# Dell is the Hottest Trade in AI.

https://www.youtube.com/watch?v=WxWayeyjqig

[00:00] 30 days ago, President Trump went on camera and said the following sentence, "Go out and buy Dell."
[00:03] Since then, Dell stock has surged 80% following two back-to-back blowout earnings, giving them the biggest tech comeback in history.
[00:13] Their stock is up 240% this year alone.
[00:16] Now, Dell is probably the most boring writtenoff tech middle-aged company that I can think of.
[00:21] They were the guys that sold my school's IT department, Serverax, back in the day.
[00:27] But a person that goes by the name of Jensen Huang saw something different.
[00:31] Dell sells servers and data center racks that are ideal for training and inferencing data models, specifically the GPUs that he sells.
[00:36] Now, the smart money is showing that AI is flowing down from the chip makers to the guys that actually put the chips together and run them.
[00:44] That's what Dell is doing right now.
[00:45] And in my opinion, Dell's put on a masterclass of how to pivot from a regular company to an AI company.
[00:52] The Dell story is incredible.
[00:54] Do you have any idea how old Dell is?
[00:58] No.
[00:59] Very old.
[00:59] 40 40 years old.
[01:02] 1988. And what's most interesting and I find and the only reason why Dell is probably sitting in this position is because of the founder who has been with them since the beginning of time, Michael Dell.
[01:11] And he's really done an unbelievable job in moving this company through all the phases that have existed over the last 40 years.
[01:18] When you think about 40 years ago, we didn't even have consumer laptops, consumer computers.
[01:21] We didn't even have the internet.
[01:24] So Dell has been around longer than just about anything.
[01:27] Michael Dell has navigated through all of these transitions and this most recent transition to AI, he has done so again successfully.
[01:32] Dell has not had the easiest route.
[01:35] In fact, if you switch the chart to all on this thing that we're looking at here, uh you'll see in 2013, they actually disappeared off of the stock market entirely.
[01:45] Michael Dell took the company private at a $24.4 billion valuation.
[01:51] Fast forward to today, clearly that is different.
[01:52] Clearly, they've figured out something new and novel.
[01:54] And that's, I think, what a lot about what we're going to be covering on this episode is the transition that was made most recently, the pivot, into becoming one of the
[02:03] major arms dealers in the AI race.
[02:07] So, who would have thought that this really old school company could turn itself into a cutting edge leading edge company?
[02:14] And gosh, there are people who've been holding the stock for decades who are very happy about what's going on.
[02:16] But let's get into what's going on.
[02:17] EJ, it's clear that Dell is kind of becoming the picks and shovels company.
[02:21] They sit one layer down from Nvidia.
[02:23] What is their role in this large picture?
[02:26] I mean, two weeks ago, we had that episode where we talked about the AI stack.
[02:28] We talked about where each company six.
[02:30] Where does Dell sit in this picture?
[02:32] Before I do that, um, what's hilarious is you asked me to toggle to all, but like we can't go back prior to 2013 because they technically went off the public market.
[02:42] Yeah, they delisted.
[02:45] So, they delisted.
[02:46] They went private for I think it was like $24 billion and then they were kind of like in stealth mode for a bunch of years and then came back uh in an IPO in 2018 I believe and that's the stock uh market that we're seeing or stock chart that we're seeing right now.
[02:59] But to to answer your question specifically, everyone knows that in a gold rush, what is the line?
[03:04] You got to be the one that's selling the shovel specifically.
[03:08] And that's what Nvidia became for the entire AI boom.
[03:12] AI models went kind of viral.
[03:14] Everyone was using chat GBT and smart investors thought to place their money in the chip maker, the GPU maker.
[03:21] Now, what is playing out with Dell specifically is someone has to be the one that assembles the shovels together into a kind of like a working mine, right?
[03:30] Into like a working product.
[03:32] And that's what Dell does.
[03:34] They don't make the GPUs, but they make the racks, the cooling systems, and they help integrate all of Nvidia's GPUs to actually make them function.
[03:41] There was a story that we covered, I think it was like a few months ago, where Microsoft had purchased, I think, $300 to $500 billion worth of GPUs that were sitting in warehouses collecting dust because they couldn't get the power to set this up and power these GPUs.
[03:55] They also couldn't get the cooling systems and the wiring to get all of this set up.
[03:58] Dell kind of like spent the last 40 years creating and perfecting server racks for a completely different kind of
[04:05] adjacent industry.
[04:06] It was like the computing side of things.
[04:08] It was the gaming side of things.
[04:10] And then recently they kind of pivoted to focus specifically on making their hardware perfect for AI models inferencing and training.
[04:16] And that's what has led to this like kind of massive partnership.
[04:20] Now there's currently a really big conference happening in Taiwan.
[04:22] Taiwan of course is obviously the epicenter of semiconductor manufacturing.
[04:26] You've got TSMC out there.
[04:28] Judson Huang flies there.
[04:30] I think it was like four to five times a quarter to kind of like chat to TSMC and make sure that Nvidia chips and designs are being constructed in the way that he can.
[04:37] He's had a lynch pin on all of this and there's a big conference going out there right now where Jetson had some really important uh news updates to share.
[04:45] Now, the first one, the one that was being teased over the weekend came from none other than Michael Dell himself, the guy that is a CEO and founder of Dell and he goes, "We have the first Dell/ Ninvidia Vera Rubin NVL72 rack um live and going."
[05:03] And I actually have a picture of this right now.
[05:04] Now, if you're wondering what uh the Ver
[05:06] Rubin rack is, it is um Nvidia's latest gen uh GPU rack, and it is going to be the rack that ends up training the frontier models from Anthropic and OpenAI starting probably next year cuz there's usually like a 6 month lag.
[05:20] They announced it about I think it was like four to 6 months ago and now we see the first live rack coming to fruition.
[05:25] It's pretty cool.
[05:27] So this is amazing and unexpected in the sense that I I would have probably expected to see someone like XAI or Anthropic or OpenAI getting the first Vera Rubin chips working and actually running.
[05:36] So it's interesting to see it coming out of Dell of all places.
[05:39] And I think the reason why we see it out of Dell is because Dell is in a very unique position.
[05:44] You mentioned like the picks and shovels and how it's creating a working mine.
[05:47] Dell is hardware and AI distribution as a service.
[05:53] It basically builds AI factories for anybody who wants them and this is their wheelhouse.
[05:57] One of the most interesting things that I guess over time you start to develop relationships with over your 40-year career is relationships with people who have sensitive information.
[06:05] We're
[06:06] talking banks, hospitals, governments who are involved in defense.
[06:11] A lot of these kind of more secure entities aren't really able to go off and use OpenAI or Claude because a lot of the data is secure and they can't let it leave their servers.
[06:20] What Dell is able to do is take this hardware ability, their ability to print out these mines, per se, and then give it to these individuals.
[06:27] So, give it to a hospital who really needs to use AI, but can't let the secure customer information get off of off premises.
[06:34] Same thing with banks. Same thing with governments.
[06:35] And what we're seeing here is a really strong pivot from them in that sector in a way that I think this they're making a lot of progress.
[06:43] I mean, Vera Rubin is one of the first things that they've done, but they've also been working on a lot of other really promising ways of getting AI hardware into the hands of not necessarily your average consumer, but further down the stack where if perhaps you don't have $und00 billion to build out a data center, you can actually just go and spend $80 to $120,000 on the new workstation.
[07:03] So just last night, Nvidia Jensen Huang went on stage in Taiwan and he was announcing a
[07:07] Bunch of new hardware.
[07:08] And one of the hardware pieces that he released was this DGX station for Windows.
[07:13] What is it?
[07:13] It's basically a black wall chip in a desktop, which is the most powerful consumer grade supercomput in the world.
[07:20] And I'm not even sure we could call this consumer grade cuz I'm not sure how many consumers are going to spend $80,000 plus dollars on a computer.
[07:26] But essentially what you have is a blackwell chip that can sit inside of a room without a custom tooling, without custom data center, and it can run a trillion parameter model locally on your machine.
[07:37] So they're kind of moving down the stack as well.
[07:39] And I think this is what investors are seeing.
[07:40] They're seeing them partnering with companies like Nvidia and really making strong progress in the world of getting AI accessible to people through data centers, through this consumer hardware, all the way up and down the stack.
[07:52] What what I like about this is typically um if you wanted to go out and train your own model from scratch, you would need like billions to hundreds of billions of dollars to to be able to do that.
[08:01] It was a very siloed uh kind of like access only to the rich.
[08:05] Um, and what these new updates or
[08:07] hardware updates are doing is basically disseminating it to like the average kind of like medium-sized startup that maybe has raised uh, let's say 10 to 100 million in funding for them to be able to buy out like a server rack or a couple server racks of this new Nvidia DGX station.
[08:21] They should be able to kind of like train some kind of a model, right?
[08:23] Run running locally.
[08:24] Now, if you're listening to this and you're still like, well, I'm just a regular consumer.
[08:28] I'm not that business and I want to be able to access and run some of these models.
[08:31] They had another update.
[08:33] Um, Nvidia announced something called uh their new ARM based processor called RTX Spark.
[08:38] So, it's basically the equivalent of a GPU uh RTX 5070 that runs on your laptop.
[08:43] So, I'm talking about like uh not even like a desktop PC, like an actual laptop.
[08:46] It runs at uh 100 frames per second and it's, you know, it's it's pioneered or oriented around modern games.
[08:52] But the coolest part about this is you can technically run some form of an open source Frontier model on your laptop now using this new GPU or this new chip.
[09:01] Um, and what I love about the directional trend that companies like Dell, ARM, and Nvidia are kind of like aligning together to build
[09:09] is running models on bare metal or locally at home using your own private data.
[09:16] And I think this is we've said this on previous episodes, but this is incredibly important because you don't want to hand over all your personal data to companies like Anthropic and OpenAI.
[09:24] At some point, it gets a little weird when you start sharing medical records or banking stuff or personal lawyering things.
[09:31] I know a lot of friends personally that do that already.
[09:32] They kind of like upload their therapist notes and stuff and get like kind of information.
[09:37] I'm like, "Hang on a second."
[09:38] So, it's so Sam Alman technically gets access to that right now, right?
[09:40] They can use it for whatever they want.
[09:42] And I think uh it's increasingly important to own your own hardware going forwards, which is so weird when we have been in an all cloud world.
[09:50] Now, when you were describing earlier on Josh, the fact that like Dell is enabling kind of like bare metal training, you know, onrem running, it reminds me of that uh story we covered with Amazon.
[09:59] Ramon, we were given the Amazon bullcase and we were talking about like them building out Amazon cloud specifically for AI servers.
[10:06] One of their main assets is they run all the governments across the world's AI models
[10:11] because they can like set up racks for themselves and Dell is kind of becoming a feasible competitor to that.
[10:17] So yeah, I want to take a little side quest here and actually talk about a little bit more of what Nvidia announced last night because it was really impressive.
[10:23] So we have that RTX Spark which is this new processor equivalent to an Nvidia 5090.
[10:29] Now, for those not familiar, the Nvidia 5090 kind of run of GPUs is the cutting edge top-of-the-line for consumers.
[10:37] And generally, if you're a gamer, that's what you always aspire to buy.
[10:40] That's how you can play your games at maximum settings.
[10:42] It's now available inside of a laptop.
[10:45] This chip used to cost thousands of dollars and be pretty beefy.
[10:47] Like, this is a huge chip that would weigh like, I don't know, something like 4 or 5 lb.
[10:53] They've figured out how to compress it into a laptop.
[10:54] And not only that, but it comes baked in with 128 GB of memory.
[10:57] So, if you're comparing this to a MacBook, you're thinking, well, the max spec of a MacBook is 128 gigs.
[11:03] Same with this new laptop.
[11:03] So, on paper, it's a direct competitor, at least as it comes to training and running models locally.
[11:10] Now, it turns out that
[11:12] Microsoft has actually went ahead and released a laptop with this new chip in it already, dubbed the Surface Laptop Ultra, and it seems as if it has the RTX Spark chip baked right in.
[11:21] It has a 15-inch mini LED screen, 128 GB of unified memory, and for anyone who cares, one pedlop of AI compute power.
[11:31] I'm not even sure what that translates to, but it sounds like a lot.
[11:32] And this is really interesting to me as someone who loves consumer hardware because I'm looking at this and I'm saying thinking to myself, "Oh my god, this is this is a MacBook Pro."
[11:41] Um, and it's not.
[11:44] This is the first time, I think, in a while that MacBook and Apple have had a legitimate competitor and something that can match it spec for spec.
[11:50] And maybe the hardware isn't quite as elegant.
[11:52] Maybe it doesn't have the single uni-body aluminum design, but on a spec sheet, it's really compelling.
[11:58] And it has all the ports that you would think of.
[12:00] It has a miniLEDD screen, which is exactly what the new Pro Display XDR has.
[12:03] This is a serious offering from Microsoft, from Nvidia.
[12:07] And I think it's it's a testament to where things are headed, which is like Apple might not have their
[12:12] monopoly for too long.
[12:14] Mac Minis have been sold out forever.
[12:16] Clearly Dell wants this, Microsoft wants this, Nvidia wants this.
[12:18] they're going to try to release their competitors and this is a really strong first stop.
[12:20] It's it's really nice to see like a competitor to Apple finally at least on like the the hardware mode now.
[12:25] Now, where my um where my kind of brain automatically goes is like Microsoft is like this antiquated company, right?
[12:31] Like all the governments use it.
[12:33] They they run like all the old Microsoft official
[12:36] and how much does their API suck?
[12:37] Like they're just falling behind.
[12:39] Dude, it sucks.
[12:40] Like I was listening to a like some kind of like podcast over the weekend where apparently Sachi Nadella is now like the PM of uh co-pilot like he's he's gone back to the PM role to try and like figure out their co-pilot thing like because the head of AI couldn't.
[12:53] So like Microsoft has been kind of been in the dark ages for a while and like this new update like this new hardware update brings them into like a a new light which is great but the number one question is like are you still going to use their software or their operating system?
[13:06] So naturally I was like okay well it's great that they have this new piece of hardware that's consumer accessible but like are there
[13:12] is there any new AI that I can use on this and the answer might have come from Jensen himself or Nvidia themselves.
[13:18] They released a new open source model called Neotron 3 Ultra and it is officially the top open-source AI model made by an American company.
[13:29] Um, so typically I've spoken about this a lot on our show, the top open source models have come out of China, which is obviously an adversary to the US, but Neatron Ultra is actually quite competitive.
[13:39] And we have like a chart here, a clip or rather a screenshot from uh Jensen's keynote yesterday or this morning.
[13:46] I the time difference is so baffling actually.
[13:48] Uh, which places Nemothron Nematron 3 Ultra on par to an extent with Kim K 2.6, GLM 5.1, and some of the other Chinese open source models.
[13:57] But it's so much quicker.
[13:59] The output speed is like by the looks of it like almost like three to 5x quicker.
[14:04] Uh which is great to see.
[14:06] Um I don't know what your view on this is Josh but I am betting that Jensen, Nvidia and Dell and all these companies that are kind of like forming an alliance to go hardcore
[14:14] on open source.
[14:17] Why?
[14:17] Because it's going to help with retail distribution across everything, right?
[14:21] Like Jensen doesn't know what consumers are going to use these models for.
[14:24] But if he can get more retail consumers like you or I to buy these different laptops and use his chips, he ends up selling more chips.
[14:29] Yeah, I know.
[14:29] I couldn't agree more.
[14:31] And there's a second factor to it too about why they're going to want to continue to improve and push the open source model frontier forward.
[14:36] It's because what we're seeing happen with companies like Amazon and happen with companies like Google is they're building their own custom AI accelerators.
[14:42] And now these companies are beginning to train on more of a proprietary stack.
[14:46] So Google now is training Gemini models largely on its own proprietary hardware.
[14:51] It uses Nvidia GPUs for a lot of it, but an increasing percentage of that training is done by these TPUs or in the case of Amazon, their tranium accelerators.
[15:03] And that's not good for Nvidia because they want the CUDA lock in.
[15:07] They want the Nvidia GPU Blackwell Vera Rubin lockin.
[15:12] And how do you get more of that lock in to happen?
[15:14] You actually just give away the software stack for free and make it
[15:15] optimally compatible with your hardware.
[15:17] So, as they're giving away this open-source uh software, they're saying,
[15:21] "Hey, you can go run it on our new RTX Spark chip.
[15:22] In fact, it's available in a laptop that you can go and buy today or starting in a couple months, I think.
[15:26] I think they're releasing in the fall, but they're starting to release the entire stack and they're offering the software as that teaser as that little value ad where you can't get anywhere else.
[15:35] If you want to run Nvidia software, the best way to do it is on Nvidia hardware and it kind of solidifies and oifies that mode even more.
[15:42] So, I think we're probably going to see this trend continue.
[15:44] We're seeing already this beating DeepSeek and a few of the other Chinese open source models on a few parameters.
[15:52] I'm sure it's just a matter of time until all those benchmarks are blown out.
[15:55] If Nvidia is really putting their full weight behind open source, we're going to have a serious push into open source.
[16:00] And that combined with the hardware that we're seeing with Dell, with Microsoft, it's going to create like a pretty compelling offering for local AI infra that I'm not sure the market has quite priced in just yet.
[16:11] So to bring the story back to Dell Josh, I I want to put my our skeptic hats on for
[16:17] It it seems kind of bubbly on the headline value, right?
[16:22] It's like okay, four-year-old company, it died.
[16:25] Now it's like crazy.
[16:27] Suddenly you're like the AI company.
[16:28] This kind of like gives me the old bird story, you know, a shoe company that pivoted to running AI GPUs.
[16:33] I'm just like kind of like what are you doing?
[16:35] What what is is Dell actually doing anything?
[16:37] Well, they recently released their quarterly earnings report.
[16:39] Um, which wasn't just a blowout.
[16:44] It was their highest recorded revenue quarter ever in it.
[16:46] And before you share, actually, I I want to note that the post that we're showing on screen is from Michael Odell, the CEO himself.
[16:52] This is a founder company who is proud of the growth.
[16:55] He is in the trenches making this happen.
[16:57] So, I think that's noteworthy.
[16:58] That's different than a lot of these other companies is like the founders been doing this for 40 years.
[17:02] Yeah.
[17:02] It's it's the same guy.
[17:04] They run out of the same office in Texas.
[17:07] It's the same HQ.
[17:09] It's been there for four years.
[17:09] Going off, going public, going off public, and then going back on it.
[17:12] It's absolutely unbelievable.
[17:14] Now, a summary of the recent quarter is they uh
[17:17] increased revenue 88 almost 90% yearover-year.
[17:23] Over or almost half of that revenue came from this new uh business unit.
[17:25] It's called AI servers revenue, up almost 800% yearover-year.
[17:30] Now, for all the skeptics that were thinking, hm, I think like Dell is kind of like just advertising things that they don't actually have, the numbers actually disprove that pretty aggressively.
[17:39] Um, they have over 5,000 major customers that are buying their uh serverax, including one of the biggest customer being Nvidia.
[17:47] Nvidia wants to distribute it through a hardware moat and Dell are the ones that are basically putting their chips together.
[17:53] They're bolting their chips together to make them efficient.
[17:55] They're the first ones to create the new Rubin racks.
[17:58] they have been kind of unofficially uh crowned a key partner to Nvidia and like Nvidia's success is basically going to drag them up.
[18:05] Um now the previous quarter to this at the end of the last financial year they absolutely smoked as well.
[18:11] So it's kind of been like two to three quarters back to back that they've been absolutely killing it.
[18:13] I can't avoid the Trump story like uh he praised Dell
[18:18] Technically and uh now the stock's gone up.
[18:20] Like I don't know what you think.
[18:22] That's what I was going to say.
[18:23] I'm like, "Hey, you know who called Dell stock going crazy back in February 10th of this year?"
[18:27] Um, Donald Trump.
[18:30] He said, "Um, go buy a Dell."
[18:32] And the close at that time was, I think, $126.
[18:34] We're sitting here today recording this.
[18:37] It is now $450.
[18:40] That is a 255% gain or even more.
[18:42] That's close to like 300% now at this point.
[18:44] And that's coming off the back of a week that happened last week where they were up 37% in a single trading session.
[18:49] That's the largest single win that they've ever had in history in their 40-year history since they've been trading publicly for 30 years.
[18:57] And it'd be interesting if it was a one-off, but this is very much a trend.
[19:00] Another story that we have here today is the newest version of the Trump pump with IBM where he mentioned IBM over the weekend and on a Sunday night it traded up 17%.
[19:09] Because Trump mentioned IBM and he didn't even say go buy IBM.
[19:15] He was just praising the CEO.
[19:19] publicly on camera and the stock is going ballistic.
[19:23] So it feels like we do have to mention that this is a trend.
[19:25] There are a few other companies you might remember Intel being a major one where the government took a 10% stake in Intel since then it's up 300%.
[19:35] And this seems to be a trend over and over again.
[19:38] So, I think a lot of people who have been interested in participating in AI, who have been seeing these Donald Trump clips, are like, "Uh, all right.
[19:45] Well, he's saying it, but the stock's already up so much, it's probably not going up any higher.
[19:49] The reality is that anytime he's called something, it's gone nuclear.
[19:53] And I guess that's part of the cycle that we're in where like you're the president of the United States, you say something, the market reacts.
[19:59] It's it's similar to like a memecoin.
[20:03] Like it feels like we're watching meme coins pump.
[20:04] It's like you watch for what one guy says and then the stock goes ballistic.
[20:08] Like you mentioned, there are numbers to back this up.
[20:09] Like Intel's and IBM have actually been having pretty amazing quarters.
[20:12] Same with Dell.
[20:14] But there is some frothiness happening here in the sense that one president can deliver one line to send a stock up 20%.
[20:18] Yeah.
[20:20] this weird cloudy murky investment
[20:23] thesis that kind of like involves being
[20:26] crowned by Trump himself, but all then
[20:28] also like the numbers and the customers
[20:29] and the purchases are from very
[20:30] legitimate people and companies. It
[20:33] reminds me a lot of the Intel story
[20:35] where I think it was like a year and a
[20:36] half ago Trump was like Intel is going
[20:37] to be an amazing company very key to
[20:40] accelerating AI uh in the US and that
[20:42] has been uh an amazing trade so far
[20:45] right it's I think the stock is up
[20:47] around 3 to 400% uh since he made that
[20:49] announcement. Nvidia purchased $5
[20:51] billion worth of equity. The Trump
[20:53] administration bought 10% of Intel which
[20:56] is now up $30 billion as of today's uh
[21:00] stock price. Uh, and so I I start to
[21:02] think about like Trump being pretty key
[21:05] to forming these partnerships. Um, I bet
[21:07] you because he told Nvidia or asked
[21:09] Nvidia to purchase that kind of equity
[21:11] stake in Intel, he's kind of doing
[21:13] similar things with Dell. And if you're
[21:15] wondering like why this is the case, um,
[21:17] the answer is pretty simple. Trump or
[21:20] America in general does not want to rely
[21:23] on uh, Taiwan or uh, Asia in general for
[21:26] building the most foundational
[21:28] technology that the world will ever see.
[21:30] And as a result, they're trying to bring
[21:32] as much of the manufacturing process and
[21:34] hardware process of AI onshore to the
[21:37] US. Now, China uh and Asia has had such
[21:40] a major head start. Taiwan is a very
[21:42] crucial piece of land that China could
[21:44] take over at any moment. It's kind of
[21:46] obvious why America wants to bring it on
[21:48] shore, but that's going to take a lot of
[21:49] time. And so, Trump is obviously
[21:51] aggressively trying to form these
[21:52] partnerships and probably get a stock
[21:54] win in the meantime. I I read somewhere
[21:57] that apparently like Trump had bought a
[22:00] position in Dell on his personal account
[22:02] before he made that public announcement
[22:04] which is just absolutely insane uh and
[22:07] probably not legal but um you know I I
[22:09] could kind of see like why the
[22:10] administration is going down that path.
[22:12] Yeah, say what you will, but they are
[22:13] making an aggressive effort to bring
[22:15] this AI trend, this AI world onto the
[22:18] United States soil to build everything
[22:20] domestically. And what we're seeing here
[22:22] is like very clear signal that it's
[22:23] going to continue down this road. And I
[22:25] think what we're seeing with Dell is a
[22:27] testament to what happens when a company
[22:29] run by a founder really leans into the
[22:32] things that they've learned over 40
[22:33] years and applies that to an industry
[22:34] that really needs help. We have this
[22:36] problem manufacturing things in the
[22:38] United States. We have a very difficult
[22:39] time building things at scale, creating
[22:41] factories that have efficient margins
[22:43] enough to compete with people who are in
[22:45] other countries and companies that are
[22:47] run like ASML and TSMC and all these
[22:49] other ones that are creating all the AI
[22:51] infrastructure. So, it's exciting to see
[22:52] a company like Dell really move their
[22:54] whole weight of the company behind this
[22:56] and actually deliver some pretty
[22:57] compelling results. Like they're the
[22:59] first ones to deploy a Vera GPU rack.
[23:01] That's a pretty big deal. We started
[23:03] talking about Mythos and a few of the
[23:04] other like cutting edge uh LLMs and how
[23:07] they were just getting started training
[23:09] on Blackwell. Like we're getting our
[23:10] first Blackwell chips and now suddenly
[23:12] we have Vera Rubin already going into
[23:14] play. Training I'm sure is going to
[23:16] begin sooner than some people would have
[23:18] thought because of this accelerated
[23:20] timeline. It's really great to see Dell
[23:22] is the the builder and we need more
[23:24] builders. So I think on like a
[23:26] principled basis, Dell's a good company
[23:28] for a good reason. It's not just because
[23:30] Trump is out here saying it. uh they're
[23:32] actually making some pretty amazing
[23:33] things and I think it's it's a testament
[23:35] to other companies who are kind of
[23:37] evaluing how to navigate this world of
[23:39] AI. Um being useful in the world of
[23:41] hardware is a really compelling value
[23:43] prop that I hope more people follow.
[23:45] >> It it looks like investors probably
[23:47] agree with this as well. Uh since we
[23:49] started recording the stock is up 8% uh
[23:52] roughly 30 bucks since market open. It's
[23:54] literally been what is it under an hour
[23:56] of market being open. Um, yeah. I guess
[23:58] to to round it out, like the question I
[24:00] want to ask myself is, is this company
[24:03] in particular kind of bubblew worthy? I
[24:06] I don't have a definitive answer, but
[24:07] it's kind of like yes or and no. Yes,
[24:09] because they kind of fell into this
[24:11] opportunity, honestly, similar in the
[24:13] way that Nvidia kind of did that.
[24:15] They've been building these kind of GPUs
[24:16] for the gaming industry and a lot of
[24:18] other industries prior to falling into
[24:20] the AI stuff. Now, now, credit to
[24:22] Jensen. Obviously he saw the trend of
[24:24] LLMs uh much early on before Chad GBT
[24:27] went viral. I don't think I can give the
[24:29] same kind of credit to uh Dell
[24:31] themselves but they have had 40 years
[24:34] expertise in building these hardware
[24:36] racks and what is the major bottleneck
[24:38] that we've covered on previous episodes
[24:40] the Gavin Baker episode and the AI
[24:42] hardware stack episode which you guys
[24:44] should definitely check out. we released
[24:45] them over the last 2 weeks. Um is the
[24:47] money is flowing from the chipmaker from
[24:49] the consensus bet which is Nvidia down
[24:51] to all the constraints that are
[24:52] preventing those GPUs from actually
[24:54] being powered on today. Electricity um
[24:57] chemical substrates as well as the
[24:59] cooling and data center racks themselves
[25:01] which is what Dell is playing in. So I
[25:03] think there's a a huge opportunity for
[25:05] uh investors that are looking for like
[25:07] that kind of like next bit of exposure
[25:09] to kind of like work their way down the
[25:10] stack. And obviously this isn't trading
[25:12] or investment advice, but it's just a
[25:14] trend that we're noticing quite a lot
[25:15] of.
[25:16] >> Yeah. And like what an amazing
[25:17] opportunity to be an investor to see all
[25:19] of this. There's so many companies that
[25:20] are undervalued, that are misunderstood,
[25:22] that are kind of pivoting
[25:24] >> that are public. Yeah. There's so much
[25:26] alpha to be made in these markets. And
[25:27] it's funny that the president is the one
[25:29] who's surfacing a lot of these, but
[25:31] there actually is a tremendous
[25:32] opportunity for anyone who's curious in
[25:34] investing or participating in the AI
[25:36] stack because so many of these companies
[25:37] that we're seeing exploding, they
[25:39] they've been around. They're not that
[25:41] big. They're just pivoting and they're
[25:42] putting themselves in the right place
[25:43] because they understand the business.
[25:45] There's a lot of those hot, flashy,
[25:46] shiny companies, OpenAI, Anthropic,
[25:48] Google, like the ones that we talk about
[25:49] every day. But again, that AI investing
[25:52] stack, as you descend lower into the
[25:54] stack, there's a tremendous amount of
[25:55] opportunities. Dell is one of them. And
[25:57] as we see more people begin to pivot
[25:59] into the places where there are those
[26:02] bottlenecks, I assume we're going to see
[26:03] a lot more of these. So, that's the
[26:05] episode. you're caught up with a nice
[26:07] little teaser of the new uh Nvidia
[26:08] hardware that was just announced today.
[26:10] I think um that's everything. Any final
[26:12] thoughts before we uh we let everyone go
[26:14] here?
[26:14] >> I have I always like to end the episode
[26:16] on a prompt and I I I have one in my
[26:19] head right now, which is are you guys
[26:22] going out to buy the new uh DGX desktop
[26:25] for 80 grand, 80 to 100 grand, or are
[26:27] you buying the RTX 4090? I'm genuinely
[26:30] curious because like I know that a bunch
[26:32] of our listenership likes to kind of
[26:34] like be hobbyists in AI and run their
[26:37] own open source models. We've got a lot
[26:38] of feedback from you guys in pre from
[26:40] previous episodes. I wonder if you guys
[26:42] are going to get indulge in one of these
[26:43] new laptops, one of these new consumer
[26:45] powered GPUs that you can kind of run
[26:47] models locally and if you are that type
[26:49] of person, what are you using it for?
[26:51] Like we want to learn as much as we can
[26:52] around the locally run air models. We
[26:54] think it's going to be a huge thing in
[26:55] the future. So yeah, let us know. But
[26:57] >> I love how how much of a baller you
[26:59] think our audience is. Like they're just
[27:01] going to go drop 80K on a desktop GPU.
[27:03] That's how they think of you.
[27:05] >> Yeah. I don't know.
[27:06] >> But I think for like the laptop at
[27:08] least, as I look at the laptop, I
[27:10] compared apples to oranges. It's
[27:11] compelling in the sense that I can play
[27:13] video games on it. And I think that's
[27:14] one thing that gamers are really going
[27:16] to appreciate is the fact that there is
[27:18] now a portable gaming machine that is
[27:20] really excellent at games. Because my
[27:22] MacBook, I'm a diehard Apple fan. You
[27:24] know this. I have all Apple products. I
[27:26] can't play games on that. I still got to
[27:28] go play on a PC or on a console. So, I
[27:30] think in terms of that niche audience,
[27:31] that's a really compelling thing. So,
[27:33] any gamers out there, are you buying
[27:34] this thing? A 5090 in a laptop is crazy
[27:37] work. I'm excited about it. I might be
[27:39] buying. We got to wait on the price. If
[27:40] it's anywhere close to $80,000, I will
[27:42] be a no, but I suspect it'll probably be
[27:44] in MacBook territory. Um, but yeah,
[27:45] that's the episode. Thank you guys so
[27:46] much for watching. If you enjoyed this,
[27:48] don't forget to share it with your
[27:48] friend. It really goes a long way if
[27:50] we've been doing well. Rate us five
[27:51] stars on your favorite podcast player if
[27:53] you enjoyed. And like always, we will
[27:54] see you guys in the next episode. Thank
[27:55] you so much for watching.
[27:56] >> See you guys.