We're sunsetting PodQuest on 2025-07-28. Thank you for your support!
Export Podcast Subscriptions
cover of episode Our DeepSeek stock trades — winners and losers

Our DeepSeek stock trades — winners and losers

2025/1/28
logo of podcast Dumb Money Live

Dumb Money Live

AI Deep Dive AI Chapters Transcript
People
C
Chris
投资分析师和顾问,专注于小盘价值基金的比较和分析。
D
Dave
活跃的房地产投资者和分析师,专注于房地产市场预测和投资策略。
J
Jordan
一位在摄影技术和设备方面有深入了解的播客主持人和摄影专家。
主持人
专注于电动车和能源领域的播客主持人和内容创作者。
Topics
主持人: 本期节目讨论DeepSeek事件对AI行业的影响,以及由此引发的投资策略调整。DeepSeek是一款类似ChatGPT的低成本AI应用,其出现对NVIDIA等公司造成冲击,引发了对AI战略的重新思考。我们需要关注计算成本效益,并对未来AI发展趋势进行预测。 Dave: 我在DeepSeek事件发生后立即卖出了所有NVIDIA股票,但计划未来重新买入。我认为微软和Meta的财报可能会影响NVIDIA的股价。投资者需要关注AI公司未来几年的资本支出计划,以及他们对DeepSeek事件的回应。 Chris: 我认为Meta、亚马逊和苹果公司将从DeepSeek事件中受益,而Google和NVIDIA则面临风险。短期内市场对DeepSeek事件的共识将主导市场走势。长期来看,我仍然看好NVIDIA,但短期内对其持谨慎态度。我考虑买入NVIDIA期权来对冲风险。 Jordan: DeepSeek的技术成就令人惊叹,这与高效的团队和创造性的工作环境有关。美国科技公司过于依赖资金和基础设施,而忽略了效率。中国公司在AI领域取得的成就,促使美国公司重新思考其战略。AI即将民主化,这将对现有行业格局产生重大影响。DeepSeek对CUDA等NVIDIA技术构成挑战,其研究人员对NVIDIA的技术提出了改进建议。DeepSeek吸引了来自NVIDIA的优秀人才,这挑战了美国在科技领域的优势地位。DeepSeek事件对依赖稀缺资源的投资策略产生影响,短期内可能导致NVIDIA出现波动,但长期来看NVIDIA仍可能获胜。我们需要关注AI公司未来几个月的表态来判断DeepSeek的影响。

Deep Dive

Shownotes Transcript

Translations:
中文

You've heard of cheap Chinese knockoffs, right? Well, that's pretty much what we have with DeepSeek, the chat GPT style app that just hit number one in the app store and sent the stock market into freefall yesterday. Except this cheap Chinese knockoff might have actually improved on the original. DeepSeek claims they only spent $5.6 million compared to the $78 million that OpenAI spent on chat GPT or the high-end

$191 million that Google spent on Gemini. Those numbers were widely reported over the weekend, along with news that DeepSeek runs on older, slower, and cheaper GPU processors, making yesterday NVIDIA's worst day ever. They lost $600 billion in market cap in a day. They basically lost an entire Netflix worth of value.

But we're not here to debate the merits of the deep seek technology. We're here to make money and decide if we need to rethink the entire AI strategy. Last week, it was cool to spend as much money as possible on processing power, as much as NVIDIA would give you. There was very little downside to overspending. Google and Microsoft were on a roll. And now this week,

Getting more bang for your compute buck is cool. So who are the winners and losers? Today on Dumb Money, we're going to reveal our stock trades in the wake of Deep Seek. This episode is brought to you by Nerds Gummy Clusters, the sweet treat that always elevates the vibe. With a sweet gummy surrounded with tangy, crunchy Nerds, every bite of Nerds Gummy Clusters brings you a whole new world of flavor. Whether it's game night, on the way to a concert, or kicking back with your crew, unleash your senses with Nerds Gummy Clusters.

You know what's smart? Enjoying a fresh gourmet meal at home that you didn't have to cook. Meet Factor, your loophole in the laws of mealtime. Chef-crafted meals delivered with a tap, ready in just two minutes. You know what's even smarter? Treating yourself without cheating your goals. Factor is dietician-approved, chef-prepared, and you-plated. Pretty smart, huh? Refresh your routine and eat smart with Factor. Learn more at factormeals.com.

This is Dumb Money Live.

Hey there, Dave here along with Chris and Jordan. We are Dumb Money. Welcome to Dumb Money Live. We are already seeing the debunkers, including Elon Musk coming out and saying that DeepSeek is actually way more than 6 million bucks in on training their models, that they used way more powerful GPUs. We've heard from Sam Altman calling it impressive. President Trump calls it a wake-up call. But before we get into the trades, we need to have a little wake-up call of our own. Take a second, if you will, and smash the like button. Help us wake up the almighty algorithm, Chris Jordan.com.

A couple more quotes here. NVIDIA says that DeepSeek is an excellent AI advancement and reminded us yesterday that interference requires a significant number of NVIDIA GPUs and high-performance networking. Microsoft said that as AI gets more efficient and accessible, we'll see its use skyrocket. Chris, you posted on X yesterday and got a million views. Let me pull that up because that's, I think, a good place to start. Where's the...

I got a million views on my run-on sentence. Well, you basically were linking deep seek to the $7 billion hedge fund that is behind it and questioned whether they might have shorted Nvidia and whether that might've played into their, their whole emphasis on how little they spent on it. So,

Let's start there. Dave, don't you think it would be criminal for them not to be shorting NVIDIA? I don't see anything illegal. It would be criminal for them to be shorting, but they're Chinese and the laws don't apply. Okay, so just think about this. This company that owns DeepSeek, they have the DNA of a hedge fund. Why? Because they are a hedge fund. They're a $7 billion hedge fund. The reason that they had so many advanced AI people is they were basically using AI to...

to fast trade for their hedge fund clients. Yes. So they are about to release on the world, maybe the biggest news that technology has come across in years. They know this is going to destroy NVIDIA in the short term. Imagine them not shorting NVIDIA as a $7 billion hedge fund. I think your hypothesis is spot on.

I wish I could just see what they did. I'm curious. How many billions did they make? I want to see the AUM of this hedge fund in six months. Yeah. Before and after. Just the overnight before and after would have been very interesting. Because how many billions did they have under management? How could they not short it? It was $7 billion before. I bet it's $15 billion now. Yeah.

I mean, by the way, they're not manipulating everything that they're just putting out something in the market that they know would negatively impact a company, period. That's all in the short run, or at least the perception of that would be negative for a specific company. So they have every right. Forget about the whole China, America stuff. They have every right to play the market based on something they created, right?

That's basically what those short reports do. That's their entire business model. Yeah. Putting out information about why they're short something after they're short something.

Yeah, it's wild. So I can't. So NVIDIA, clearly a loser. We're talking winners and losers today. And we had eight little logos in our thumbnail. NVIDIA, Meta, Google, Amazon, some of the normal ones. I also want to add Apple and Microsoft to our discussion today.

Fine, but we can't just roll right into the winners and losers. This is just too important. This is too big. Plus, we don't have our 100 likes yet. Remember when we used to hold the show hostage until we had 100 likes? It's not about that. We've got to discuss the narrative here. I'm pretty sure I've read every intelligent take.

on the entire planet, okay, over the past four days on how this deep seek news would impact the AI sector and NVIDIA. And guys, my conclusion is that there's absolutely zero consensus. I'm talking about the smartest people in the world, people that are AI scientists, the companies themselves, like debating this

in papers debating this. Guys, I was up both

Well, 3.30 in the morning last night, I spent five and a half hours reading every last tweet from anyone with any sort of AI background. And I'm telling you, there's no consensus here. There's no consensus. And the rumor is that all of these companies are putting together a task force to try to figure out how to basically accomplish the same thing. I know for sure, well, at least I read a report that Meta is doing that. And why would you do that?

Jordan, I don't mean – I agree with you. There's consensus on what they're doing to go forward, right, to improve. What there is no consensus on –

I'm talking about financially. There's no consensus around whether this massive infrastructure moat that has been created, that quite honestly has been driving a lot of the valuation in some of these companies, including NVIDIA, obviously,

the degree to which that is going to break down at least temporarily. There's no concern because we have... Yeah, I think there's two things you have to look at though, right? And you have to look at what are these guys actually doing and what does that mean for NVIDIA in the future?

Right. Can we agree on one thing that right now, at least in the near term future, NVIDIA's valuation has been based on AI having scaling issues? Yeah. So it's based on AI, but really what it's based on is CapEx. Right. So these capital expenditures by huge companies and that that that capital expenditure growth has to keep on. And I think that that's what's in question right now.

Exactly. Because if you can't determine with conviction that the money you are spending on that infrastructure is going to have a positive ROI, then you are going to be

more cautious about spending money beyond what you've already committed until you can determine that the ROI is actually there and that could create a massive issue forward at least temporarily for companies like Nvidia right because yeah and they might get the spend that's already committed but are they going to get another 150 billion on top of that the following year right

That's the question, right? I think it all boils down to what is... And there's no real answers yet. I don't think even Meta or Amazon or any of these guys know what their capex spend is going to be going forward based on these results. No, I totally agree. Also, let's just talk about DeepSeek for a moment. First of all, what they did was just...

objectively incredible. And I think this has a lot to do with the fact that, you know, this guys, when you put people into a situation where they're forced to be creative, that that's where the best work comes out throughout all of mankind, throughout all of history, right? This is what they used to do with artists back in ancient times.

ancient days, right? Like they would literally have them working 19 hours a day under extreme pressure and they would create some of the best work we've ever seen in the art world. Like this is, we know this is true in America. Could we just say it? Like we have become lazy when the average developer, like senior developer is either a millionaire or a deca millionaire that has to have some impact.

Every employee at NVIDIA is a millionaire. It has to have some impact. I mean, you could only buy so much with money. Listen, we make up for it at scale. We make up for it by having more of them. We make up for it with money and infrastructure. But I think that our whole going into this, we knew that we had access to these advanced chips. And so we load up on it, knowing that what we're going to be able to develop

will eventually be able to be streamlined and made more efficient, but we didn't have a need to sit and focus and make it streamlined. We're just going to throw more processing at it and see how fast we can advance the technology. And I think that that's kind of like, you have two things going on. You have China that was forced to get creative and

figured out a way to make it more efficient. We would have eventually gotten there, but we didn't need to because we could just keep spending and throwing technology and processing power and GPUs at it. I think what we're looking at, and I think there is a lot of consensus around this, is that AI is about to democratize, right? So if we go back, I don't know, to the mainframe days, right? You know, IBM was building out these insane mainframes

like perfecting that entire industry sector right as computing was about to be democratized, right? And like they did not see that coming, right? And you have companies, you know, going back to the 90s like Google and what Google did, I think it was to old...

to Oracle back then. Google basically is a very similar situation, right? Where you have Google basically democratizing that type of compute in a way that nobody thought was possible back then. And it just radically changed the entire sector for the better, by the way. And it did. This is where the whole Jevons paradox that everyone's talking about does come into play. It did ultimately impact some companies negatively,

but for the most part, it created an industry that was a hundred X the size. And basically democratizing AI and making it a commodity is what Palantir has been saying would happen all along. It's what you do with the technology, not so much, you know, developing the actual chips and, and developing the AI. So I think that,

There's a lot of companies that will be the winners and losers, and we'll get into all of that. And I think we have to, like, we can focus on, like, what is it, 95% savings? But I think what's even more important is, like, what happens to this massive innovation cycle that we're about to get into when all these cost curves completely change?

And because like it's, you know, you could focus on, okay, I've heard a lot of people saying, well, how is deep seek going to make money? Right. At such, you know, cause it's, it's just so, so cheap now, right. This democratization of what they've essentially created. Other people are saying the more important question is,

How is open AI going to make money? And if open AI doesn't have a solid plan to monetize going forward because of this democratization, that's where the cost curves can really impact

the money flow that ultimately is what Nvidia has been relying on because what, 90 plus percent of the spend has been on these massive pre-training models that appear to be less valuable going forward than they were a week ago. - Yeah, but are they, right? Because at least if we read some of the articles that I know you shared one, I've read a few others.

that believed that basically DeepSeq was using OpenAI to help train their models, right? And so in that case, can you really do this for $5 million? Or were they able to take advantage of a mispriced situation where they were able to use the API of OpenAI to get

But that ultimately doesn't matter if OpenAI can't make enough money. Unless OpenAI puts that down and you can no longer do this type of thing anymore. Well, that's what I'm saying. So it's a reset. It still took OpenAI to make the investment to get where they got for DeepSeek to be able to get where they went.

But it still doesn't matter because if open AI can't generate a return, then because someone else is able to do that and then democratize the money that they're spending and there's no moat around that, then they won't be able to continue to raise money. And by the way, this is like,

this is nvidia's biggest customer pretty much more or less right like they won't be able to continue to raise that capital so that that's part of the democratization process it gets broken down and then gets rebuilt in new ways right we have to figure out how to rebuild it in new ways with completely different economics here's the other demonstration that i'm concerned about in this whole thing and this leads directly to nvidia and that's that you know these um these

These researchers, the people that put together these models for DeepSeq, in their clusters, they did not use, regardless of chipset, maybe they used the H800s like they claim. Maybe they had some H100s. I have no idea. I have no insight into that. But the claim is that they did not use CUDA and that they basically went down to the metal, right, of these things and made some optimizations there.

Does that lead to the moat falling? That's one of the big moats that we said that NVIDIA had was this CUDA. Or we've just gotten more efficient. Imagine what we can do with more efficient code, throwing it at even more GPUs. Imagine how much faster we'll be able to advance. That's not the question, right? So I agree. First of all, I agree. We've never had enough compute ever to do what we wanted to do.

this moat that we thought that it was for NVIDIA, or does this open the door for competition in the GPU space now that we know that the software

like kuda dave part of the reason why they circumvented cuda is because they couldn't actually drive the type of innovation with the modeling they were doing because cuda was restricting them so they had to break it they had to do it themselves in fact part of this paper that they wrote

was instructing, this is wild, they were instructing NVIDIA on all the things that they think NVIDIA should do going forward to actually improve, like add to the acceleration of what people can do on their chips that NVIDIA was doing wrong. They're actually providing directions to NVIDIA for how to do things better. And quite honestly, I don't think anyone's been able to debate the fact that they are right.

They were right. They were better. They are sharper. Here's what's really interesting. I actually saw I think he was the third employee at Deep Seek, a very young kid from China.

worked for NVIDIA and NVIDIA was trying to put together a package for this kid to actually, you know, stay at NVIDIA. And he ended up taking instead a job at deep seek when they were a minuscule company. And so the brain power, most of us don't understand this. A lot of our brain power here in the U S, uh,

are people that are coming out of China, right? And for the most part, we've been able to retain a lot of those people. But this deep seek moment is big. You know, there are a lot of people thinking right now,

Should they just keep working for Zuck or go home to China to work for what appears to be one of the smartest groups in all of technology, one of the most bleeding edge tech groups in the world? OK, now at this entity, DeepSeek and potentially at other entities across China, that's not really being discussed. But it's an important issue, I think, going forward, maybe not an investable issue.

but I find it something that's really fascinating. Just the level of talent that we just, we brush off like, oh, you know, it's China. They don't really, they don't do things right. Well, no. This is the first like very public thing that we've seen come out of China that is making us rethink our superiority in technology. But Dave, again, a lot of our people here working at our companies are

are from China. This kid was from China, but working at Nvidia, but now he's a deep seek and he's part of the narrative of what they created at deep seek. I just, listen, it has nothing to do with investing. I find it fascinating though. Okay. So to what degree I think are we going to punish

scarcity investments here right like that's the big I I think we have to start with Nvidia we can talk there's a lot of invest there are a lot of trades here I know there's a lot of trades but to what degree Jordan does is the scarcity trade

just less interesting than it was a week ago. Not to say that there aren't other things that NVIDIA will do to make up for it. And then as the industry scales up to be meaningfully larger because of this democratization, at the end of the day, maybe NVIDIA still wins. Maybe they actually win bigger at the end of the day. But as traders,

Do we experience some meaningful amount of volatility for the next few weeks, next few months, maybe the next year, while we have to rebuild this industry from the ground up in a totally different way that potentially destroys the existing business model at NVIDIA? I'm just throwing that out as a thesis. Well, that's the thing. I don't think we have answers to those questions. And I think that we have to pay attention to what

you know, some of the companies that utilize these chips have to say over the next few months. I think it'll take them at least a month to figure out the answer to those questions themselves. I agree. Good news is we have earnings like just right around the corner. This is what makes us so great, guys. The fact that we have earnings this week. Yeah, we have. Microsoft is tomorrow, right? And then Meta and then

I thought Meta is tomorrow. Is it not tomorrow? Tesla is tomorrow, right? I thought Meta and Microsoft were tomorrow. I saw Tesla was Thursday, but then I saw... Microsoft, Meta, and Tesla are all on Wednesday. Apple is Thursday. Google is next week. And NVIDIA is not for four more weeks. This is nuts. But tomorrow, Microsoft, Meta, and Tesla all have...

AI, they all have a lot of Nvidia processors. They all have a big CapEx budget and they all will be asked. The very first question out of an analyst mouth is going to be, what are you doing? What's next? What happens to CapEx spend?

I think that is one of the most critical things for investors to try to get in front of this week. So basically, we're in a situation where I'm just going to tell you straight up with high conviction that there's no consensus and nobody knows. I'm going to tell you that NVIDIA doesn't know because you could read between the lines of their own statements.

They are putting out statements that are very broad. If they actually had any type of conviction that the transition over to inference, the transition over to in time, what are they calling it? Like in time testing, which is essentially training on the go, how that is going to

potentially make up or maybe even be better than anything that they might lose from pre-training, which is where the bulk of the spending is going right now. If anybody knew that, they would say it right now with conviction and no one's saying it with conviction. And that tells me that nobody knows anything this week. So basically, what are they going to say? One open source model drops in out of nowhere. Nobody saw this coming.

And it disrupts the entire industry. All right. I'm going to tell you my first trade because we're talking about NVIDIA. Did you? Okay. So first of all, you made trades yesterday. Was this the very first trade you made?

Very first trade I made at 6 a.m. in the morning, the second that the pre-market opened up, I did two things. I woke up, I put on my reading glasses, because I could barely see, especially in the morning, and I sent out that tweet that went crazy viral about the hedge fund shorting NVIDIA from China. And then I went on and I did something that I wished I would have done on Friday, which

i had a pit in my stomach on friday night i was traveling that day i was coming home from a lot of robot work uh and i knew within an hour of getting into this deep seek mess i knew that nvidia was going to be destroyed on monday and i was like i i just the post market had closed it was too late for me to sell my nvidia

And when I saw Robinhood on Sunday night and Nvidia was only down like four or five bucks, I was stunned. I was like, if it stays down four or five bucks, this is my dream. To sell it only down four or five bucks is insane. What are they not seeing on Robinhood? They were so misaligned with reality on Sunday night. And sure enough, I just watched the overnight markets. And by the time I woke up,

They were slaughtered. But I did sell 100% of my NVIDIA at 6 a.m. on Monday morning. 100% of my NVIDIA. Now,

That said, my intention was to get back into NVIDIA shortly after. Okay. So I haven't gotten back in yet. Where is it at right now? It's up. It bounced 5% today, I think. Yeah. It's up to 126. It's up nicely. It's probably roughly where I sold it right now. I don't.

i am honestly i don't know i don't i don't have the answer here uh as to what i want to do with my nvidia i just know that i have to own nvidia long term yeah uh i don't want to own as much nvidia as i own before

I don't. I want to wait it out. I didn't make the same knee-jerk reaction and sell all of my NVIDIA. I did not sell the stuff in my taxable accounts, but I did quickly, immediately, in the overnight trading, sell it in my IRA.

Okay. Yeah, I did that as well. I sold it across every one of my accounts. But again, I intend to get back into NVIDIA. I am happy to pay up for NVIDIA and getting a little bit late. I don't necessarily see this as NVIDIA jumping, you know, 40% here in the next week. I could be wrong and I'll be pissed off if they do. It's up 6% today.

Yeah, I mean, it's coming back after yesterday. I think that there are more questions and answers in the short term on NVIDIA. Now, I might buy some NVIDIA options to protect myself through the rest of the week, right? So if it does make a run on Microsoft and meta earnings based on statements from those CEOs, and I think that's why it's trading up here. I think most of the consensus here is that these companies are going to

come out and protect themselves. They're going to protect the decisions they make. And in doing that, they are ultimately going to protect NVIDIA, right? Because they don't want to look dumb. And it's going, what they would say- Well, I think all they can say so far is that they haven't changed their, you know, CapEx plans, right? And I think that that would be reassuring enough.

that would be somewhat reassuring, but I don't think that's what people care about, Jordan. No one cares about- - I do. I mean, that's literally all that matters for Nvidia is what the capital expenditure is by these AI firms. - No, no, no, no, no, no. You're not going deep enough.

Nope, nope. It's not just that. It's not just what their CapEx has for this next year. It's how they're thinking about CapEx going beyond that. Remember that. That's what I'm saying. So whatever their plans were, if they haven't changed those plans yet, and they're still planning to invest...

But Jordan, an analyst is not going to ask about that because they'll probably already say that. An analyst is going to ask beyond this, how does this impact your future thinking over the next few years? And then they're going to read, well, they're probably not this smart, but if they were, they're going to read between the lines as to how

that CEO answers that question, do they hesitate? Are they ambiguous? Are they cautious about continuing this type of spend? Or are they like, well, we're just going to have to see how this plays out? Well, we know for sure that Meta is already investigating, right? So maybe they say that, but we know that all these companies are putting teams on it to try to figure it out.

Meta has been all about their open source version. They're the one I think that might benefit the most from seeing these kind of cheaper versions of inference training that they'll be able to implement and use in their models, in their LAMA for integrating across all of their platforms, right? Yeah.

Agreed. Agreed. I think Meta is not a stock. I kick myself for not being in Meta, but I feel like Meta is a benefactor.

So I'll kind of tell you the consensus. So I don't know if you guys have read the Ben Thompson kind of report that came out. And that's the report that I think a lot of people are going off of. He claimed that Meta would be the largest winner because it has the most consumer touch points for AI and can serve all users cheaper.

Okay, so that's the thesis there on meta. And by the way, now that this thing is running, I'm like simultaneously going in to realize I should probably pick up some of those safety options on NVIDIA. So I'm just going to read you what his thoughts were on these stocks. He thinks Microsoft is going to win.

Microsoft would theoretically win because they're spending less money on data centers, GPUs, but the customer demand would still grow overall.

In a similar way, Amazon would be a huge winner because it hasn't been able to make its own AI model, so it hasn't even spent any money. But now there are so many high quality ones it can access or its customers can access at a really low cost.

that still gets served through AWS. So think about this, guys. This is just like an AWS dream. It's like they can just load up another new thing and just put it in there for free.

This is Amazon's dream. Amazon wanted there to be tons and tons of models that anybody could access through AWS so that Microsoft wouldn't have a chokehold because they own the one model, right? Open AI, and you're forced to go through Microsoft Cloud.

Amazon is like, hey, you can get all the models through Amazon. There's tons and tons of models. Amazon still makes their money and they don't have to invest one penny to secure an exclusive with this one bleeding edge model. This is such a win for Amazon. I'm going to tell you right now, I think Amazon might be my takeaway winner from all this democratization of AI.

I think they are my number one winner. I agree that Meta does have the most consumer touch points. And so, yes, they would theoretically win there. But who has the most commercial touch points? Amazon. And who has the most margin on sales?

taking something that was developed for free, running it through their data cloud and letting, letting, letting the meter run while people use it. Yes. Taking something for free and monetizing it is an AWS specialty. So Amazon and meta were my top two winners in this winner and loser game. But Dave, you're missing the one that everybody has been talking about. You're missing one that everyone's been talking about. Right. And this, this is a big one. Okay. So, uh,

Who is the company that basically said, we'll spend even less than Amazon on this stuff? They went all in on edge inference. They went all in on the type of edge inference where you could basically take all of this AI and actually deploy it on a personal basis. Apple.

Apple has to be one of the biggest winners, and that's why Apple stock is doing what it's doing. Before I hit the share button, I pull up the stock that I think you're talking about. It's like a game I play with myself, and Apple is the one I picked. And that is my third, because they are optimized for edge inference, and they are the consumer touchpoint. And this lets Apple intelligence, which their whole dream is having something that can just navigate.

natively run on your iPhone without having to talk to the cloud. Now they can have one less thing that has to go up to what are they calling it? The secure Apple cloud or whatever.

They're going to have to give us a little bit more RAM on these phones. Okay. They already did. The current version has more RAM across the line, and that's going to be just their new thing. They're going to bake in slightly more RAM across everything. They've done it on iPad, the new iPad releases. They got rid of the lowest tier and added a higher tier across the product line, and they'll continue to do that.

Totally agree. So the company, can we also agree that the company that has the most uncertainty in the short term, okay, because they're two big moats, CUDA is one of them, Jordan said it, CUDA is at risk, and their other moat, which is networking multiple GPUs,

That's also at risk. OK, because that kind of flies in the face with the way that deep seek has kind of re architected how they're going about this modeling. So if those are their two big moats.

They're both being challenged right now. And it's undebatable. We just don't know how it all plays out. And there's uncertainty with NVIDIA. Do I believe NVIDIA wins long term? Yeah, absolutely. Long term, I believe they're going to figure it out and win. But I'm just a little nervous to have NVIDIA be my number one position. So they went from my number one position to not winning any.

uh and i'll by the way i will land somewhere in the middle i'll land somewhere in the middle uh of the two so you know who else people are really by the way this is also really important we can discuss who the winners and losers actually are but kind of like the what the presidential trade for the next few days and the next few weeks it's really important it's more important

to assess who the market thinks the winners and losers are going to be than to actually pick the winners and losers. And that's why I'm spending so much time actually reading you what is the consensus take, because all the stocks that are moving today are the consensus take. Now, I actually kind of somewhat agree with the consensus take on a lot of these, but

you have the consensus is going to move the market in the short run. Now, if they're wrong, the consensus is going to kind of move around too. So it's not like the, what we say today may not be true tomorrow. It may not be true a week from now. This is just like a snapshot moment in time of initial market reaction to something that has kind of just blown everyone's minds.

That's exactly correct. And while this stock is up a little bit today, they're kind of on the rocks right now, and they're kind of getting beat up a little bit. Google, you know, they've specialized so much in this TPU hardware, which I guess is not really seen as a beneficiary.

of what's happened the last week with deep seek. I don't really know enough about the underlying drive down the cost and increase the availability of these things and you decrease the value of search. And that's the that's the rub on alphabet right now. Yeah. So we may talk about this for a long time that what 90% of the revenue is still search. And guys like, you know, Dave, let me let me call you out for a second.

Yes, please do. When this whole ChatGPT first came out, one of the first shows we did, I came out and said with conviction that...

that at some point in the not too distant future now there are still people that read newspapers we know that and there's still people on yahoo your old company dave but i said there would be a massive massive transition and it wouldn't take that long for people to get off of search and search would be completely democratized in a way that we can't even comprehend due to ai and you didn't see it at the time you're like oh come on like i'm not i'm still using we're going to be using google like we're

Why would I need to use AI for search? But it's been a couple of years. And do you now, are you now closer to where I was in thinking that, I mean, I don't use search almost ever anymore. And it seems like a lot of people are just blind.

at least people on the bleeding edge have completely left Google search. Do you not see that? No, I, I stand by where I was at the time because it kind of depends on what you're looking for. If you're just like, if you're just looking for quick information, uh,

The original way of searching is slow and reading through, you know, two pages of results and clicking through that doesn't work. But Google has built in their Google browser.

search result at the top of the search page so much that it's almost annoying because if you're just looking for pages, if you're trying to use Google the old way, it's really not as good. But I feel like just for a quick thing, rather than opening up and trying to have a conversation and asking questions,

I can just use Google search the way I used to with a three word query. And it responds with an AI response where when I, when I start talking to chat GPT, or when I start talking to deep seek, I feel like I need to give them like, here's, here's the context of what I'm looking for. Here's the, the parameters that I want you to consider. Here's, I feel like I have to craft a,

prompt more in chat GPT because if I just typed in you know restaurants near me Google knows what I mean chat GPT does not okay but I'll challenge that too though because because I do think that the people that end up sticking with the Google ecosystem will eventually do that and they'll just continue using Google Gemini whatever it ends up being but I think we're

two out of every three people are going to, like myself, I've migrated to open AI, right? So like I'm using chat GPT. Some other people are using perplexity. When you are on a different platform all day long. So my chat GPT is open all day long, Dave. I'm so accustomed to going to it that even though it might be a lot quicker for me to go to Google to do something, I'm actually going to do it the slightly slower way on chat.

on ChatGPT because I'm already on it, Dave. It's a change of behavior. And it's just so weird for me to go back to Google. Yeah, I do think, and I think Chris is right here, there are a certain number of people that have changed their behavior. And like I said, if you're on a laptop all the time or a desktop, you just have ChatGPT open, right? If you're on the go, like, you know,

Maybe like Dave with his Jordan. It's on my phone. I know it's on your phone, but dude, you don't want to have a conversation typing stuff out on your phone, dude. I talk to all of my AI agents. Yeah, you can use your voice.

I'm just looking at my search history on Google right now, and there are just pretty much everything I use Google for. I couldn't use... I mean, like, obviously, I'm searching for, you know, Microsoft PNG logo transparent. That's not something you can ask ChatGPT for. Another search, Amazon. Okay, these are... This is me building a thumbnail last night. I have to search further back in my search history to find something. Okay, but Jordan... I mean, Dave...

Excuse me, both of you. Dave, now you're being very boomery because like, let's talk about Jensen. I saw this video of Jensen last night and he talked about and this is this is so true. This is 1000 percent going to happen. And by the way, I think Google is going to control 25 percent of what I'm about to say, but they're not going to control 90 percent of it.

Jensen said that everyone would have basically an R2-D2, right? And what he means by that is that there will be so much personalization with your AI assistant. Everyone's going to have an actual robot. That robot though, that R2, he said in the R2, some of it will be in your glasses. Some of it will be in your phone. It will then migrate to your robot

Okay. It will migrate to your laptop. It will migrate to so many different things in your life, but everything in your life is going to turn into a personalized AI that knows everything about you. So at that point, you're going to have to pick the ecosystem that you want to live in. Maybe it's the one that is attached to your physical robot because you don't

decide that your humanoid is the most important piece of hardware and you decide to go with that ecosystem, the ecosystem that provides the best humanoid. And it's so well integrated with this AI agent that is the default that that's the one you live with. But whatever one you choose, I think becomes...

Okay.

And they do already. It's already in 18.3. I don't know if you've updated your iPhone lately. Google is going to lose that search business as it gets integrated into the other ecosystems that people choose. And then the nerds are going to use some other random one, right? Like open source one or something. I'm in the middle of this. I think it's a whittling away, right? It's a whittling away and it just,

each little thing that comes out, it just whittles away Google search over time. There are things that I do prefer for Google search because I don't want like,

If the AI knows me too well and I ask for something out of context, I don't need to have an argument with AI. I just need the question answered, right? I just need to know. But this comes back to I don't talk to technology. Like you said, text-to-speech or speech-to-text. I'm out. Dude, I'm not like – I don't want to have a – Sometimes you don't want to have a full conversation. You want to just say three words and get the answer. And I feel like Google does that well. And I feel like eventually –

I think that that use case, my use case, where you're trying to search for something and you don't have to explain yourself. I think that's what I'm saying. So like, let's say I don't like Thai food. I actually love Thai food, but, and one of my friends is like, Hey, is there a Thai place close? And I'm like, Thai place close in Google. And it gives the results. But if I tell that to my AI agent, he's going to be like, well, you don't even like Thai food. Why don't you? I'm like, no, no, no. It's for a friend. Then I'm, I'm having a conversation with a computer and I'm out. I'm out on that. That's funny. Yeah.

Your data is like gold to hackers. They're selling your passwords, bank details, and private messages. McAfee helps stop them. SecureVPN keeps your online activity private. AI-powered text scam detector spots phishing attempts instantly. And with award-winning antivirus, you get top-tier hacker protection. Plus, you'll get up to $2 million in identity theft coverage, all for just $39.99 for your first year. Visit McAfee.com. Cancel any time. Terms apply.

My dad works in B2B marketing. He came by my school for career day and said he was a big ROAS man. Then he told everyone how much he loved calculating his return on ad spend.

My friend's still laughing at me to this day. Not everyone gets B2B, but with LinkedIn, you'll be able to reach people who do. Get $100 credit on your next ad campaign. Go to LinkedIn.com slash results to claim your credit. That's LinkedIn.com slash results. Terms and conditions apply. LinkedIn, the place to be, to be. Now, one of the AI that I find myself just...

Using because it's already open kind of the way you're talking, Chris, is I use grok all the time because I'm always in the X app. Like that's just where I, that's my mindless doom scrolling activity. And when something comes to mind, it's already open and there's a tab there. And so I just ask grok and I think grok does a great job of summarizing just about everything specifically. I use it when I'm trying to find something that's currently topical in the news.

Listen, I actually think that Grok and Tesla, along with Optimus, I think that they are going to be one of those major ecosystems. I think they're going to have part of the market. I think they're setting themselves up to have part of that market. And you're right, Jordan.

It's just going to be whittle away the same way that cable providers are just whittling away over time, right? I see the same thing happening. The big question for Google is...

Can Google make it up if it's a slow whittling away and Google instead gets 25% of this instead of 90%? Is the 25% they get in the future meaningfully more valuable than the 90% of search that they have today? That's an open question. I ranked 10 stocks from most positive to most negative and Google was my most negative.

If OpenAI were a public company, that would have been number 11, but Google is my most negative. While this new streamlining makes things cheaper for them to operate, I feel like you're right. This is going to move towards making their search less relevant, and that's where they make their money. It's where they make their money today. But Dave, now to defend Google, Google actually has maybe the most

amazing data set for everything that we're talking about. Google has the only real time, real, real time data set. They have all of your emails, right? They have all of your search. They have my cloud has my emails now. Wait, you posted out of Google for email? Yeah.

Man, they have so much rich data that's current and real time. But they do. And as I was looking at my search history, I realized just how they know every video I've watched. They know everything about me.

But they have given the controls to delete, and I think I have mine set to delete after three months. So they don't have that much deep historic data into me, but they do have... No, nobody deletes. They soft delete. You can't see it, but they still have... I'm sure they don't delete it, yeah. They don't delete it. But Dave, moving forward, if they have the existing Android...

user base. Okay. And the rich real-time data that comes off of that is so critically important, the same way that it's critically important for Apple. And that's why Apple, I think, is going to be a major leader in this space because Apple also has critical real-time data and

and access to just layer in the best of AI because at some point you are not under the chokehold of having to go to one of these three or four entities that can spend $100 billion a year on a frontier model. If that is no longer the kind of barrier, if that's no longer a moat, then doesn't the person that owns Android and iOS want

And the hardware that goes along with it, because they don't have to own the hardware, but like they own that access point. Don't they ultimately win?

Yeah. Right? No, I think so. The difference is if you believe that, if you believe that, then you have to like Apple because Apple's not losing the search business. You get what I'm saying? Yeah, Apple never had that. They've always been hardware and software and not advertising dependent. Although I'll take that back. In many ways, that payment that Google makes to Apple every year is kind of like Apple is in the search business, you know? So...

So it kind of, if you peel the layers back, Apple is also at risk to losing search. And by the way, I haven't heard anybody talk about that. Everybody's talking about Google losing. Apple makes almost as much from their search off their users than Google makes because of that payment. Google's paying, I think the majority of the money they make, they're paying to Apple. But right now-

So in the new Apple intelligence ecosystem right now, it is open AI's challenge EPT. That is the only option, but the way they've built it, it's like just in the settings as an option. And it looks like in the future, they plan to support other AI's.

So maybe in the future, OpenAI will have to make a payment to Apple for a premium placement. Google will make a payment for their AI placement. Like all of the little plugin AI bots can pay Apple for the prominence in the device that everyone prefers.

Yeah. Until the government comes out and says that they can't charge anybody for anything and they have to allow all of them to be in there for free. Well, the EU, the EU will definitely say that. And our current administration is going to try to deregulate and as much as possible. So we've, we've got like a, Apple will do better in the U S and worse than the EU. It all, it all gets really, it all gets really confusing really quick. So, so,

By the way, I do think that deep seek is likely to come out with more news in the very near term future.

uh this group is sharp they are moving quickly jordan wouldn't you agree and i think a lot of people that even read that paper i didn't read it in its entirety but a lot of people that read the paper had a feeling that there was more to come and i think that's really meaningful right because they already like they dropped this

a few days ago. And then we had this like perfect storm of the X debate over what this all means that tanked the market on Monday. It's like they timed it knowing we were going to have a weekend so that we could have a market impact on Monday. They have already come out with the, another, another model that, uh, that competes with Dolly and image generation. So they, they're staggering their release, but they've been working on this stuff for months.

So when I'm looking at this space, I just want to read you a few lines that I kind of clipped from all my research I thought were pretty prolific. This, you know, as this migrates, especially the open source nature of it and the fact that everything can now theoretically be run locally, right?

and we have this massive democratization of AI, new classes of product that were impossible will become possible now. So this will create space for things that could not ever exist previously. You could make a case that innovation is no longer linear. It's becoming exponential. We're watching the birth of a new information paradigm.

One related to Apple I thought was really cool. Cloud is building everyone's AI, but Apple's building your AI.

That might be the number one tailwind for Apple is the degree of trust and personalization that Apple has on its customer base. Yeah, from a brand trust standpoint, that's the reason I left Gmail and switched to iCloud for my primary email provider because I trust Apple more than I trust Google.

And I think that I would trust Apple's Apple intelligence, even if it's written using, using like, I don't know if they're sitting back, back there, like scrapping everything they've been working on and trying to put deep seek rebranded as Apple intelligence as the thing that's coming out in April. But I would still trust the Apple version of that more than the Google version.

Agreed. But Jordan, all those things that I just said, as an engineer, would you agree that if I'm trying to assess who is the ultimate winner there with the least amount of risk? Because everything is going to change going forward if that stuff is true. Wouldn't you say that Amazon feels like the ultimate winner? Because if we're truly...

truly going to innovate exponentially and, you know, products that were never possible before, creating space for things that couldn't exist previously, a new information paradigm. Doesn't Amazon ultimately capture a lot of the kind of back-end

kind of behind the scenes kind of broadening of the layer that all these new companies would ultimately have to sit on, right? Or a lot of them would, they won't capture all of it. But don't you think they just benefit from all that without having to make a lot of big investments in the space? Maybe, I mean, but if you look at DeepSeek, like they have their own cluster, right? They have their own H800 cluster, right?

but this is not about hosting models and hosting innovation that's kind of what they do so yeah i mean i can see that a little bit but also this research team was just pinging the heck out of open ai's api to be able to even pull off what they did right and i think that gets a little bit i think that gets too little of attention that it seems like they did this all on these 800 clusters um and that's partly true

part of the training that they did was asking questions to ChatGPT and then getting those answers back, really relying on Microsoft Cloud. So I don't know, it's kind of, it's an interesting way to go about things. And so I don't really know who the winner is. - I mean, and the reason why I was leaving Microsoft off, they would benefit in the same exact way, the difference. - Yeah, so as far as like a cloud service provider and being able to utilize GPUs in those,

uh cloud services then yeah they both they're both those uh cloud services win well the only reason why i left microsoft off is because it seems like they're making such massive investments now they can cut those investments off right they don't have to keep funding open ai to the extent that they have been funding them in the past uh but amazon

kind of wins in multiple ways because one, it just depends on where people like to host things. I prefer AWS, but I'm not everyone. So any, any,

But they're going to get a team's going to make their decision on what, you know, what systems to use. But Jordan, what I'm saying is they're going to get a piece of it. I don't know if it's 30 percent or 40 percent or 50 percent. They're going to get their piece. Yeah, they'll get it. As long as people are continuing to invest in trying to create new models, then, yeah, I mean, it's just it's just Amazon selling compute.

Without having to invest hundreds of billions of dollars to lock up the one frontier model that is going to have a chokehold on the market because there will be no one frontier model that does that. Well, there still is one frontier model, and that's whatever OpenAI's best model is. Because DeepSeek still isn't at...

you know open ai's pro model level yet it's not as up to date it's fantastic for what it is um but like i said earlier the only reason they could get it that far is because they were pinging supposedly they were pinging open ai's api to to be able to um

okay but now when you look at open source when you look at the impact that this has on open source ai and the talent that will likely move to open source ai you look at models like llama right like how can that not be uh you know kind of a favorite going forward over the next few years to be a front you know a

open source frontier model. Yeah, I still just come back to these guys were hammering open AI to get answers and to help train their model. So they were using open AI. If they were going to use a different, if they were using Lama to help train their model, then yeah, maybe, but they weren't. They were using the best.

Okay. I don't know. That's the one thing that gives me a little bit of pause when I want to lend too much credence to Deep Seek. But to remain the best, you have to make money.

to keep building, to stay the best. And the problem is that open AI just came out and said that they were going to contribute $19 billion. Okay. Uh, to this new half trillion dollar initiative, they don't have $19 billion. Yeah. So they have to go out and raise money. So this, you know, there was some chatter in my email box this morning from some people in my network, uh,

that are saying that they are going to potentially definitely pull forward in a big way, whatever they have, because they have to impress investors now because they need to raise this money. So open AI is going to have to pull something out of their hat and whatever they have, whatever they can show, they're likely to show it. And they're likely to show it really soon because if they can't raise that money,

It's game over for open AI. They're done. They're done as a frontier model. Like they have to be able to keep investing. It's Sam Altman's entire strategy hinges on the fake it till you make it.

raise massive amounts of capital and then nobody can catch up to you because you will always be on the bleeding edge with more GPUs pushing the boundaries and then you'll be able to hire better talent because they will want to work for the one right that's on that that is pushing the boundaries and you have the best processing power to do it and it's like he's creating his own moat

But if that ever... Like, Elon did the same thing, by the way. But it's a risky game because if at any point you're not able to raise the capital...

it's game over. And even if you are able to raise the capital, if we're entering this new era of AI where maybe technique becomes equally important to GPU, to compute, not saying it's more important, if it becomes almost as important or just more important, then we could get a flattening of the frontier models and his edge might not be, maybe he'll be better, but not that much better. And it just won't make that big of a difference at the end of the day. You

That you're not going to pay a lot more money to get like, you know, that much better of a model. So there's one thing I think we have to discuss, guys, and it's Trump. So this did not impact the market today. And I haven't read the news this morning, so maybe he'll pull back on this a bit. But Trump made a statement yesterday saying,

about semiconductor tariffs. And he made a statement that it could be 25%, it could be 50%, it could be 100% that he wants people making chips here and that he doesn't want to pay them to do it, but he's going to penalize them if they don't do it. Now, I don't know how he actually pulls that off because there's only one

basically one company making all the chips. Don't they have a fab open in Arizona? Well, they have an older fab, Jordan. It's an older fab, but they already have committed. Are they not doing the four nanometer process at that fab, at the Arizona fab? I don't know enough about it. I do not think it's even remotely close.

to what they're doing in Taiwan. But they've already committed to a major project in the US. Here's why I think it didn't impact the market, because people just are like, this would be insane. I don't think you could pull it off and they'll probably forget about it. But let's just read some of the other people in Congress.

of what they're saying, because this is a massive risk factor that we all need to be keenly aware of. And quite honestly, probably following these people on Twitter, because anytime that they make a follow-up remark, it has potential to really knock down the chip manufacturers and NVIDIA, the chip designers. So

Rep or Senator Mark Warren said defending existing export controls related to advanced chip technology said that there's more regulation that might be needed. Rep John Molinar from Michigan, he's the chair of the House Select Committee on China. He said that he wants to see the U.S. act to slow down deep seek.

The U.S. cannot allow CCP models such as DeepSeek to risk our national security and leverage our technology to advance their AI ambitions. We must work swiftly to place stronger export controls on technologies critical to DeepSeek's AI infrastructure. That would be devastating for NVIDIA if they really put in stronger controls because we all know what is like 20% of NVIDIA's revenue coming from Southeast Asia? We all know where it's going. Like everybody knows. It's like,

Not a big speaker anymore. House Speaker Mike Johnson said that they steal our intellectual property. They're now trying to get a leg up on us and AI. It's a serious threat. So the president takes that seriously. And I think he will deal with that in an appropriate manner. So the threats just keep on coming. I think there is real risk here, guys, of...

them putting tighter controls on NVIDIA, and if that starts to generate buzz, could be a big negative for them. And if this whole tariff thing on semiconductors ends up being real, or they even discuss it much further? I just don't see how putting tariffs on semiconductors helps secure the United States position. It doesn't. Right.

we're already listen they're already planning a major project here it will take a decade to build out yeah uh i think trump just doesn't want to pay them to do that we were giving them like huge financial incentives and trump is saying hey i'm not going to pay you to do that i'm going to penalize you if you don't do it so there could

Maybe that works. And like, but it's not super significant. I don't think it's going to ultimately end up in actual tariffs being put on that sector. I don't think his advisors would actually allow that to happen by using it as more than a threat, a negotiation tactic. But it is something we should at least be aware of. Yeah.

Okay, so I know we've been kind of negative here on NVIDIA, but again, I think there's a pretty good case to be made that as this transition happens, right, and we get away from kind of all the pre-training models, all of these DeepSeq R1 models that are using RL are

are actually really compute intensive. And if the entire AI sector transitions to reinforcement learning, you could make a case that that will ultimately benefit NVIDIA in the long run, except the way in which they were conducting the reinforcement learning at DeepSeek was so innovative that

It's not reinforcement learning as we've known it today. Like they're just doing it way more efficiently. Still, it probably still even done more efficiently, it probably still generate, it takes more compute if everything goes to that type of a model, right? As opposed to just,

kind of the more basic models, the way that we've been doing it previously. So I think long-term, because you got to remember guys, like the frontier models eventually were never, were always going to, I don't want to say end, but there was always kind of an end point.

on when we stop spending quite so much money on frontier models, right? So this makes a case that there's something very large beyond the frontier models as it relates to reasoning. Because reasoning takes compute. Even if we do it more efficiently, reasoning does suck up compute. So long-term, that is a huge positive for NVIDIA. I just don't know what the next few weeks looks like, next few months. I don't know. Anyway. Yeah.

I agree. But I've been doing nothing but think about this. NVIDIA is the loser here, but I think that long-term, NVIDIA is a stock that I will own. I own now. It's one of my top holdings still. I only sold it in one of my two accounts. So how do we think about Tesla here? Tesla earnings are this week. Tesla, I think, should, and let's not talk, like the majority of Tesla's valuation right now, their market cap,

While I believe it should be based on Optimus, it's not. It's not even based on automotive. It's probably based on FSD, right? Let's just say a trillion dollars of that market cap is FSD right now. I think that's a fair assessment. Can you make a case that this breakthrough benefits FSD in a pretty meaningful way and would accelerate FSD over the next year?

I think you have to, right? Yeah. No, I think, I mean, I don't know exactly how they would implement this to improve FSD, but I feel like anything that can optimize and give us better use of existing GPUs is going to benefit the companies using them. I think it's one of those that spend a ton, and I think will continue to spend a ton, but I don't see how it would hurt. It could only help.

I would think that any company that has massive AI infrastructure is going to take a look at the paper and try to figure out, hey, what can we learn from this? Is Tesla the dark horse? Is FSD the dark horse beneficiary of DeepSeek? It could be. It could be because it has the most to gain in the short run. It's not an LLM. How much of this can you abstract to other models? But

Maybe. I don't know. I don't know enough about that to give an answer. But we know that it could help their LLM Gronk. Yeah, but Tesla doesn't have a stake in X. That's true. They just have it installed in the vehicle. Why are my glasses so crooked? Are my glasses crooked? Yeah. Your camera is crooked this way and your glasses are crooked this way.

That's so annoying to me. The other dark horse is Corsair, our episode last week. We talked about this over the weekend. We did an entire episode last week on Corsair, and yet it was all about how their chips could benefit from... Excuse me. It's all about how their home PC unit could benefit from

from Nvidia's new chipsets for gaming we wouldn't we didn't even have a discussion around AI builds on those machines the hobbyist AI build does this you know does this spark an interest in people wanting to run all these open source models it does it it does Jordan I'm going to prove it right now so one of these tweets that I kind of captured

Hold on. I took a photograph. I was up so late last night. I don't even know what I was... I'm going to put my glasses back on to see here. You took a photo of your screen? Instead of a screenshot? I took a photo. You called me a boomer earlier? 100%. Okay, here we go. So...

Running your own private model will be the key. People won't want to share their private insights with the world or let their AI provider resell their insights to others. There are a lot of people talking about this right now. This could end up being the biggest tailwind ever for this tiny company, Corsair. I don't know, but it just seems like it could be.

Because you could run AI now on your desktop. Well, you can do it right now. I actually have Ollama running on my MacBook Pro. It'd be better if you had a GPU. It'd be better, right? Yeah. So, I don't know. They could be the ultimate dark horse here. We'll see how it plays out over the next few months. But for anyone that did invest in Coursera or is evaluating Coursera, I think that has to be part of it.

of your research it was completely off our radar deep seek existed when we did that show but nobody was talking about it yet yeah so matthew sharp is saying that tesla trades on auto sales and this has potential to be a really rough quarter maybe that's part of the reason why tesla is trading down today uh this bsd guys oh jordan you're a pickup truck guy

Have you seen this new BSD pickup truck? It's pretty amazing, dude. Pretty amazing. I watched a video last night. I watched a few videos on it. And I was like, is this thing really legit? Like, it's not being sold in America. What's it called, Han? BSD. Wait, I don't even see a BSD. Like a...

I don't see anything. Am I saying the name of the company wrong? I went off no sleep, man. Is it made by BYD? BYD. BYD pickup truck. Oh, God. We got roasted for that one. So it's the Shark. The BYD Shark. The BYD Shark. It is really impressive. I mean, I saw off-road test on the beach.

hauling a boat behind it on the sand this thing is an absolute beast and it could be i'll tell you it looks a thousand times better than the cyber truck and and it costs like they're saying the north american price if it were to sell here could be as low as 30 000 as high as 50 or 60 000 it is

This is legitimately a concern for the auto business that I don't care about. The problem is Trump will just tear up the daylights out of it. It will never make it over here because of that. But that's the thing, Jordan. It's not just about the U.S. They're selling these in Australia now. They're selling these in South America. They're selling them everywhere.

They're about to start selling these all over the place. I know nothing about it, but I've heard that BYD makes a great vehicle, and this looks great. So I don't know. Jordan, if you watch some of the reviews and the test drives, it's so impressive, especially for the price point that it's at. It's just absolutely wild. But again...

Who cares about the automotive division of Tesla? Tesla should basically just be like, we don't even sell cars anymore as far as I'm concerned. Could you imagine? That's what they announce on their earnings call. Their market cap would probably go up. They're like, guys, we're not going to waste our time and resources selling cars. That's just the dumbest thing ever. Why would we waste our time selling cars? Why would we employ anyone to do that?

No, we're obviously going to business in the entire company. We will sell cars that are FSD only that are purely for robo taxi or will manufacture robots. And that's all we're doing from now on. That would be a smart thing for Tesla to say. I don't think they're going to say that. And I'm concerned that this earnings thing.

Could be rough. So I'm going to hold a moderate levered position in Tesla that I'm not afraid to lose if earnings are terrible. That's how I'm viewing this this week. Interesting. Yeah, I'm not touching my Tesla stock. I own way too much of it and will continue to. Yep.

I'm just seeing if there's any other comments. In my ranked list of most positive to most negative, Tesla was number four. What are your top three most positive and top three worst? Top three most positive are Meta, Amazon, and Apple, followed by Tesla. And worst, OpenAI, which isn't a stock, but Google, Nvidia, Microsoft, Apple.

Now we've got Dave mispronouncing NVIDIA. What is going on? Chris is pronouncing it right, and Dave is mispronouncing it. I don't understand what... Did you... NVIDIA. Chris? NVIDIA. Yeah, I know. I've just heard him say it wrong so many times that it's just entered my brain incorrectly.

Okay, Jordan, what's your top pick on post deep seek? Who's going to benefit the most? Who are you most concerned about? I'm most concerned mostly about long-term Google. Pluses and minuses there, I don't know. I think you're right on Amazon though. Amazon wins. AWS is the king.

I hope Apple wins. I really hope Apple wins. I'm really wanting this non-Apple innovation to boost Apple's innovation. Here's the thing with Apple. They have the ability to push something out very quickly to a ton of people. And you can't underestimate that. But, you know, people aren't really excited about the next iPhone every year. So they need something.

And they tried with Apple Intelligence, but the problem is Apple Intelligence wasn't ready. And so people who bought the phone only have half of the features. And they just released 18.3 and 18.4 is where we're going to finally get the improved Siri that we've been promised. And I'm waiting any day now. We should be getting the beta version of that. I cannot wait, but I am waiting. We didn't talk about this company much on this episode, but Palantir is one of my top picks post-DeepSeek.

yeah so my list is moderately positive for them the better margins yeah exactly jordan i think more importantly not that the entire commercial and industrial world wasn't already woken up to having to employ ai and everything that they do before this but more than ever more than ever

This is going to, I think, empower Palantir to get into the kind of small to midsize business market that I think is like their future the next decade. I know they're not really focused on it right now, but that is the next big thing for Palantir is for them to go downstream as this becomes mainstream.

Exactly what you just said. So as AI becomes democratized and more approachable and more affordable, companies like Palantir can make it more accessible and a better way to work.

Way more companies. This is how I see Palantir evolving into more of a sales force. I don't really have an opinion on sales force right now, but this is how I see Palantir making that transition. Palantir needed this. They needed AI to be cheaper. I think it's a really big deal for them.

Now, Chris, you sent me when we were preparing for this episode, you sent me a couple of other tickers we have not talked about. Did you want to talk about Tencent or Alibaba at all? Yeah. So there is a there is a thesis that this is pro China in a very big way. And I think just that was more of a short term trade on, hey,

you know what like china can do really great things this is obviously going to benefit the big chinese tech giants uh probably the same way that we're talking about this benefiting meta here in the us dave uh how could it not equally benefit those two companies in china and you could even make a case that because of the price point and the price sensitivity around accessing ai in china

uh where the average income is less and the average business has less money to spend on ai that this could be even more prolific in china than it is here in the us because ai for a lot of the chinese market even for the lot of the chinese commercial market was just unaccessible but if this democratizes it the big

tech giants of China are going to massively benefit from that. Right. So I think it's, I think you have to include them in your, in, in your research. I haven't purchased either yet, but I'm contemplating it. I haven't either. I'm going through comments here. Make sure we don't miss anything guys. I think we're missing a lot. Actually. I like, there's so much more I want to talk about, but this can't be a six hour episode. Um,

No, we are already an hour and 20 in. It doesn't seem like it. It seems it some, some shows it seems like when we hit the hour mark, it's like, I'm exhausted. This one, I'm just, I can keep talking about this forever. Jordan, we keep talking about this in this, you know, like an iterative reinforcement learning model that deep seek is really pioneering. That's sucking so much cost out of the system. Like what that's. So this is something that the, the,

OpenAI does too, but they just don't make external API calls to ask its own model questions in the middle of training. They can just do it internally. It is mildly unique, I think, to use API calls to do some of this, but no, I don't think that's unique for them.

But I guess the way and the extent to which they were leaning on it, right? And as opposed to the supervised training. Well, to rely on like a third-party model, yes. But like I said, in the middle of training, from what I understand at least, OpenAI does the same thing. So they'll self-ask questions within the model to get clarifications to then build out the model. So it's kind of a circular thing. Jordan, did you watch the HBO show Silicon Valley?

Yeah. So is this a Pied Piper moment of inside out compression with tip to tip efficiency? I hope it's more like not hot dog. But that's the other one, the one that you're mentioning is inappropriate. I wasn't I'm not gonna do the hand gestures. Oh, you know, inside. Oh, my God. Anything else?

I am going through my notes. I want to make sure I touched on all the big stuff. I think we pretty much... Oh, God, there is other stuff, though. I just can't do too much right now. We can do more than one show a week. I know. We'll call it a day. This is an exciting moment, and we'll see what... Oh, that's what I wanted to say. Yes, thank you.

i do think you know we talked about how this could impact fsd i actually do think this is going to have a massive impact on the models that will run robots and will be really really impactful for optimus as a public company right so

uh that no one's really talking about that now i would love to hear elon speak about that on the earnings call let's see if he actually does i think elon's going to lean in on optimus for the first time in a really big way during this earnings call i just don't know the extent to which the investment uh sector is going to care about that or care more about the crappy numbers they might release related to their auto division but if he does talk about it

it's just so big right like this is it almost like as much as i don't want tesla to have bad numbers i'm i'm kind of hoping that whatever they have makes elon kind of feel like he needs to say more because when he when he does get aggressive on these earnings calls and give us a better vision of of elon vision for the future

You have to discount that with his timelines, but I think that it does excite investors. And I think that if he shares with us what he's thinking about what these models could do to his models, I think it, I think it's big news for Tesla stock. Yeah. Agreed. Agreed. So let's see him wait. Let's wait and see what he does says this week. I am. My position is not massive going into earnings. It's, it's big enough that,

That I'll be, I won't be super disappointed if they kill it, but I would prefer Tesla go down this week. It is a big earnings week. Microsoft meta and Tesla all tomorrow. Apple the next day. Google next week. It's almost enough for us to do a live stream of these earnings calls. But do you know what is so boring? Listening to us listen to an earnings call. I hated that. I'm going to be listening to Amit, who I hope does live stream these, though. Yeah.

if you're watching all right anything else i'm out we're all out thanks for watching we're done money we'll see you uh as soon as we make another show