Silicon Futures for May 2025 (Microsoft Build, Humain, Computex and more!

Jim McGregor of TIRIAS Research joined neXt Curve to cover the first half of another action-packed month in the world of semiconductors and accelerated and non-accelerated computing on the neXt Curve reThink Podcast series, Silicon Futures. The duo also shared their thoughts on IBM THINK 2025, CadenceLIVE 2025, and other events of note in the first half of May.

This episode covers the follow topics:

➡️ Highlights and key takes from CadenceLIVE 2025
➡️ Jim's recap of IBM Think 2025
➡️ Highlights and key takes from Google I/O and Microsoft Build 2025
➡️ Highlights and key takes from COMPUTEX 2025 and Dell Tech World
➡️ Highlights of May 2025 including HUMAIN and AI Middle East
➡️ Celebrating 40 years of Qualcomm!

Hit Leonard, and Jim up on LinkedIn and take part in their industry and tech insights. 

Check out Jim and his research at Tirias Research at www.tiriasresearch.com.
Check out Karl and his research at Cambrian AI Research LLC at www.cambrian-ai.com.

Please subscribe to our podcast which will be featured on the neXt Curve YouTube Channel. Check out the audio version on BuzzSprout or find us on your favorite Podcast platform.  

Also, subscribe to the neXt Curve research portal at www.next-curve.com for the tech and industry insights that matter.

Transcript

Leonard Lee: 0:09

Hello everyone. Welcome to this next Curve Rethink podcast episode where we break down the latest tech and industry events and happenings into the insights that matter. I'm Leonard Lee, your host and executive analyst of Next Curve, and in this Silicon Futs episode, we'll be talking about a lot of stuff actually that occurred in the month of May. Moving track of time. It's amazing how much is going on in the semiconductor industry. But, I'm also joined by my co-host, Jim McGregor of the Highly Accelerated 3D package. did I already say accelerated Curious researchers? Yes. You

Jim McGregor: 0:52

already did. Accelerated? Yes. Oh, geez.

Leonard Lee: 0:58

I think wanna start this again? Help me out do this.

Jim McGregor: 1:03

Okay, so let's just say welcome to the Next Curve podcast from two exhausted analysts from a wicked month of May. Wicked with tons of announcements. Tons of shows. Yes. And all the opinions you hear are the analyst's own opinions and not those of, next Curve or Terry Research. Actually, they are Terry. Yeah, they, because I am Terry Research, we'll exclude that part of the disclaimer. This is awesome.

Leonard Lee: 1:29

Yeah. Okay. This is working out really great. So we're gonna be covering a number of, key events that occurred, in May, cadence Live, Microsoft Bill, Google io, copy next 2025, which Jim and I were both at. And then, we'll commentary on Qualcomm's 40th anniversary. So. remember to like, share, react, and comment on this episode, and also subscribe here on YouTube and buzz route to listen to us on your favorite podcast platform. And also make sure to follow TEUs Research because you, you don't know Jim on the guys at TEUs Research.

2:09

And they're the solutions. Wait,

Jim McGregor: 2:11

We're, no, we're the gate squad. we're the insanely technical guys.

Leonard Lee: 2:15

I'll vouch for that. So, hey, why don't we get started? Jim,

Jim McGregor: 2:19

what do you say? I definitely, first thing I'd say is you forgot a few. You forgot Dell Technologies world. IB think, Microsoft Built, yeah. No, it's, it is a ridiculous one. I, I guess the one wrapper we could put around everything is. Ai. AI and ai. More ai. Yeah,

Leonard Lee: 2:36

absolutely. And you know, it's interesting how, the ground truths are sanctioning rapidly and I think maybe used as pockets. Are may a few steps behind, but we'll get into that in just a bit. let's start off with Cadence Live really quick. I attended, Carl Flynn was there as well, our buddy who couldn't make this episode. it was a really packed event. they had a wonderful agenda. Unfortunately, we weren't able to access many of the sessions independently, right? So we were assigned specific sessions But in terms of two takeaways. AI supercomputing, is really driving, the momentum for cadence. one of the big announcements that came out of the event was the millennium m of 2000, which is, based on Nvidia Blackwell, platform. And this is a big deal because, for those of you who don't know, cadence has their own. silicon that they, as well as systems that they build, to run their EBA software. And so this was an announcement that was accompanied by a, seemingly omni precedent appearance by Jensen who joined, uh, on who was the, CEO of Cadence, to announce his, partnership as well as product. So that was the big news item coming out. But, just a couple of real quick, points, Jim, a few of the big things, EDA two systems, design, automation, so SDA, so this whole concept of, really going beyond just the chip, two systems, right? Ip. Ip the

Jim McGregor: 4:22

systems, yeah.

Leonard Lee: 4:24

Ip, the systems, right. And then, this concept of digital twin and becoming, say, a more prominent thing, especially in this, partnership between Cadence and Nvidia, where NVIDIA's bringing a lot of their DNA around digital phones and merging it with, a lot of the design tooling. And assets for simulation that, cadence has. Right? And so the, this is gonna have a huge thing passed in the future of, systems designs beyond the standard conductors.'cause one of the things that they were talking about was drug design, right? Which is different from drug discovery. So, those are some of the things that came out in terms of acute types here. This is a really interesting, Comment that, honor, who is the CEO, mentioned, about packaging. Jim, I want to get your reaction to this. He says that there is going to be a move toward 2D packaging before 3D and 3.5. So this whole idea that we're moving towards advanced packaging, where we're stacking things, he seems a little bit less bullish about that being sort of the priorities going forward. Who's thinking we're going? Wide, and you know what I'm saying?

Jim McGregor: 5:41

Well, you have to remember that, when we think of the most advanced packaging, we think of the latest GPUs, the latest processors and stuff like this, and yeah, they're going to use 2D and 3D packaging. Matter of fact, some of the latest. Stuff we've seen from TSMC and Intel talking about, you know, CEL and some of the MT stuff, that is going to be for that segment of the market. But percentage wise, in terms of the unit shipped per year, that's a very small percentage. The vast majority are still these microcontrollers and microprocessors used in a wide variety of applications. They are going to chip technology very, very quickly and we're gonna see that being able to integrate new ip, be able to, spin new designs very quickly. And yeah, a lot of that's gonna be 2D but I would say all of that 2D and 3D packaging is intertwined. So, you know, I would agree that a lot of it's just gonna be 2D'cause there's not a need to do a lot of dice stacking. When you look at Your monitor or your TV or some of these other applications may not need that level. But I would still say that, all that packaging technology is definitely intertwined.

Leonard Lee: 6:51

Well, yeah, and his point was, the challenges with thermals, right. Thermal constraints, but his caveat was memory is probably gonna be stacked. it's just that, the rest of it, because of the thermal, constraints and considerations, he just seems more momentum, toward, scaling. in two dimensions versus three. So, I just thought that was a super interesting comment that we made.

Jim McGregor: 7:16

I think the cadence is some of the stuff that comes out especially with Cadence and Nvidia, Nvidia developed their own cool litho, so their own Yeah. AI solution for developing chips and doing the layout of. Chips and obviously that's integrating with all the EDA tools that we see out there. So it's interesting'cause the individual tools are getting smarter with ai. But at the same time, I think we're gonna see agents eventually that are leveraging all these tools, all these models, all this capability, not just from the E-D-E-D-A vendors like Cadence, but also, some of this technology from NVIDIA and others that are developing very, very advanced models. yeah, and now I wanna be, very careful because, the semiconductor industry still doesn't use AI to do a lot of the steps in developing silicon today, say,'cause they don't necessarily trust it yet. But yeah, I would say by 2030 I'll be amazed if there's actually any physical designers doing layout. I would imagine the tools by that time will be so advanced.

Leonard Lee: 8:18

Well, it'll be supported. I think one of the challenges that AI is going to be able to support designers with is dealing with the increased complexity, right? Yeah. And so, that's something that you and I talked about a couple of years ago maybe. that is why ai. Is important and why? it has a role in scaling and accelerating or at least sustaining the pace of getting to market, in a way, supporting Moore's Law that is really the reason why EDA is so important, we're now increasingly, SDA is and will continue to be increasingly important.'cause nowadays we're talking about systems not just. Simple like,, ip, right?

Jim McGregor: 9:07

That, and I think people need to realize that the learnings here translate to a lot of other segments, like you said, drug development or discovery. a MD in one of their keynotes at, or their press conference, pulled out, radiology. And, analyzing healthcare and using AI in healthcare, that's one of the areas I've said for years, if we just had access to all that medical information, AI can do things much better than, a radiologist can do. Matter of fact, We're now using AI a lot of times to control robots, to do surgeries for like hip and knee and shoulder replacements, and rightfully so, they're less likely to make mistakes. Especially'cause you have to tell'em what leg to actually operate on, but No, just kidding. but no, it's so between the ai, the robotics, they're so precise. It's incredible. and while you were at Cadence Live, I was on the other coast at IBM think. Yes. So, and obviously that was all about, that was all about AI as well. they introduced some new granite models, which, are, they're, and they're open sourcing a lot of this information. A lot of the stuff that they're doing around Watson X, they have, they, they keep evolving orchestrator and this is an amazing tool. Now it's to the point where it's actually going to be an all AI development solution. You can actually use Orchestrator to help develop models or. And they used a great example where they took a whole bunch of different models and created an agent within five minutes for a specific function. So incredible stuff that they're doing. Obviously they have their new mainframe, the Linux one out there, and that follows on their announcement with the Z 17, mainframe. So. They continue to expand their portfolio and their partnerships. They announced, expansions with Oracle and Salesforce. So they continue to drive. You know what, Francis and I like to say, they're kinda like an AI ready button for enterprise, they have, yeah, they have the engineers, they have the models, they have the hardware. They have pretty much everything. Just says, okay, if you're really an enterprise, boom, here you go. Wow. That's it. That, well, there, there's a lot of stuff about, everything else that they're doing with their partnerships, but it was really about how they continue to accelerate AI for enterprise.

Leonard Lee: 11:16

Okay. Yeah. now let's talk about build. I thought build was really interesting this year. Build Zelle type world and copy text happen all at the same time. And

Jim McGregor: 11:30

no, and while there was also a Google io and there was the embedded, vision Summit. so yeah, there was five events going on the same week. Yes. Let's go back. Let's see if we can unpack those, first off, Microsoft build, it's, I would say build, and, Google io are on the same theme. It's about developing tools, enabling developers, and enable, to develop, new AI solutions, as well as, new assistance or new agents. To be able to use those. Obviously Microsoft, the big thing for them is co-pilot. So trying to advance co-pilot and further copilot ca capabilities. they have the, Azure AI foundry and co-pilot studio, really trying to enable agen AI solutions. yeah, and Google obviously is looking at AI powered tools for a whole bunch of different applications. Anything from no AI filmmaking image generation to, their normal search and, and email. So,

Leonard Lee: 12:27

yeah, actually a lot of these events, you started off with IBM Bank, they're all similar in terms of, what they're trying to. Jockey four. Right. Which is, these companies, IBM, Google, and, we should throw AWS and M mix as well. Microsoft all, um, v your, attention in the market for these AI frameworks. AI tools. Yes. Above the supercomputing, above the GPU as a service or neo clouds or whatever you wanna call it. And yeah, agentic AI being that, theme Azure. That was a common thread. Definitely. Yeah. Right. And we thought with, Microsoft, Google, but interestingly, with Google. In particular, as well as Microsoft, actually, there's this whole notion of an agent web. But also this idea of the infrastructure that was supported at Google's cloud. Next, they have 2 million miles of fiber and so they have this concept of a cloud

Jim McGregor: 13:47

Right? An infrastructure,

Leonard Lee: 13:48

So when we're talking about Ag Gentech web, there we're seeing signs of. this manifesting itself, or at least this concept forming from infrastructure, right? Mm-hmm. Because a lot of the hyperscalers, including like meta and such, have been investing their time. And overseas, cabling, right for internet. Putting these fat pipes under the ocean, connecting continents and, running fiber at,, pretty much, anywhere that they can. and then now you have this adventis AI layer on top and, lots of mention of MP, right? It really is a rag protocol. But then you also have agent to agent protocols, out of anthropic. But the bottom line is this, is that, there's this whole layer. At the platform level. Right. at the orchestration middleware level. That's, starting to form that at global scale or, or at least the ambition was to make it global scale, which is super interesting. And one of the things that came out of, Microsoft builds that I thought was really, curious and also potentially very compelling, is this idea that they would be hosting trusted MCP. Servers, sort of like a registry, Where their developer communities, right. So anyone in all likelihood it's gonna be multi-cloud. but then it would be also multiagent platform, if you will, to kind of stitched together with the MCP protocols. So, there's a lot of interesting stuff happening. very ambitious early days. But it's really interesting how the likes of IBM, Microsoft, Google Meta, Amazon are thinking about this. Next. Stage and,, distributed or computing.

Jim McGregor: 15:45

Yeah, and we have that side of it where it's really focused on the software, the services, the solutions. And then we have the other side that are kind of the Dell technologies world and Computex, Focused on really the hardware and everything it builds up to enabling that. The solutions and build those up. Dell announcing, all their, solutions from I PCs to servers, including new workstations. and the same thing at Compex. Matter of fact, Compex was kind of interesting show this year. Last year was the, major introduction of the A IPC, or actually it was all copilot plus pc. This year, it seems like the industry went back to a I pc. It's not all about Microsoft anymore. yeah, and it's, mostly major announcements were about, workstations. Matter of fact, a MD, Intel and Nvidia all announce had workstation announcements, while. the only really a I PC focus was Qualcomm, and they didn't announce a new part. They teased it. Matter of fact, they teased two new parts, one for the PC and one for the server. So we're gonna see that new Yeah. Architecture that they acquired. The Orion, going not just into, not just their next evolution of that for a I PCs, but also a server part that they're gonna be discussing later this fall at their Snapdragon technology summit in Hawaii, which I'm assuming we will both be at. So, yes. Some of the big announcements there. Computex is always great to go to. It's in Taiwan. Yeah. You know, it's a major source of semiconductor manufacturing and electronic assembly. Matter of fact, I would say 60% or more PCs are assembled there. Not to mention, everything else from embedded systems to TVs, to you name it. So Taiwan one is a very important market, so it's natural to have it there. So Conex had not just, uh, obviously there was everything pc, there was PCs, there were PC monitors, there were PC monitor holders, there were keyboards. There were even components for keyboards. There was everything PC there, but. I would say about 25% of the show floor that I went on actually was also embedded, computing solutions and factory automation, which makes a lot of, a lot of sense. The factory automation, since they do so much assembly work there. Matter of fact, I don't know about you, but. I was promised a whole floor of fans, PC fans, and I never found it. I promised me the answer.

Leonard Lee: 18:13

I think it was

Jim McGregor: 18:13

you.

Leonard Lee: 18:14

And this is why you have to show up. And I'm really glad that I went to Compex this year. I wasn't gonna go. You kind of, you know, convince me, but I think it's a bellwether now. It's become a bellwether event, right? Because everything seems to converge. On Taiwan, and I'm not just talking about semiconductors, I'm talking about the greater electronics industry. And I think this is a really important point for our audience to understand. the reason why you didn't see as many fans is because there's so much water cooling, or, liquid cooling.

Jim McGregor: 18:49

Right. Mostly direct. Oh, oh, oh. Oh, really? Really? I didn't see a whole floor of liquid cooling either. No, you didn't. But it was dispersed everywhere.

18:57

Uh, why?

Leonard Lee: 18:58

Hey, did you notice there's a bunch of companies that are sprouting out, even at the Vertex booth, which was huge. The comparatively speaking, right. immersion. Full immersion. Right. Full immersion. So, which I still say is a bad idea. Yeah. It, it's a totally different dynamic, right? I mean, yeah, the equipment's different. The cooling systems are different, right. But it was interesting to see, that many of the vendors were starting to showcase some of that. And, who knows? last year, in the roundup of Blackwell. There was all this direct, liquid cooling. And that wasn't necessarily a thing before. People weren't really talking about the cooling. And so I actually like the tip why of the fans are disappearing, dude. And also a lot of the PCs, a lot of the PCs liquid. That's, I can see a lot of fans, man.

Jim McGregor: 19:52

Well, on the desktop, not the mobile,

Leonard Lee: 19:55

the desktop work stations desktops. Right, right, right, right. But, um,

Jim McGregor: 19:57

And. And the workstations were definitely the hit there. Oh yeah, yeah. Intel talked about their ARC Pro B 50 and B 60 for workstations being an alternative to, obviously a MD and Nvidia. Yeah. Especially in terms of more memory at a reasonable price so they can handle larger models for inference processing. A MD stepped out of the box. Both with a new thread ripper. Which I love the thread ripper. Both. That's the ultimate workstation. processor, the 9,000. And then they also talked about their new, radio on, cards for, the workstation and high-end PCs. and then, not to be outdone, Nvidia obviously was talking about. Their, RTX PRO Solutions and they're new and yeah, their DGX workstation and their DGX, spark the little desktop workstation. Now, there's been some rumors about those, having some problems, but they said both of'em would be shipping in weeks. So, according to them, they're on track for both of those products, and I'm really excited about it. I've actually signed up to be one of the, early, acquirers of A DGX spark. I'm excited about having that on my desktop. Cool.

Leonard Lee: 21:04

Yeah. I had a chance to meet up with, Rob Herman, who was the workstation boutique at Lenovo, and he gave me a preview of the think. in station PGX. So, I asked them, Hey, kind of borrow one destroy. So who knows?

Jim McGregor: 21:21

might get one too, Jim, but, they're pretty, a pretty awesome solution. no, I'm very excited. Yeah, and wonder, they were great, right? Yeah. And there were great workstation and PC solutions there from Aser. From asus. matter of fact, ASUS probably had the best booth there with both Asus on one side. R on the other. it was incredible.

Leonard Lee: 21:40

Yeah. And so one of the things that I noticed at Compex is last year there was like a wall of MBL 70 twos. Right. this year it was all about MVF and you mentioned, the RTX platform mm-hmm. For enterprise and enterprise is also another. Big theme, especially in light of inference. And, one of the things that I really liked about the intel, demos, the GPU and gaming demos, was as you mentioned, the workstation, Aspects that they, showcased, running large language models in particular, those is one where they were running, the entire, uh, Boote R one, model on a, it's more like a workstation. Were like four B 50 cards in there. I think that was a really cool demo and very, very telling. Mm-hmm. In terms of. Um, how actually inference compute looks different than the compute for massive scale, model training, right? yes. And so it was a, it's a lot of good stuff there and unfortunately I think I'm gonna have to go every year now., I don't know. Food is good. So you, the food is spread, transportation is cheap. So yeah. Might not be all that bad of a thing, but, yeah. And so MGX, that, that, I thought that was really interesting. So things are changing, the focus is changing. and yeah, I think you got a lot of that out of comp text. Oh, yeah. In that Dell Tech world, the thing that was,. The surprise, was de the new, what do you call it, Dell Pro Maps. That's features Qualcomm, yes. Accelerator. the AI 100. It's like a Ference card. So, yeah. I'm gonna reinforce your statement that workstations are a thing and I think they're becoming That sort of, bridge. between the super computing and the cloud with enterprise. and edge Well, and it's gonna be enterprise super computing. I don't know, but it's a smaller scale.

Jim McGregor: 23:58

It's also gonna be interesting because, as the industry's changing and accelerating so quickly, it really is, to see how things change. last year we saw the, introduction to UA link to compete with, NV Link, NVIDIA's, GPU to GPU interconnect. But now Nvidia is saying, Hey, we're opening it up to anybody use with NV Link Fusion. So, and so it's gonna be interesting. Interesting to see. I, I think, and we saw Nvidia open up the NVL 72 rack architecture, to OCP, late last year. So It is interesting to see, how everything's evolving and, even some of these islands, these competitive islands we saw, seem to be breaking down just to advance the industry going forward. and I think that's necessary. we saw the huge write down by Nvidia for the age twenties, not being able to ship those into. China and, Jensen made the, comment to the president saying, listen, if you wanna make us a dominant force in AI throughout the world, our technology needs to proliferate throughout the world. And I think, all of the tech companies that we're there, no matter what country they were from, would agree with that. nobody wants to see boundaries in our industry, and it's interesting to see how the industry is really coming together, to develop advanced solutions.

Leonard Lee: 25:14

Yeah, it's this, there's AI and then there's super computing. Yes. And one of the things that I've noticed. And I've been saying this for a long time, we're not really talking about AI here. The vast majority of what we're talking about here in terms of the revolution of the booms is super computing. AI still has to prove itself. now AI happens to be the workload, but, fundamentally what we're talking about in terms of infrastructure build out this is super computing fundamentally. this is going to be the challenge for the entire industry, including Nvidia, is now how do we actually make AI, stick? I think that's the big challenge right now. Even though we do see, companies investing a lot, the monetization is not apparent. And these are the things that I think are gonna be a challenge for the next year. there's still a big question mark this year, but there's no doubt, I think. this moves toward the edge as well as, this effort to bring some of the supercomputing down to the enterprise level. I think it is gonna be an important dynamic to keep an eye out on, And that will hopefully help AI diffuse. But the question of AI leadership, I don't think is entirely. Hinged on Supercomputing. There's also the models, right? Who has the best foundation models, because that as the word to Jeff, is the foundation for who has the best ai.

Jim McGregor: 26:48

oh,

Leonard Lee: 26:48

have the

Jim McGregor: 26:48

best ai. I would agree with you. I would say the two based innovations for the AI industry, and to really help the AI industry have been llama and deep seek, you know, two open source models, solutions that people can grab and use very effectively. But I think we've passed the point. I kind of disagree with you. I think AI is proving itself. That's fine. Very quickly, just looking at the growth of tokens, especially with the recent reports from Microsoft and Google, it's huge. And that's just tokens, that's not images and video and everything else. I think that, we are finding a way to monetize AI both with consumers and with the enterprise. And I think agentic ai, gen AI was I think, the first step. But I think agentic AI really leapfrog that with, companies realizing the value of AI and how to start implementing AI now, it's still a struggle. I agree with that. Still a struggle, but I don't think anybody can look at their business anymore and not see some value, with Gen ai, Orent ai, and I think it's coming just so rapidly.

Leonard Lee: 27:52

It's a learning process, right? Mm-hmm. And that's what I'm seeing. there's a push and then there's a, a bit of, let's say a throttling back just to see, okay, how far did we get? but I think it's a real, my observation is a learning process for everyone. Again, going back to this is moving so freaking quickly, it's really hard to keep track of everything. And if you're just looking at one layer, and you're not considering the impact above the, a particular layer or below, you're really going to miss something important. but. I wonder what June's gonna be.

Jim McGregor: 28:29

Well, the one thing that I would say, and first off, we've seen huge advances in, the models. Deep seek was a monumental moment and a push towards that reasoning slash ent AI solution. But also look at what we've done on the hardware site. That DGX Spark is equivalent to a DGX one. And so from 2016 to 2025, we went from this massive, computer solution down to something that fits in the palm of your hand. in terms of compute performance, it's a. It's

Leonard Lee: 29:02

basically the nook up the size of a nook. Yeah.

Jim McGregor: 29:04

Yeah. It is incredible what we've done. Yeah. That more than anything shows the innovation that's going on in our industry.

Leonard Lee: 29:12

Yeah. And physically in independent of, AI itself, I think that is the big revolution. In going back to super computing, how, basically this whole AI push has driven a revolution in democratization of super computing. And, if you wanna call that accelerated computing, and if to borrow from Jensen's parlance that, that is the. A phenomenal thing, right? Yeah. And all the capital investment that has gone into shrinking all this thing stuff down into basically a DGX bar. Yeah. One pet flops that's like great. Pretty incredible. so, hey, the other thing in New York, Qualcomm celebrated their, 40th anniversary. Yeah. Which is pretty amazing. I remember when that company first started. Yeah. Radios, trucking radios. Yeah. Erwin Jacobs was a professor at my alma mater.

Jim McGregor: 30:19

Tracks or something like that? yeah. Trying to remember what, yeah, so I mean, that, that is amazing. they were, they have been the powerhouse really behind cellular technology, especially since the two G generation to 3G to 4G to 5G industry leader, and. Just also, a leader in terms of, heterogeneous computing and low power of, power efficient computing. Yeah. So, yeah, it's fun to walk into their headquarters and see the walls of, patents that they have. Yeah. Those are always fun to see. It is amazing what that company has done and how important they've been to our industry. Yeah. And how they continue

Leonard Lee: 30:56

to expand and diversify their business We define themselves actually. And so, yeah, Qualcomm, now you have Snapdragon and Dragon Wing and, who knows what else, because, they're starting to get into data center, so I wonder what they're gonna call it, the data center part of their business.

Jim McGregor: 31:13

I don't know. We'll have to wait and see.

Leonard Lee: 31:15

Oh, yeah. And, all the stuff in the Middle East. Humane. Right. That was the other big news item. So yeah, we'll see how that pans out. there are a lot of big numbers thrown around.

Jim McGregor: 31:26

you know, well, you know, Stargate is coming along pretty well, so, stargates doing so well in terms of, what is it, something like five or$600 billion investment into a sovereign AI solution for the us. I'm sure that we will see similar investments from, other countries around the world.

Leonard Lee: 31:44

Yeah, well, you know, I think the Middle East, there is definitely a value proposition in terms of the location because of, their access to energy. that part I think is very compelling. So, the Trump administration, basically rescinding the AI diffusion. rules was an act, it was rules, instituted by the Biden administration before they left office. that definitely had a big impact in pivoting, this AI narrative, right? Quote unquote AI narrative. So that was the other thing. let's see where things go. some big numbers thrown around though. Yes. congratulations Qualcomm, for your 40th Anything else before we tap off this episode of Silicon Future? I think the

Jim McGregor: 32:33

rest of my mind is just melting

Leonard Lee: 32:37

literally every week. Yes, there's like an entire year of headlines the pace is just outrageous. I mean, that is definitely a fact.

Jim McGregor: 32:50

I think the entire industry should take a three month break just so we can catch up.

Leonard Lee: 32:54

I was thinking of taking a three month break just to think about all this stuff. It's difficult to have time to synthesize, you know? It is. But, Jim, always great man to chat with you and disagree, constructively disagree on things. Otherwise, it'd be so boring. You would be so

Jim McGregor: 33:14

boring. You realize that, right? Just remember, it's curious. Research what's right. No, just kidding. Okay. Yeah,

Leonard Lee: 33:23

no, you're not kidding. that's the funny thing. That's the really funny thing.

Jim McGregor: 33:28

No, it's always good to do stuff with you. I'm glad that we get to, go around, work with these companies, talk to these companies and see all the exciting stuff that's. It's incredibly upsetting. It's a great time to

Leonard Lee: 33:39

be a geek. Oh yeah. Oh yeah. And evidently you're, not a bad sector to be in the geek sector. So with that, hey Jim, tell our audience how they get in touch with research and,

Jim McGregor: 33:53

You can always reach out, you can always, reach out to our website@wwwdottiresearch.com. That's T-I-R-I-A-S research, or you can reach out to me personally at Jim Attis research and all my contract information's available on the web. also look for us on LinkedIn. X Facebook, mastodon, blue, sky, red, on and on and on. So, look for us, wherever you choose to social, network. Okay, awesome. And remember,

Leonard Lee: 34:25

please subscribe to our podcast, which will be featured on the Next Curve YouTube channel. check out the audio version on buzzsprout or find us on your favorite podcast platform. And also subscribe to the next curve research portal@www.next-curve.com for the tech and industry insights that matter. And we'll see you next time. Take care. Cheers.