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:
- MWC Wrap-Up
- Salesforce Agentforce 2dx
- Broadcom Q1FY25 Earnings
- Marvell Q4FY25 Earnings
- HPE Q1FY25 Earnings
- Tariffs Impact on Tech
- TSMC and Intel
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.
Transcript:
Daniel Newman: Hey, everybody. It’s Friday, and welcome back to another episode of The Six Five Podcast, episode 252. Daniel Newman and Patrick Moorhead here.
It’s been a big week, Pat. We had a long travel day yesterday. How you doing, buddy?
Patrick Moorhead: I’m doing surprisingly well and I have no idea why. I came back from Davos and I was just an absolute train wreck. A little disappointed on the gym situation. Gyms that open up at 7:00 or they have no equipment in them. Right now, I’ve got a bum shoulder, so I have no idea why I’m feeling really well. I did not sleep on the plane either, on the way back.
Daniel Newman: Yeah, there’s a rumor that I sat next to you just by happenstance. The happenstance, for everybody out there, was I happened to have my EA talk to Pat’s EA, and we happened to schedule to be on the same flight, and then we happened to put our seats next to each other. But it was a total accident by happenstance.
Patrick Moorhead: Yeah.
Daniel Newman: Yeah, it was really nice. I did sleep, not as much as you. I slept about two hours early in the flight, and then maybe dozed off for another 30, but it was 10 hours.
Patrick Moorhead: Yeah.
Daniel Newman: It was a long flight. American Airlines, fix your damn wifi. You guys are awful. $35 for five connected minutes across the whole flight. I spent more time trying to get connected than actually connected.
Patrick Moorhead: Yeah, it was pathetic.
Daniel Newman: You guys suck!
Patrick Moorhead: Yeah.
Daniel Newman: I just want to point that out, Pat.
Patrick Moorhead: I know. Well, I didn’t get anything done. You can’t get anything done. I wanted to grind for 10 hours and I couldn’t. Yeah.
Daniel Newman: It’s amazing how dependent we are to be online to get work done now.
Patrick Moorhead: Yeah.
Daniel Newman: There was an era, that wasn’t even a thing. You would have stuff to work on. You’d get on the plane and you’d work on your stuff, and you’d be disconnected. Then you’d reconnect. There was obviously a point before that where you had your chicken scratch. You were there for that. I don’t think I ever worked in that era. Anyways.
Patrick Moorhead: Well, in that era, what happened is you had Outlook and everything was synced. You had all your files that were on your PC. By the way, if there’s not a better proof point for on-device AI, I think it would be great to have DeepSeek distilled on my system that I could have downloaded all the documents beforehand for Salesforce, Broadcom, Marvell, HPE, everything.
Daniel Newman: Yeah, it’s great. But hey, man, we had a busy week. We had a lot of fun. We covered a ton of ground this week at Mobile World Congress in Barcelona, Spain. The Six Five was there, we had a lot of interviews. Check out and definitely follow our Six Five account and see them. We had good conversations, Pat and I together. We had good conversations independently. Pat had a banger of a sit down with Intel’s Product CEO Michelle Johnston Holthaus. We also had a founder of Palo Alto Networks. Then we had a bunch of off-the-record type conversations with CEOs and executives, and we’ll get into that more in a minute. But we also have, there was big news this week from Salesforce. We’re going to talk a little bit about their 2DX launch. We’re going to talk a bunch of earnings. Then we’re going to have a little fun and we’re going to go a little off the script here. We’re going to do some tariff talk. We really cannot separate what is going on right now in DC and what is going on in tech, so Pat and I are going to try to end the show today by breaking that down for everybody a little bit. Then Pat’s probably going to go on 10 TV networks and talk about it today. Right, buddy? You doing?
Patrick Moorhead: Only one. Only one. Yeah, I’m headed onto Power Lunch. Hey, bestie, we’ve been on for five years, episode 252. That is crazy.
Daniel Newman: Five years, we’ve been through it all. Market ups, market down, tech innovation. Tech deceleration. When we started, chips weren’t cool. Now chips are super cool. We’ve gone through it all. You know what, we appreciate every one of you that have been part of our community throughout this whole time. All right. Let’s tee this baby up, Pat. You wrote a great post yesterday. You were cranking, pushing stuff out. You did your whole MWC tear down. Why don’t you give everybody the quick what did Pat think, see, and feel this year at Mobile World Congress?
Patrick Moorhead: Yeah, it is funny. I put out what I thought was a very thoughtful banger and get 4000 views on content. I put out trash, what I ate, how many avocados I scarfed down, and it gets 25,000. But anyways. Here were my takeaways, and bear with me. This came, again, from the distilled conversations on-the-record and off-the-record with folks. Definitely a heightened sense of optimism. This industry has been through its challenges. If you look at the valuations of carriers and carrier equipment makers, it’s absolutely in the toilet. There were some people who took advantage of 5G and some who didn’t. But a real sense of optimism around this AI at the edge, the generative AI in the physical world. If it sounds like we’ve been here before, we have. We called it Industry 4.0, and it is basically industrial IOT. But there was a belief, and I do believe that there was a higher likelihood that the industry will succeed on this than it did in the prior generation.
Also, kind of a pivot it seemed related to 5G. A lot of talk there was in consumer services before. I think we’re still waiting for that next killer 5G consumer app that never came. If you remember the thesis from 3G to 4G, or sorry 4G to 5G was, hey, look at all the stuff that 4G brought, like Instagram and Spotify. We’ll have to wait for the developers to jump on that. As we saw, the challenge was you needed to have some pretty hardcore APIs plugged into every carrier to make it work. And developers weren’t too excited about the idea, if you were a global company, writing to 25 different APIs. Of course, there’s a lot of companies that have come in and a lot of industry organizations, like GSMA, for these APIs, but what we’re seeing now is really a pivot to the enterprise, which I just thought was fascinating.
You’ve got carriers moving into enterprise. You’ve got enterprise networking folks moving into the enterprise. You’ve got carrier equipment providers moving into more enterprise networking. But listen, we’ve got the SA Core, and I think it’s time to cook. We’ve got the APIs. I think compared to two years ago where we were talking about hypothetical use cases that an SA Core and stuff like bit slicing could provide, we actually saw them. Real, live enterprises using 5G advanced services through an SA Core to get stuff done. Now it’s time to scale. Final two things. To make a long story longer, a lot of talk about GPUs at the edge, or accelerators at the edge. That can be an Asic GPU or CPU with an accelerator to do some of this heavy lifting. Have we had GPUs on the edge, accelerators on the edge? Absolutely. Could they run generative AI inference? Absolutely not. I will leave it there.
Daniel Newman: Yeah. It was a definite pivot in sentiment, the overall consensus from the executives we talked to across carrier service providers, enterprise infrastructure companies, and then of course device makers. It felt a whole lot more positive. We had some really good conversations. Like I said, a number of them were off-the-record. There is a sense that the edge and the symbolic relationship that it needs to have with the cloud is returning to prominence once again. There was this period of time where, I call it the edge in terms of the telco edge, was kind of being looked at as a bit of a dumb pipe, and everything was thought to be done in the cloud. We saw that DeepSeek moment where, again, there was a lot of misinformation in the DeepSeek moment, but one of the things it did really start to emphasize is what can be done on a device, what can be done on a PC, what can be done closer to the edge. What could potentially be done branch versus AI all being built and delivered out of the data center. Which makes sense, especially on the training side, but it doesn’t really make sense when you start to scale out inference. If you heard Jensen’s CES keynote where he talked about physical AI.
Patrick Moorhead: Yeah, absolutely.
Daniel Newman: Physical AI is going to be a continuum of really bringing smarts outside of the data center. This is really I think what Cristiano Amon had meant when he talked about the connected intelligent edge. I think what they’re doing with, what is it, Dragonwing, Pat, is really interesting in particular. We talked about this on the show last week. Really, the idea, and this really came to my head as I was having conversations, is this continuum of data that needs to be captured and processed at the edge that’s going to be important if we’re ever going to really have true autonomy in things like vehicles, or in really the intelligent cities that, as you said, Pat, in Industry 4.0 1.0 that we heard a lot about that we didn’t really get. Is going to come down to figuring out … It has to be done at lower costs.
This is why I think companies like Qualcomm are more interesting because Nvidia’s stuff is all super high-end processing. The humanoid robots, autonomous vehicles, sure. But what about the data coming off of an oil field, or a farm, or a smart city, or a manufacturing plant? Then of course, this is where the Ericssons and Nokias, and what they’re doing with their enterprise pivot makes sense. Is they’re in these places, they’re providing fixed wireless access, they’re giving 5G. That dumb pipe becomes a really important, meaningful connectivity solution for the edge. And really, despite 5G’s long in the tooth, it really hasn’t been long in the revenue tooth. We’re really just starting to see that happen now. That’s my take. It was a good time out at MWC.
Pat, let’s hit the next topic. We did sit down with Jeff Amann, the head of industries and telco for Salesforce at MWC. We had a great convo. If you didn’t check it out, go check it out on our Six Five. But Six Five was … Oh, sorry. Six Five. Salesforce was busy making a bunch of announcements this week. One of the big one is … Let’s go back a bit. This is a company in a massive transition. We talked about their earnings recently. We talked about growth had really slowed in core business. But where growth hadn’t slowed is in its Data Cloud and in its Agentforce. You’re talking thousands of people, of companies that are signing up to use this. You’re talking about high, high growth rates. Close to triple-digit, around triple-digit growth rates of their Data Cloud. It might even be higher, I actually don’t have it memorized, but it’s really high growth rates. This is all about that data alignment to be able to deliver on agentic. This is a company that realizes it’s going to need to pivot. It realizes that the future is going to be generative extractions. It’s going to be coordination of agents and orchestration. It’s not going to static, rigid, difficult to use SaaS software. But like every business, it needs to make its transition. We’ve done a lot of work on this. We did a big study looking into the economic impacts that could be saved through agentic AI. We’ve identified that it could be in the range of $6 trillion in the next 12 months or so.
This is because companies can put a lot of what I call the middle, Pat, the middle of the org. Things that are heavy, slow, gelatinous parts of an organization that don’t necessarily have critical value. I know that’s always really hard to say about human capital, but in many ways, we got a lot of … I heard this little bit about people moving bits of data around an organization. How many people take a little bit and move it to another person, and move it to another person? This kind of role is what an agent will be perfect at. Salesforce basically, with 2DX, they’ve got these proactive AI agents, that’s what 2DX enables. They’re working autonomously on business processes. Meaning you need something done, this agent can do it. There’s a lot of how good is it that’s still running through people’s head. You and I had that whole it ain’t going to replace salespeople.
Patrick Moorhead: Yeah.
Daniel Newman: That’s for sure. Actually, Salesforce has admitted that by hiring 2000 salespeople to sell agents. But I do think when it’s needing to … A lot of things we do, Pat. We process a video and then we want to create a tweet about it. Could that be done largely autonomously? Yeah, I think it could. Or you get a new sale, could that invoice and that follow-up interaction, can those things be autonomously programmed? Absolutely. They’re also really focused on this DIY versus Salesforce, meaning could you do it yourself or using Salesforce. They’re continuing to really focus on giving low code and pro code optionality to people so that developers can build more complex agents, but also companies don’t need to have a big developer group to get this done.
The App Exchange is the marketplace. Marketplaces are always going to be super important. Not all agents are going to be built on Salesforce. Salesforce has a big partner ecosystem. Wanting to plug in agents that can orchestrate across the org is going to be something of focus. There’s just some other stuff around developers and ecosystem. But I think right now, Pat, my feeling is we are getting closer and closer to this agentic moment. It’s not going to happen all in 2025, but companies are really meaningfully looking at where this can drive efficiencies into the business. We see it, multi-trillion dollars. I think Salesforce has been really aggressive to make this pivot and I think they have to be.
Patrick Moorhead: Yeah. So, not only did they bring out information at TDX. Also, Morgan Stanley conference, they gave even a deeper dive. I think maybe stuff that would be even more interesting for you and I. For instance, they talked about 5000 deals in Q4, 50% resolution rates in pilot sectors. Cost savings of $7 to $9 per interaction, which I thought was pretty interesting. There is this thing called Universal Credit Pricing, which replaces the conversation models that I thought was interesting.
It’s called 2DX, but maybe it should be called Agentforce 2.5. In this world of compressed acceleration and everybody trying to one-up each other, here we go. I did find it interesting that, at the same time, Marc Benioff was putting out tweets about this. Microsoft’s Satya Nadella talked about how it was using its agents. Personally, every day, he’s using it as a CRM and he’s not using Salesforce, or even Dynamics 365. This spat between DIY and leaving the driving to Marc Benioff moves forward. From a targeting perspective, target audiences here were enterprises, developers, and admins, and partners. For enterprise, it’s all about automating workflows across customer service, HR, and supply chain. For developer and admins, really it’s leveraging their ecosystem. Then partners is obviously making money via this new agent exchange.
Last comment I’ll make. I really like the Salesforce product discipline, product management of, “Here’s the problem.” What problem are they trying to solve? Labor shortages. Operational efficiency. Integration complexity between low code and lots of code with some of these pre-built integrations. All in all, not a step function forward I think, which is why it’s not Agentforce 3.0 but 2DX, but definitely a step forward and should give Microsoft, ServiceNow, Oracle, Adept, and UiPath something to think about.
Daniel Newman: By the way, did you see … This was really interesting. Did you see with DOGE finding all these HUD software licenses?
Patrick Moorhead: Yeah, it was pretty wild. To be expected.
Daniel Newman: One of the big ones, I’m a huge fan of ServiceNow, but I think the number was, God, it was 35,855 licenses and they were using 84.
Patrick Moorhead: Right.
Daniel Newman: There had 11,000 Acrobat licenses and they were using zero.
Patrick Moorhead: Yeah.
Daniel Newman: Jeez, wow. Okay, sorry. Step tangent from the reality that we’re in on this show, but God. My God, dude! All right, let’s talk Broadcom, man. They absolutely smashed it. Had a huge reaction after the bell, but it’s tempered today, only up 3%.
Patrick Moorhead: Yeah. So, just by the numbers, EPS beat by 6%, revenue beat by 2%. But it was the guide and the future comments that got everybody super, super excited. Listen, revenue up 25%, non-GAAP BPS up 45%. AI up 77%. Infrastructure software, a lot of VMware, up 47%. Pretty awesome. But in the end, it was all about ASIX. These are the competitors to GPUs that are typically higher efficiency, lower cost, and a little harder to program. Particularly if you are locked into CUDA. A couple questions that analyst interactions was all about the confidence in the future growth. Hock rattle off a bunch of future wins that they haven’t baked into the financials or the forecast. What’s funny how little conversation there was about VMware. If you remember, the past couple quarters, half the call was dedicated to VMware. A little bit of talk about pull-ins. Hock said, “Hey, new material timing effects that’s going to impact Q1s or Q2.” I think net-net, it was all about confidence in future growth in their business. Which, by the way, was in sharp contrast to what we saw with NVIDIA, AMD, and we’re going to talk about Marvell.
Daniel Newman: Yeah. Thanks for leaving that out there. Look, everything was really tidy. This is a company that has its strength in its diversity. It has the ability to take and lead in a direction that’s future-forward. But at the same time, their real ethos is always about being able to support a product and a customer, and that might sound a bit in contrast to how the market reflects it at time. But in reality, they don’t tend to want to compete. You look at Apple moving away from basically everyone, but in actuality the partnership between Apple and Broadcom has deepened in a lot of ways, and that’s been announced.
You look right now what’s going on, Nvidia and Broadcom have a really important partnership. I really do think that ethernet will be the scale. You can call it an Apple-Android thing, but I think in the end, the AI data center scale will be built on ethernet, not on InfiniBand. There will be its applications for it, but that’s just how I see the future. That’ll create this really important symbiotic relationship that’s going to exist, even if they are GPUs. Now of course, there’s I think a lot of enthusiasm about the XPU business. It’s not really very big yet. That’s one of those things that people are … I don’t think the actual XPU part, I think the total AI business is four billion now. Four billion for the quarter. I think it’s bigger for the year, it’s 16. But I think 75% of that is still networking.
Patrick Moorhead: Yeah.
Daniel Newman: I think the vast majority of that number is not the XPU. A lot of people, when people like you and me, start point out that the XPU may or may not be a non-factor, it may have a real competitive posture at some point to NVIDIA, people want to throw cold water on us about it. I just keep saying it’s not a zero-sum game, but it’s going to be both. First of all, we know that accelerators are going to show up in a big way at the edge. We know accelerators are going to show up in a big way for hyperscalers that want to control their own destiny. They’re going to want to build their own chips, build their own network infrastructure. We were already seeing this happening. This doesn’t mean that they’re going to just not use NVIDIA, though. That’s the thing is everybody wants it to be one or the other. That’s not how it’s going to work. It’s going to end up being both and there’s going to be a really big market for it, and I see that happening over the next few years.
That’s something that the investor in Broadcom should be really encouraged by is this is a net-new revenue stream. They’re building for multiple hyperscalers with more on the horizon, multi-generation build outs. This is something they talked about. And expect every generation that these commits, they’re going to get a little bit bigger. Then you look at what they’ve talked publicly about co-packaged optics and memory, custom HBM. Now we know Marvell has led first, or at least came out first, but you could actually be sure they’re going to come out and they’re going to build that as well. Then you start adding it all together.
Then you take mainframe software, VMware, the core chip business that they’ve had for a long time. Everything PCIE, network route switch, back of the rack, front of the rack, all that stuff. That’s really stable business, and most of the time they tend to have, like with SerDes, where they have IP control. They cannot be actually encumbered by competition in that particular area, so they end up winning every design. Anyways, a little bit like what Qualcomm does on certain parts of the phone. It’s a really great business, great results as expected. Market is just in shambles. When we talk tariffs, we’ll talk about it a little bit more, about why it’s in shambles and these things aren’t up more. It was still a good freaking quarter. God, people. Buy the [inaudible 00:24:59].
Patrick Moorhead: It’s interesting how the market has wavered. It was up big time and it’s up today 6%. Maybe that-
Daniel Newman: Is it six now? It was three a little while ago.
Patrick Moorhead: Yeah.
Daniel Newman: I think it was three when we started talking.
Patrick Moorhead: Okay.
Daniel Newman: Hey, Hock, Charlie, take note. It was three when we started, six when we finished.
Patrick Moorhead: There we go.
Daniel Newman: Anyhow. Let’s move onto Marvell. You and I have both talked a lot about Marvell as the alternative. Marvell had a strong quarter. It had a triple beat. I’m going to go really quick and get to the end here. Sometimes it takes me a while and I like to rattle on about things. This was a whisper number miss.
Patrick Moorhead: Yeah.
Daniel Newman: The miss was a whisper number miss. They came up above the consensus guide with a beat. But the delta between that beat and the whisper number was substantial. The stock fell 15%. It’s closed down almost 20 the next day. They had 62% year-over-year growth in their whisper number of 1.875, but the whisper on the street was 2.1. It was an almost 10, 10-and-a-half percent miss on the whisper. This is basically people are saying that, yes, you’ve gotten into AI. By the way, look at their business. AI data center. Whew! Every other part of that business. They have completely made the pivot. They have not been shy about this. They will not be insulted if you say this. But their business pivoted to this, so now they’ve retrained the whole market to look at their whole company as an AI data center play in many ways, so people wanted to see that number growing more quickly. This is that push-and-pull about how fast AI infrastructure is growing, how fast are XPUs growing, how fast are non-InfiniBand, non-NVIDIA solutions in-and-out of the rack growing. Marvell is supposed to be the other biggest winner here. We know about AWS and Trainium. We know they had some other builds that are out there. What kind of scale, what kind of growth, what kind of utilization? And to some extent, I think the difference, the biggest polar difference between Broadcom-Marvell is Broadcom gave this visionary number and guide like, “This is just getting started.” That was the feeling. “Oh, this is just getting started.”
Patrick Moorhead: Yeah. They also hinted about these two other design wins that aren’t even baked in.
Daniel Newman: Totally.
Patrick Moorhead: To the financials. That was a difference, too.
Daniel Newman: Totally.
Patrick Moorhead: Apple and ByteDance, right?
Daniel Newman: Correct. We know ByteDance might not make it out the other end unfortunately, due to China-US ongoing flibber-flabber. We’ll talk about that a little bit more later. But we know Apple’s out there. The thing is we know that every generation to generation, these designs go back up for bid. I just think there’s this sense in the market that Broadcom is winning the bigger multi-generation, and they have the proof point with the TPU. Marvell has to prove that it can win the multi-generation, win the volume. That once they do win, come out in volume. By the way, it’s not just XPUs. They’re getting some data offload parts, smart networking parts that are related to AI as well, but the volume and the numbers are just not as exciting to people. It’s just not as much at the trendline. Pat, I think the opposite reaction wasn’t in this quarter. It was a miss on the whisper and it was a miss on providing that market confidence that these XPU projects are going to scale. Same fate AMD faced, by the way. Just not giving that certainty of what’s ahead.
Patrick Moorhead: Yeah.
Daniel Newman: The market right now hates uncertainty. I’m going to say that a few more times before this episode’s over.
Patrick Moorhead: Yeah. Yeah, I’d like to say they got AMD’d, but none of it makes sense. All-time record, Q4. Record Q1 guide. Great execution. 20% sequential revenue, Q3 and Q4. I think investors just thought there was something else in Q1 that a company like Broadcom delivered. There’s still, when you look at growth rates, they’re still the only company that’s grown more than them has been NVIDIA. It appears that they are on that 20%, on track to get that 20% share of that $75 billion data center that they outlined at their AI Day.
I always like looking at the sentiment of the analysts. Yes, I’m using Perplexity to help me with that. But it’s so funny, I don’t trust it, so I went in and I read the transcript. I think on a positive side, data center growth. Gee, how could that not be? On the negative side, she had Barkley’s questioning near-term Amazon ASIC constraints. Characterized high watermark expectations. Cantor Fitzgerald called the guidance conservative, suggesting that Marvell was in the penalty box until visibility improves. Vivek Arya highlighting 30% year-on-year decline in enterprise networking. Which again, I think came down to industry inventory digestion, but also that this enterprise market is still not exactly banging it out. All in all, not a great quarter and a great forecast, but the market, they just wanted more.
Daniel Newman: Yeah, I think that’s absolutely right. Unfortunately, it’s insatiable right now. Buckle up, pull up your pants, wear your suspenders. If you’re not up for some volatility right now, go to cash because it could be really ugly. But I always say that out of all that typical, these phases of ugliness, comes the best opportunities. Anyone that knows from buying the … Again, it’ll depend on the stimulus and the bets, there’s so many things at play here. All right, let’s move on, buddy, to HPE. This was another really interesting print.
Patrick Moorhead: Yeah, it was. They’re down 15% today, getting hammered out there. They had a slight beat on the revenue side, they missed on EPS. But I think the biggest thing is they were below expectations on the next quarter and for the year, not only on revenue but also on non-GAAP EPS. I think this just demonstrates the lumpiness of this AI market here. There was a lot of discussion, either cited or non-cited, on the Blackwell transition. It was an issue. That’s driving not only some revenue challenges, but also some profitability and inventory challenges out there. If I stand back and look at how they did, order growth, double-digits across multiple segments. Green Lake, 41,000 customers, $2 billion on subscription. A lot of positives here that I don’t think made it out here. I think a lot of what we’re seeing with HPE is a combination of expectations management, but also the challenges of working in this AI economy.
Daniel Newman: Yeah. Are you handing me the mic?
Patrick Moorhead: I’m handing you the mic, bestie.
Daniel Newman: Thanks, buddy. This is another one that’s just super guidance-driven. They beat on revenue, they barely missed on the bottom. Barely, like a penny. I get it. A miss is a miss is a miss. But it was very, very close. Great growth of server orders, up 29%. Again, people want that to be 275% right now. This is the is AI booming or is AI booming?
Patrick Moorhead: Yeah.
Daniel Newman: Kind of thing that everybody’s got going on. But they saw their AI systems orders doubled. They got 8.3 billion of AI systems on order, but the stock dropped 20%. Look, again, it wasn’t even a whisper number, this was just a straight guide miss. Right now, you can’t miss the guide. Literally, you cannot miss the guide. Nobody cares, unfortunately right now, what you’ve done. They want to know what you’re going to do next. They basically came out and said, “We’re not going to hit revenue. We’re going to miss by somewhere between 2 and 700 million next quarter. We’re not going to hit the EPS for either the full year or the quarter.” They were very uncertain about tariffs.
We talked a lot about this this week and we’re going to talk more about it in a minute. The non-certainty of where they can manufacture is creating some consternation and it will, it’s going to be friction for a while. The 2500 job cuts, actually I think the market will like that. As much as they don’t like a lot of things, right now the market likes efficiency. This has been a turnaround key. I don’t know if I would have actually announced it at earnings. I would have waited a minute, let the sell-off happen, and then come back and announce it because people seemed to like this. They shouldn’t, it’s not a good thing. We don’t want it. But at the same time, if the company can be more efficient, raise its margins, get closer to its EPS through cost reductions, that’s part of the role a company has to play. I expect them to do that.
The sell-through of the servers in inventory was a bit of a challenge. And of course, seeing Supermicro not capitulate, apparently getting through their little audit and then staying listed. I think there was a big opportunity had something gone the other way for Supermicro. I think amazingly, because like I said, I fall on my sword. I don’t understand why a big five accounting firm would have fired them if there was nothing going on. But there’s a lot of gray matter here, there’s not a lot of regulation and ability to regulate. I tend to believe the accountants and lawyers of these big corporates are smarter than the accountants and lawyers at the SEC. I think that there’s a lot of ways you can move things around to get things done.
Pat, in the end, it’s going to be maybe the reset here, bring it down a bit, beat next quarter and get back to growth. During this AI growth period though, if you’re a company that’s selling AI solutions, AI systems … Oh! One other thing, by the way. Great ARR growth. They’ve surpassed $2.1 billion of ARR, 45% growth. Green Lake is working. The ability to deliver as a service some of these AI solutions might be one of the biggest and most compelling differentiators that HPE is going to have. Dude, we are absolutely tearing it up today.
Patrick Moorhead: We are. We’re on point, man. You and I should be jet-lagged, death.
Daniel Newman: How did we do that?
Patrick Moorhead: It’s close to the end of the day in Spain.
Daniel Newman: We may actually be done not only on time, but early here. This should be a long topic.
Patrick Moorhead: We should turn this topic, we should talk about this for about 30 minutes so we’re late to our meeting.
Daniel Newman: There we go, there we go. Let’s talk tariffs. Now that you and I are both geopolitical, macroeconomic correspondents, but you just cannot separate what’s going on in the macro, what’s going on with the Trump Administration, what’s going on with all these global negotiations on tariffs from the tech trade. For people that are out there that are trying to figure out tech. This goes back to that quote that we got from one CEO at the event where it was basically, “We move manufacturing from Point A to Point B because of tariffs. Now we don’t know if Point B is going to have tariffs. We can have a mitigation, we can carry through this, we can hit numbers, we can grow, and we can deliver good results to the street, but we can’t do this if we don’t know what’s going on.”
I call the trade right now, it’s all an uncertainty trade. We are in this capacity right now where we just simply don’t know what’s going to happen next. One day, tariffs for Canada and Mexico were on. Next day, “Oh, we’re going to hold of for those for a month. Automakers, we’re tarrifing everybody starting in April. Oh, maybe we won’t do all of that.” We’re just seeing this going back and forth. Of course, we’re seeing this Ukraine situation, the DOGE situation. Are we cutting all these costs? Are they going to even let the government cut all these costs? There’s all these actions being taken within the courts to try to stop the freezing of some of these different aid payments. By the way, I listened to the whole Rogan Elon Musk episode last night. I got a whole great tear down, if you have any time or anyone out there wants to get a great rundown on NGOs, and all the waste and spending.
But look, these things are all symbiotic. In the end, what I think the tariffs are, I’m going to put this out there on the record, is I think the tariffs are one big negotiation to bring down interest rates and bring down inflation. That’s not really all that big of a bet, but Trump said he’s been pushing on Fed Chair Powell for some time to bring down interest rates. It’s hard to justify bringing down interest rates when you have stock markets near all-time highs. You have assets like Bitcoin at all-time highs. You have homes still trading at very high levels despite increasing inventory. But there’s also a whole bunch of bad macroeconomic stuff. The tariffs are interesting because, first of all, what most people don’t understand is that we are tariffed exponentially more than most of the tariffs we send out. Even to China, as we double our tariffs, we’re still paying a bigger tariff to China. When we send cars to Europe, we pay a bigger tariff. They don’t when they’re sending them to us. It’s weird because it feels like nobody, Pat, really knows this. Everybody’s like “Tariffs are terrible. I can’t believe we’re going to …” Everybody tariffs us. We pay tariffs to everybody. Why is this so crazy?
But part of it, when you think about China, you think about Korea, you think about some of the threats to Taiwan. The risk of Taiwan, if Taiwan decided to bring some implementation R&D here, I don’t think they’re going to bring their biggest R&D here, but they’re also planning to bring N2 here potentially. Trump’s reaction in his State of the Union report was if they do that, they won’t pay tariffs. There’s an incentive in that. He said cancel the CHIPS Act, we’ll come back that some other time. But the bottom line of it is is that the reciprocal tariffs are really interesting because what if the reciprocal tariff says, “Everybody’s going to be paying these 40% tariffs,” so what if everybody finally comes together and says, “Hey, it’d be really weird, but we could bring all the rates down.” Then suddenly you start bringing inflation down. Actually, what people aren’t talking about, have you seen the 10-year yield? Have you seen the 10-year note? Let me see really quickly what it’s at. 10-year yield at. The 10-year has dropped by half a point in the last … Not half a basis, it’s at 4.25. It was at almost 5%.
No one’s really talking about the fact that all these government spend cuts, all this tariff talk, we are seeing the number break that needs to break. The interest rate, the bond markets bet on what interests are going to be. If interest rates come down, guess what? The middle of America, not the asset holders, not the people that own Bitcoin, not the people that own stocks, but the middle of America can actually start to afford things. They can afford homes again. 8% interest rates, no one’s moving. People that have 2 and 3 percent, they aren’t going to sell their house. Cars, you buy a $50,000 car and you’re paying 10% interest. You’re paying $5000 a year of interest on a car. Over five or six years, you almost pay twice the price of a car. That’s really hard on the average American. We’re not even talking about bananas, and eggs, and all the things in the grocery store. We have to beat inflation.
I think what he’s doing is radical, it’s revolutionary. It could be really wrong. I’m not saying here it’s absolutely going to be right. But I’m saying that just this endless spigot of stupidity, and spending, and foreign aid, and sending money to places that we don’t have, running a debt that we can’t support and paying over $1 trillion a year of interest is not sustainable. I’m signing up for doing something different and trying something else. But this tariff thing impacts tech. It just simply does because, whether it’s moving semis, moving laptops, moving all goods, automotive and everything else, Pat, we cannot do it. My rant is over to you. You’re muted, dude.
Patrick Moorhead: Thanks. Sorry, my dog walker came and my dogs were barking. Anyways. I think your idea where all tariffs come down, I think that works with the exception when you have government subsidies. TSMC, very heavily subsidized. Tech companies in China, very heavily subsidized by the government. It works up until a point. The one thing though that bugs me is this idea. We gave up on US manufacturing back in the ’80s. We didn’t reskill anybody. What I can’t find is a thesis where it makes sense, for example, to do iPhones here in the US. Here’s a crazy, fun fact for you. China’s Foxconn imports Malaysians to work in the iPhone factories. There’s reports that say there’s up to a million people who work in these factories. Where in the United States have you ever heard of anything like that in the past 50 years? What would it take to rebuild that?
The other thing is supply chains have to come with it. The 500 companies, for instance, that you have to have in place. I’m just picking an arbitrary number, making it up. But a lot of different companies that require, particularly in the electronics supply chain, to build a product. You’ve got PCA, you’ve got PCB, you’ve got final assembly. Heck, you want to talk about plastic injection molding, and metal bending, and aluminum C&C. Okay. That takes years and years of investment, and you can’t just say, “Oh, I’m going to put a tariff on it.” What, maybe five, six years that can be established? You know me, I’m not one for nuances. I’m one for more black-and-white-
Daniel Newman: Certainties.
Patrick Moorhead: Yeah, yeah, more of that. I’d like to see the full thesis. I’d like to see the full game plan. Don’t get me wrong, man. I love the idea of our government getting more efficient. I love that … We grew our government over 20% during the pandemic, Daniel.
Daniel Newman: Yeah.
Patrick Moorhead: We just kept spending it when the pandemic went away. That is just absolutely disgusting to me. It’s your money, it’s our money, and that stuff just needs to change here. Yeah, those are my thoughts on it, bestie.
Daniel Newman: Maybe, just since we have a few minutes left and I’m sure a lot of our audience would love to hear about this, because this all really does tightly correspond with the Intel, TSMC, TSMC moving here, Apple moving here, less so but also similarly. Apple, for the 17th time, has committed the same $500 billion to the United States. But it’s a victory lap, people forget things very quickly. Let’s hope they do build here. Let’s hope they make phones here. Let’s hope they build more of their PCs here.
Patrick Moorhead: Yeah.
Daniel Newman: This is a good thing. This is the whole thing I want everybody to hear, at least from me, is that this is all potentially good stuff. Why is this good stuff? Well, first of all, getting TSMC to bring its most advanced node here, Pat, and I saw your tweet. Even if they are only bringing implementation R&D and not necessarily the full skillset of their next generation and newest IP, we at least secure the current level of technology for the next handful of years. Meaning we don’t go back into the stone ages, effectively, if for some reason to, the music stopped, and China invades Taiwan, and they’re airlifting all the systems out, and they’re blowing everything up. Again, fatalistic potentially, but we just don’t know.
But additionally, competition is a good thing. I get the feeling, I know Trump keeps saying, “Cancel the CHIPS Act,” and this leaves Intel in this weird middle. But I actually believe this is a bit of the deal thing that he likes to do. Look, I think Intel has to become the US chip making champion. I’m not talking about instead of TSMC being here and making lots of chips, or even Samsung coming around. By the way, I think the same thing for memory with Micron. Micron needs to be the US champion for high bandwidth memory, and of course just other NAND and traditional memory that we don’t talk about that often anymore.
But we need competition. It is a monopoly right now. We have one company controlling pricing that has capacity limitations. They’re controlling co-auths. They’re controlling leading edge two, three, four nanometer. They’re mostly sold out for multiple years. This is leaving very little capacity and it’s giving them endless control over pricing. We don’t want monopolies. Monopolies create price inflation. You want to get rid of price inflation? Don’t have monopolies. To some ways, maybe yes, cancel some pieces of the CHIPS Act. But I actually think fund Intel.
Patrick Moorhead: Yeah.
Daniel Newman: Bet big. And also, just know it’s going to take some years. Even if they get the situation right, the governance right, the wall between products and foundry right, they get the people hired in, maybe TSMC takes a small stake like you’ve suggested. Bring some of its talents to it to create extended capacity, competitive processes. They can make the AI parts in different foundries, different fabs, at different scales across the US. Then you can obviously take this to Europe, take it to other parts of the world, so that we’re not so dependent on one company in one location doing everything. This is a good thing. We should be egging on more investment here, more manufacturing here, more guarantees that we are securing the national future of this country and our companies. This country. And of course, maintaining economic leadership. Economic security is national security. We do not have economic security here, Pat, as long as we can’t make enough of the chips we need to support our businesses and commerce, and then of course our defense industry.
Patrick Moorhead: Yeah, a couple comments. Just to review. Foundries create two things. They create intellectual property around transistors, and then they make chips. Of course, they package it, but a lot of that packaging is outsourced. They all have to come together. A scenario would be that TSMC invests that 100 billion, builds the factories. Let’s say we’re on two, and then the follow-on. If China invades Taiwan, essentially, the shock clock time stands still. All we can build on TSMC is the intellectual property that’s currently out there and being produced. Today, C.C., the CEO and chairman of TSMC, basically put the true colors out there, I love it. He says a lot. I don’t know if he knows how it can rubber band or boomerang back. He essentially said, “Hey, don’t worry about this whole thing. We’ve got 10,000 of our best engineers still in Taiwan and we’ll have 1000 in the US.” Essentially, those 10,000 people will be IP creators and implementers, and then 1000 here in the United States will be implementers of that IP created in Taiwan. The Earth would probably stand still for two or three years, where nothing would move forward. Here is the other scenario that I’ve been thinking about. Call me crazy. By the way, tell me on social media if you think I’m nuts here.
Daniel Newman: You’re nuts.
Patrick Moorhead: Let me lay this out. A lot of the core machines from ASML and Applied Material, Tokyo Electron, a lot of the stuff that you see when you go through a foundry are the same machines. You could hypothetically run an Intel transistor inside of a TSMC factory or foundry here in the United States. And by the way, AMD, Apple, NVIDIA, Marvell, Broadcom would have to obviously spend a lot of money to build those designs and that flow for that Intel process. Maybe the world wouldn’t stand still for two or three years, but maybe one. Now if it gets really heated, the US government should, let’s just say the CHIPS Act goes away, pay the designers enough to put their designs to create a product that could go into Intel 18A and beyond. That would be a very nice hedge and insurance policy. It doesn’t cost $500 billion. Anyways, I was working out and was thinking about how that would work.
Daniel Newman: Nice.
Patrick Moorhead: Tell me on social media if you agree. Or if not, that’s okay, too.
Daniel Newman: I don’t agree with you on much, Pat, but I actually have been saying for a long time that a lot of this can be solved with policy. The fabulists may not like being told they have to produce a certain amount of their chips, but this is actually one of the true cases of real national security is making sure that survives. Like I said, I don’t know if that means some sort of additional levels of security of how IP is protected between products, business, and the foundry business. I’m not a huge fan of the break-up idea. I think there’s more value in the company staying together, but I do think there could be some very strong church and state that is set up to make sure that is absolutely protected. If there’s a huge amount of liability, I’m sure something can be done with the blockchain with IP or something else, to absolutely make sure it never gets into the wrong hands. But hey, there’s a lot here and a lot to be said, a lot of unknowns. Pat, have a wonderful day, buddy. Any last words before I shut this thing down?
Patrick Moorhead: No. I think it was a pretty good show. I don’t know how we went through there. We need to be jet-lagged more often. I think it was that leg day I did. Did I tell you I love legs now?
Daniel Newman: Yeah.
Patrick Moorhead: Now that I can’t do upper body.
Daniel Newman: I’m really not sure how to take this though, but legs are the absolute best and the worst. I love all those videos on TikTok that show me doing leg day, or how to do leg day, how to have fun doing leg day, and it shows somebody going over starting a Bulgarian squat on a bench.
Patrick Moorhead: Yeah.
Daniel Newman: Then putting on the weight and turning over, and they’re doing bench presses. I see those videos all over because legs are painful. For everybody out there, this is the Pat Moorhead fitness corner is legs are the largest muscles in your body. You build your legs up, it builds hyper-speed to your metabolism, it builds your strength. Of course, those compound lifts are great for your body and your core. Some of us, still chunkers, but I’m working my way down. Buddy, I’m going to be ripped like you pretty soon.
Patrick Moorhead: Hey, have you ever noticed Dan is cutting and his face is looking very thin, but he’s keeping the guns. He’s keeping the guns, man. Your midsection is getting tiny, too. You may or may not have shown me your six-pack.
Daniel Newman: I may or may not have one. We’ll have to do a great reveal. Everybody out there that loves our pod, I think it could destroy the network if we started doing shirtless gym pics. We’re going to go ahead and keep doing them with our shirt on. Unless you want to give us an overwhelming response that you really want that, and then maybe we could be convinced.
But hey, for everyone out there, thanks so much for tuning in. We very well may, if not this week, the next week, be moving to our new format. We’re very excited about it. It looks great so far. But for the 252 episodes of this show, from Patrick Moorhead and myself, thanks so much for being part of our community. Subscribe. See you all later. Bye-bye.
Author Information
Daniel is the CEO of The Futurum Group. Living his life at the intersection of people and technology, Daniel works with the world’s largest technology brands exploring Digital Transformation and how it is influencing the enterprise.
From the leading edge of AI to global technology policy, Daniel makes the connections between business, people and tech that are required for companies to benefit most from their technology investments. Daniel is a top 5 globally ranked industry analyst and his ideas are regularly cited or shared in television appearances by CNBC, Bloomberg, Wall Street Journal and hundreds of other sites around the world.
A 7x Best-Selling Author including his most recent book “Human/Machine.” Daniel is also a Forbes and MarketWatch (Dow Jones) contributor.
An MBA and Former Graduate Adjunct Faculty, Daniel is an Austin Texas transplant after 40 years in Chicago. His speaking takes him around the world each year as he shares his vision of the role technology will play in our future.
