looking for and I think NetSuite’s doing a really good job across the stack. It’s a competitive market through parts of the NetSuite portfolio like the CRM side’s always been a little tougher. It’s got a very strong foothold in the ERP side. Having said that, with the CX features it’s building, it’s becoming more and more compelling as we go. All right, I won’t go on. Read Pat’s article. It’s really good. It’s really, really, really good and it covers a ton of ground there. So let’s talk Pat about NVIDIA.
You actually wrote an article going back, gosh, Pat a few months now. And we all like to do a victory lap. I like to talk about when I called out NVIDIA becoming the next trillion dollar company like three plus years ago, and sometimes it’s getting things right three or four years back, sometimes it’s getting things right three or four weeks back. But Pat, a couple months ago you basically said Joe Biden and the various chip controls he’s putting into place simply aren’t going to be enough to stop the routing of advanced accelerated computing chips from getting into China. And I think so far, Pat, it’s proven to be correct. We knew this going in, we know that most of these kinds of import export controls rarely work. And that’s because unless the world wants to cooperate, and I mean the whole world wants to cooperate, there’s always the two and three and four steps.
It gets sent first to Italy and then Italy sends it to Macau and the Macau says, “Well, we don’t have any rules, so we’ll ship it to China, we don’t care.” And by the way, nothing with Italy, I’m just randomly using that as an example. But my point is that we know that the next big frontier, the next war is going to be fought on the technology front. And while yes, there are certainly ground wars being fought right now with weapons, we know that also if you’ve watched the Israel’s rocket, I don’t know what they call it, Pat, but basically the shield that they used, this is heavily technology driven, the ability to build these defense systems and it’s machine learning, it’s training, it’s intelligence and the advanced semiconductors are going to be key to that particular front. And of course there’s a technology leadership battle being fought on a global basis right now, right?
The country with the strongest AI is going to have the strongest economy, is going to have the strongest growth in productivity. It’s going to have the world’s admiration. And so we’ve been at a long time philosophical arms race around technology with China and us sending our most advanced technology and intellectual property over there for them to build and iterate and innovate upon is a risk factor for us too. So I’m not talking a lot about the specifics here, Pat, but one thing I will mention is basically companies Intel, NVIDIA, AMD, were all looking at building these kind of tuned down AI chips, the most notable, the 800 series from NVIDIA. And they were basically designed to get around the chip controls. The chip controls had certain specs. They said, “Well, we’ll build something right up to the edge of the spec and then we’ll make it work.”
And then what basically this new regulation included was, “Well, we’re going to just keep changing the goalpost, then you’re going to mail something that’s right up to the edge of the goalpost, we’ll move the goalposts.” Because what we don’t want to do is send the technology over to China that they’re going to get one way or another and make it easier for them to get and make it easier for them to compete both from a national defense and from a technology leadership standpoint.
And so I just think, Pat, this is going to be a merry-go-round. It’s going to go round and round and round. We’ll move the goalposts, they’ll continue to get it through gray markets and different channels. We’ll continue to try to make it harder through different import export controls. But in the end, China is fighting to keep its market and economic leadership as well as its defense leadership. It’s not leading the US, but it is the second-biggest economy or the biggest, depending on which metrics you’re using for imports and exports and they don’t want to be left behind. And the US has a lot of say in how fast China can move in AI. So here we go again. It keeps rolling.
Patrick Moorhead: Yeah, listen, I don’t want enemies of the United States or potential threats to get any technology that could harm my family and I’m also a realist that says that you actually have to be able to stop something. I remember when I was at AMD I was in my first year there and we had put a blockade on Iran and some weird picture showed up on USA today and in the background was a rack of AMD processors that they weren’t even supposed to have. I got firsthand account of what you can keep from coming over the border and what you can’t. So how big is an H100? Dan? How big do you think it is?
Oh, you’re on mute, but that’s okay. No, I’m going to get to the punchline here. It is about the size of a carton of cigarettes. It’s 11 inches long and about five inches tall. And I mean, think about how many ways you could bring that into a country. You could put thousands of them on a boat and show up on the shore. You could go through airport security and heck, you could relabel it. I mean it’s kind of funny for me to think that you can stop this and there is no phone home capability either. So you could do some sort of a lock where you had to phone home to NVIDIA to get authorization to make this thing work. All you need is drivers, they’re sitting right on the NVIDIA website and before you go to the next step that says know your customer, we’ve been through that.
There’s another article I wrote you can find on my website that talks about the insanity of know your customer type of things and how that doesn’t stop anything. So the other element is that although we really don’t talk about them, there are a lot of local powerhouses that are chip makers. One of them is called Bahrain and no, it doesn’t have the performance of an H100. It’s close to an H100 though, right? And what chips do you think that all of these models like OpenAI, GPT-4 were built on? They’re built on an H100, not an H100. So I think that this is a nice political gesture. I think in the end, A100s will get through H100s will get through border control at AB.
I think this is going to incent China to build its own ecosystem. And will they be as efficient or performant as an H100? No, they won’t because they can’t fab them on leading edge. But I think as we saw, they can do some miracles with lagging edge like they did with the Huawei chip. And that’s it. That’s all. I got. Insanity. Read my two articles, tell me.
Daniel Newman: Yeah, you hit it pretty well there Pat. And it is insane and it is going to continue because this isn’t something they’re just going to relent and be like, oh, we can’t ship it. I guess we’ll give up. We’re out of the AI arms race. They’re going to find a way. And so it’s going to be a bit of a cat and mouse game that we’ll continue. All right Pat, we’ve got a little bit more time and you’ve got only a little bit time. You got that race pass on? Are you going to the F1 this week?
Patrick Moorhead: I think we talked about that, Dan. I think you said it. Are you sleeping?
Daniel Newman: Yeah. No, but I mean is everybody watching the whole show? Are we that awesome? Does everybody, I just needed to bump up a reminder that Pat is, anyway, all right, we’re having fun. We’re smiling, we’re laughing. I’m laughing alone today. That’s fine. I’ll laugh alone. Whatever. All right, we got our eyes on some pretty cool AI technology yesterday. Pat, what were we doing with AWS yesterday?
Patrick Moorhead: Yeah, so we got the grand tour. We talked to developers, we talked to architects about the entire AWS custom silicon portfolio, two key locations for this development work. One of them is in Austin, Texas, which we visited and the other is in Cupertino. And I literally got every single one of my questions answered about Graviton, about Trainium, about Inferentia. And I got to take pictures. We got to take pictures of boards of the backs of chips that will tell you almost everything out there. Some really cool stuff that, well first of all, the Inferentia2 and Trainium1 are the same chip. Now the blocks are programmed differently for different things versus training versus inference and it’s packaged a little bit differently on the board, but it’s the same chip and I think that’s really cool. Ironically, when it comes to Generative AI, even with NVIDIA, their H100 does both inference and training.
I thought that was pretty cool. And gosh, one big takeaway was they designed for 10 years. They designed connectors, they designed their chips to last 10 years. And I thought that was just a huge takeaway for me. They showed us racks too, which was crazy. And the most impressive to me was this foundational model training beast. It’s two 4U racks stacked on top of each other with a 3U that as the CPU module interconnected by copper cables over PCI express. I would’ve thought for sure it would’ve been optical. That’s how you connect high bandwidth stuff. But they worked the signaling to the point where they almost hit the performance of, well not the full bandwidth of optical, but at just an incredibly lower cost with the same amount of reliability and having nearly the speed of NVLink. And I just thought that was just a big thing.
The other fun fact here is they actually do system wafer testing. So that’s taken a 12-inch wafer putting it on a fixture at a foundry in Taiwan that you might know. And it runs neural nets on it. And I’ve never heard of that in my life. I’ve heard some basic testing, right? Making sure power, making sure your pins are lighting up, but they actually do functional testing. The reason they do that is Interposers and HBM and the packaging costs so much, why test on a full processor? Now that makes sense, but the fact they do it on a complete 12-inch wafer blew my mind. And by the way, it’s not punching out the chips and testing those. It’s actually putting the entire wafer on a fixture and doing a functional test.
Got the Inferentia2 tray, pretty cool. It’s 4U. You don’t rack and stack because the CPU’s on there, boy, that looked like an AMD CPU with 12 Inferentia cards going over PCI express. And again, I can’t believe they let us take pictures of the boards themselves. We saw Lattice, some big Lattice semiconductors, probably security potentially an FPGA for IO. And then I saw a bunch of Marvell, which is likely networking.
So again, unprecedented access. I’ve been covering these guys for a decade and that’s how long the company has been doing this. But big picture, I’m going to get out of my geekdom real quick is AWS has found a way to have their cake and eat too. They have merchant silicon, everybody wants to work with them. NVIDIA, AMD, intel for CPU, for GPU for AI because they are the largest IAS provider hands down. So you want to go strong with AWS.
And then they have their own custom silicon CPU inference, GPU and where it all started, which was on the networking plane, which is ironic. So they can provide customers lowest cost, they can provide the customers highest performance. And I’m just scratching my head wondering how do the other providers compete on this level? You can’t just dive in and be world-class. It took Apple 10 years and it took AWS 10 years. And the question I have is how long will it take everybody else to get to this level if the rate they even want to go there, because AWS has to be spending a billion dollars a year on development. It has to be right. Even if I go 250 million a design per year that they have to rev, they have to be spending this.
So I’m going to end it there. I’m really excited about getting this access. Never thought we would get it. Nobody lets you take pictures of boards, really. But they did. And they don’t even let you take pictures of their racks because their racks are non-standard from power and a width and a height basis. So again, okay, I’m pretty excited. I’m going to calm down. Dan, you got the ball?
Daniel Newman: Yeah, you did. You spilled my beans. You shared all my secrets. Yeah, I had a couple of people when you and I asked, can we share these? And they’re like, yeah, go ahead. And I was like, sure. Then I tweeted and some people were like you said, I actually had 15,000 people look at that tweet and they’re like, “I can’t believe they let you share that.” People are commenting on that. And it was really cool and I guess their point is like, well, it’s not like other people can’t do this and take these out of a rack and take a picture of it potentially, but generally speaking it’s not a known practice to give that away. Pat, my take on it though is kind of like they’re like, catch us if you can catch us, if you can. There’s a quote, I believe it was Herb Kelleher, the CEO of Southwest Airlines early on. He said, “I could leave my entire strategic playbook on the seat of one of my planes and I’m not worried about anybody. I don’t believe anybody else can execute like we execute.”
And I think that was a little bit of what we were hearing was they absolutely believe in their execution. They believe in their ability to deliver price performance. And your comment about the get it all here, merchant silicon all the way to custom homegrown silicon, they’re like, we’ve got it all. I mean, look, this company, it’s not a small margin by which they’re the world’s largest infrastructure player. It’s not small as much as we want it to say it is, it’s actually pretty substantial, especially because the way numbers are sort of represented by all the different cloud companies when it comes to pure enterprises running their infrastructure on AWS, it’s a fairly significant gulf. Having said that, of course everybody’s competing. But Pat the theme vertical integration, the theme, we understand that if we can be more and more vertically integrated, we can be higher growth, we can deliver better experiences across the application stack.
And of course there’s a model for this. The company is called Apple and it’s a different business, but it’s the same of understanding how as you vertically integrate, you control your destiny more. But they’re approaching it in all ways and I think they’re doing a good job from a balance standpoint. You hit a lot of the geekiest stuff here, Pat, but from a story and an evolution here, Pat, they’re taking things like PCIE and using it to create competitive networking to what I’m not going to say it doesn’t outperform NVLink, but the point they’re kind of saying is, yeah, but for 1/10th the cost we can give a whole lot of performance and things like that are pretty impressive as we know that the scale of AI eventually is going to need to address price markets from lower to upper. By the way, your background is so cool. I’m just telling you that the building behind you.
Patrick Moorhead: Coming up folks.
Daniel Newman: Yeah. God, that’s so cool. Anyways, all right, I’m going to let that go there because I know you need to get on the road and we’ve still got one more topic and gosh, we just cannot do this show in under an hour anymore. We have so much to say. It sounds like a Dave Matthews Band song. All right, let’s wrap this up. We are heading into earning season. SAP was one of the first. You had TSMC this week actually outperformed, but SAP, another company that you and I tracked very closely announced its earnings and it was a bit of a mixed bag. It didn’t actually meet the street’s expectation, but the company reaffirmed its longer term guidance and got a really positive bump from it. This quarter it saw its cloud business grow 16%, the street wanted 18%. It saw its backlog of cloud grow.
It saw its margin grow, it saw its revenue grow, it saw its operating profits grow and its EPS grew substantially. So winner or loser. Well of course what you do doesn’t always matter when the street wants more, but my take is this. This is a company that’s quarter after quarter doing things right. It’s got a very difficult situation with lots of on-prem customers with massive custom code bases that are in a position now where they need to move from old iterations of SAP to S4 in the cloud. So they can take advantage of the most advanced AI and other cloud enabled features. It’s a tight rope. If they push too hard, they’re going to be the bad guy. And because of the complexities of lifting and shifting from old to new with SAP, they’re going to create vulnerabilities.
At the same time the company saw some massive customers this quarter make commitments and move over to SAP, which it’s always fun to watch the tit for tat that goes on between Oracle and SAP every quarter when they each announced the wins that they took from each other’s companies. But I think SAP is taking a very balanced customer led approach. They’re basically trying to listen to the customer base and make transition with the customer. But at the same time, with things like Joule and what they’re doing with advanced AI features being in the cloud, they’re subtly pushing the envelope of saying, if you want to get the best of SAP, you’re going to have to be thinking about migration to the newest technology. It’s a solid quarter. I mean, this is the playbook that Christian Klein, CEO, is trying to execute.
I had some time this quarter to speak with him about it and what I’m seeing in the numbers tends to match with what he indicated in the interactions that we’ve been having. Strategically speaking, I think it’s a tough but achievable road ahead for the company. The company needs to be executing, needs to be making its customers feel that there’s enough value in the migration and in the up leveling of their SAP stack. At the same time, they don’t want to give customers reason to feel they’re being bullied or pushed after making massive investments for multiple decades to build SAP instances that run their businesses. And by the way, taking down your ERP system or changing your ERP system is a massive commitment. Anyone that’s ever done one knows SAP has to really walk this tight rope. But I think according to the numbers, according to the reaffirm guidance, it looks like that’s the direction it’s headed.
Patrick Moorhead: Yeah, SAP is a company that you run your business off of, right? I mean, you’re building things, you have a transportation company, you have an airline, you have a power company. I mean, this is technology that you don’t want to mess with and it’s double-edged sword, right? It’s strategic and it’s sticky, but you don’t want to touch it. Okay? I believe that SAP has been too gracious with some of its customers and quite frankly it’s holding them back.
I think what we learned in areas like Windows and what we learned with Oracle Fusion was that at some point you have to cut and run and you have to move even though the Register might write nasty articles about you. Because what happens is you end up splitting your R&D and I urge any SAP customers out here to at a minimum get on the Common Core.
You can get on the Common Core three different ways. It’s not like SAP is saying there’s only one way to get there. There’s three ways to get there. But if you’re rolling and running on something that you were using in 1996, and can’t even remember the name, you need to move or you need to get onto something else or you need to write it yourself because it’s holding back SAP.
And I think this momentary thing with the quarter for a company like SAP to me is meaningless. I mean, I’m not an equities analyst and quite frankly, I don’t even care what the stock does. But what I do care about is that it is investing R&D and it is taking care of its future forward customers that aren’t holding it back.
So I’d like to see a little bit more push the company even has programs like RISE and GROW to help customers get there. So I think we’re, I don’t know, a couple years away from a cut and run by the way. And there’s no other vendor who will just let you sit and squat on-prem who is worth their salt because there’s no money in it. So all in all, I’m really happy with what the company has done with the cloud and to get all the good AI goodies you have to get on the Common Core and get in the cloud. And we’ve gotten multiple ways of getting there.
Daniel Newman: Yeah, you hit it buddy. Even a little bit more assertive than I am. It’s a really tough balance. You don’t want to incent anyone to leave, but you need to incent them-
Patrick Moorhead: Where are they going to go? Where are they going to go? Who squats? I mean what like an IFS?
Daniel Newman: What I mean is it’s pretty well understood that if you, our three instance is S4, that lift is the same as moving to Oracle.
Patrick Moorhead: Yeah.
Daniel Newman: So what I’m saying is when they have to lift off the prem and their custom code, they’re going to have to basically start from scratch, which opens the door. And that’s all I’m saying is that that threat can hold us back. You know how many times we think we don’t want to lose that customer. And you make a decision based on that, even though you’re holding back from saying or doing the right thing. We always are balancing that. It’s a tough one, but I think the company’s handling it well.
Pat, you have to go. So I’m going to end this podcast right now, but we covered a lot of ground this week. It was a great show. We talked some RISC-V, Qualcomm, Google, Lenovo analyst event, Oracle’s NetSuite event. NVIDIA, are their chips going to make their way to China? We think so. AWS and the Annapurna lab tour and SAP’s quarterly results. Pat, great show bright and early, but I’m pumped now. I will see you at the track. It’s qualifying day, by the way. It’s a sprint race this week. So we have three awesome days.
Patrick Moorhead: Hey, a very important question before we close, Dan.
Daniel Newman: Yeah.
Patrick Moorhead: Are you doing Schumacher or Barrichello?
Daniel Newman: Don’t make me talk about that. I don’t think I’m going to talk about that. But what I’m going to do-
Patrick Moorhead: It’s race day! Are you kidding me?
Daniel Newman: I’m going to represent from head to toe. And who are you cheering for? Pat, who are you cheering for?
Patrick Moorhead: Listen, I’m a Red Bull fan before they started winning.
Daniel Newman: Front Runner. Pat’s a front-runner. Just want everyone to want to know that because he is number one. But I am number two apparently. And I’m voting for Lando who has been number two. It’s just in my veins. Call me number two. Call me the five. You’re the six, but whatever. All right, everybody hit that subscribe button. We appreciate you all being part of our community root for Lando, not for Max.