Deep Dive
Shownotes Transcript
Will M&A pick up in 2024? Will this year mark the return of IPOs? Listen to Strategic Alternatives, a podcast from RBC Capital Markets to get fresh insights on the trends and market forces impacting deal flow across sectors and find out how companies and investors are preparing for a potential surge in
in deal activity, and what signals to watch for this year. Listen and subscribe to Strategic Alternatives, the RBC Capital Markets podcast today, wherever you get your podcasts.
Guy, I'm sure you're already up to speed on all things current. Of course I am. Their app makes managing your money, saving, even building credit super easy, Dan. They have a new feature called Paycheck Advance. You can get up to $500 before payday when you switch your paycheck and qualify. Now that I didn't know. So now when you sign up and you set up direct deposit, you unlock so much more.
And it's super easy to get started. Just go to Current.com slash OK. That's O-K-A-Y. That's Current.com slash OK. Current is a financial technology company, not a bank. Banking services provided by Choice Financial Group, member FDIC, and Cross River Bank, member FDIC. Paycheck Advance is offered by Finco Advance, LLC, a financial technology company, not a bank.
Eligibility and available amounts may vary and are subject to change at any time. For full terms and conditions, visit Current.com or call 888-851-1172 for more information. All right. Welcome to the Risk Reversal Podcast. I'm Dan Nathan, joined by Deirdre Bosa. She is the host of CNBC's Tech Check. Debo, welcome.
Hi, so happy to be here. It's been too long. You've been really busy. You had mega cap tech earnings. I see you every night on our program, but you are all over the air. It seems like tech is the only thing people care about these days. I mean, I guess we could have said that for the last few years.
I thought that was just me because I live in San Francisco, but I've felt like that in the last few months more than ever. And not just earnings. I mean, earnings is part of the AI story, which has just gone bonkers over the last few months. Yeah. You know, it's funny about the AI trade too, in a way. It's like...
I felt that starting in 23, there was a very few ways in which to play it in the public markets. And we knew that some of those private companies, they were bubbling up as far as the amount of cash that they were able to raise, the valuations that they were trading at. And you and I have talked about this. There was this whole bubble.
bubble or pocket of risk that was going on in the private markets, but there were very few ways to play it in the public. And the hyperscalers was one of the big ways, obviously NVIDIA. Now we've seen, and I know we want to talk about this a little bit, some of the custom chips and as these models evolve. And I know you want to talk about large bottles.
versus little models, that sort of thing. I know you've been reporting on that. We also just kind of want to hit the fallout from DeepSeek. And then I think the most important thing that's happened to me over this earnings season or from my perspective is how investors reacted to some of these big CapEx numbers from the hyperscalers. So I'd love to get your take on all that. And if we have some time, a name that's near and dear to your heart, Uber, Bill Ackman of Pershing Square takes a stake. And the last few months of this stock have been
really interesting if you go back, what, three or four months from the Tesla robo taxi announcement and what happened to the stock afterwards and what's happening now. So where do you want to start? You want to start at what we've heard over the last two weeks. It started out with Microsoft and their results and their CapEx. Then we got to Alphabet, then we got to Meta, and then we got to Amazon just last Thursday.
And let me just frame this for a second. You know, Microsoft was down 6% the day after, Google was down 8%, Amazon was down 4%, and Meta was like flattest or whatever. And a lot of it had to do with their earnings growth, their sales growth, and then the OPEX number. So what were your big takeaways from the earnings season? My big takeaway was this continues to be sort of the AI story, right? Running through all of these companies, how much are they spending? When are we going to see return on investment? Which has really been the story for the last few years. But
I would argue that the AI story has changed so much over the last few months that we now should be and can question everything we thought we knew about this trade, right? How much it costs, who the winners are, are the winners the most capitalized? And
DeepSeq and what DeepSeq represented, I would argue that it's not just DeepSeq, but this idea of distillation using smaller, more targeted models, being able to essentially copy some of the big ones that cost hundreds of millions of dollars to build. This has changed so much, and yet the CapEx numbers that we saw from Microsoft, Meta, Google, and Amazon were revised up in many of these cases.
And we still don't have a clear answer on return on investment, right? When we're going to see this, but there's more questions on the ROI side of things because we've seen from a number of labs and universities now that you can do this cheaper and more efficiently. So the stakes have almost grown larger. It's good for NVIDIA, right? I think that people were most concerned with the NVIDIA in all of this and they're
No matter how much you think it costs DeepSeek to produce the model, the idea that you're just going to have more AI more efficiently is going to be bullish for someone like NVIDIA. So that was sort of like my big takeaway. It's still the most important story, but I don't think that the street has a good grasp on
how this is all going to shape up. Even though we have those CapEx numbers, I'm a little skeptical. I think that those could change. You know, when you think about NVIDIA and you think about like that knee-jerk reaction two Mondays ago down 17%, I mean, it gapped lower. It didn't see an uptick all day long. And I think a lot of the, you know, again, the knee-jerk was that this is really bad for them if this model was really...
trained on whatever the amount of-- they said $6 million. We know that there was a lot of compute that was used external to China that was using Nvidia GPUs. And my take then was that this is a shot across the bow to our US national champions that thought they had such a big lead, you know what I mean, on everyone else in and around the world.
And I heard a really interesting take that if this was like Mistral out of France or something like that, that, you know, Silicon Valley would have been like welcoming to this. You know what I mean? It'd be something. But the fact that it was the Chinese was something that they supposedly made really good innovations with much worse equipment.
But can I tell you, Silicon Valley was extremely welcoming of DeepSeek's innovation. There was sort of the Wall Street of it. What does this mean? Panicking about NVIDIA, you know, the big market mover. What does this mean if NVIDIA doesn't have the same, you know,
What does it mean for everything else? But Silicon Valley was interested in first and foremost, the technology, the breakthrough, this idea of distillation. So everyone embraced it. And actually the mega caps did as well. Microsoft was saying, hey, we're offering it on Azure in a blog on the same day that it was, you know,
accusing DeepSeek of copying open AI illegally or against its terms of service. Amazon adopted it and even Andy Jassy tweeted like, "Hey, have at it. Here it is for you." I know you know, Dan, that Perplexity was one of the best at incorporating it and showing how because this is an open source model, you can strip out the censorship, you can host it locally, taking away some of those Chinese concerns.
there was like really these competing narratives. Wall Street was a little bit afraid of deep seek and made it into China versus U.S.
where on the Silicon Valley level it was appreciation for what they were able to do because of what that will do to the race and the breakthroughs that Meta and others can build upon but also like the Silicon Valley attitude was let's just keep innovating like let's get to work here why did they discover this and not us let's build upon it because it's open source and continue to you know
adopt DeepSeq and what it taught us. Yeah, let's take a step back for one second because you had a great tech check on this and you were quoting Arvind Srinivas, the CEO of Perplexity. And you made this really great point in the back and forth was that there's two ways to think about this model, right? And so it's open source and you can incorporate it into some of the tech that you're doing.
you know, Perplexity put it right in the R1 into their different choices of models to use. And then there's the app. So there's two different things, right? So if you want to download the app and use the app, you know, there's a privacy issue in that regard, but not really as it relates to the model and open source. So give us the 411 on that because I think that's a really important distinction.
Right. I did a whole tech check on this, just the distinction. So if you're downloading the DeepSeek app, which by the way, you know, shot to number one on the App Store for at least a week after sort of Wall Street got wind of it, probably safe to say your data is being sent back to China and being stored on Chinese servers and you can't ask...
the model things like what happened in Tiananmen Square, anything around that. So you're getting the fully censored, fully Chinese hosted version of DeepSeek. But the model itself is also called DeepSeek and it's open source, meaning that an American company can basically download it to its server or its computers because it's extremely smaller and efficient and
and then host it in the US. So there's no worry about sending the data back to China. The risk is in the company that's hosting it. And they can strip away the censorship like Perplexity did. It's weights are open, meaning that you can tweak this model and make it a lot more secure and a lot more free, we'll call it, free from censorship.
Yeah. And so, Len, let's take it to TikTok for a second here, because, you know, perplexity has kind of made this really odd bid for TikTok. Obviously, they would need a bunch of partners. They were also talking about, you know, the U.S. and some sovereign wealth sort of situation, which we don't have a fund to do that to help with this bid.
give us the sense of like what is similar about the situation with deep seek and what's different because you know when i saw that news that monday and i had a pod with gene munster and we were kind of going back and forth in this is like if we're really dialing up tensions as far as a trade war
with China, the low hanging fruit right away is just to ban TikTok, right? And if you think like the flip side of that is the Chinese have opened investigations into Google, into Apple and that sort of thing. So, but the fact that we are reluctant to do that seems like, you know, it's become a political hot potato if you think about it over the last six or seven years. But
there is a differentiation between deep seek and TikTok because the version here in the U S while the data can go back to China, we don't know if it is right. A lot of those servers are hosted here. You know, there's this, there's a separate TikTok in China. You know what I mean? So there's one made just for us here in the U S. So let me say that if you're worried about TikTok,
about TikTok, then you should be deeply, deeply scared of DeepSeek and a Chinese-hosted DeepSeek, I should say. Again, we're talking about the app that comes straight from the Chinese AI lab that may or may not have connections to the CCP, probably though.
TikTok, you worry about influence campaigns, right? You worry about sending people social media preferences and biases back to China and that the Chinese can then turn around and run these campaigns that subtly influence the hundreds of millions of TikTok users here in the US. Within chatbot, right? Chatbots have the ability to shape ideology, right? Which is a much scarier proposition.
I don't know about you, Dan, but I use chatbots all the time, right? And I ask a lot of stuff that I would not necessarily like to be made public, right? It's really like you're giving a lot of information to these chatbots. You're basically showing the chatbot how you think, what you're interested in, how you process information, and then you're getting it back, right, to process information yourself. So I think that's like...
the fear for a Chinese chatbot versus a Chinese social media application is much greater and has much bigger implications for American freedom and ideology. Yeah, it'll be interesting to see now that there's kind of this barbell sort of situation between a chatbot and then, you know, an app that people are
absolutely addicted to that does, you know, it is susceptible to these sort of influence sort of campaigns, no doubt about it. I mean, you know, Trump himself has said it helped get me elected, which is one of the reasons why, um, but it does speak to political influence. If you go back to 2016 and you know how there was all this kind of assumption that Facebook was co-opted by, you know, foreign adversaries, whether it was, you know, Russia or China, um,
you know, or the North Koreans or Iran. I mean, it could have been anybody kind of pushing different narratives to affect our electorate. Those are kind of scary things, you know, when you think about it. Let's just put a bow on some of these, some of the CapEx concerns and we can get away from deep seek
But, you know, to me, I think the reaction by investors and I kind of alluded to this was to sell Microsoft, to sell Google, to sell Amazon. And to be fair, Google and Amazon were trading at all time highs, you know, heading into the print. Expectations were really high. Microsoft's been a big laggard over the last sort of.
year and a half, and I think partially because it was one of the early beneficiaries due to their partnership with OpenAI. Let's start with Microsoft-OpenAI relationship. Where are you on this? Because it seems to be in flux. If they're
Or like on an old school Facebook relationship, it would be listed as it's complicated. I don't even know if Facebook still does that because I haven't logged in in forever. But the relationship went to one that was like extremely beneficial from both, right? Microsoft gave $13 billion to open AI at a time when everyone, not everyone was an instant believer, right? It
might have had trouble raising 10 plus billion dollars. And in exchange, it got this exclusive cloud computer agreement. It seemed sort of like this match made in heaven. And as you said, Dan, like Microsoft saw a lot of the early gains in terms of its share performance in the first year, let's say after ChatGPT was released. But now, and this goes back to what we've seen with like the DeepSeq breakthrough,
And even before then, models are just increasingly commoditized and Microsoft wants to spread its bets because companies that are going to be using Gen AI APIs incorporated into their company organizations, they want choice.
They don't necessarily want to use OpenAI. They may want to use Anthropics Cloud. They may want to use Gemini. They increasingly want to use DeepSea too because it's a lot more cost efficient. So Microsoft has good reason to spread its bets. And then, of course, OpenAI raising, what, $40 billion from SoftBank. Microsoft is no longer the priority. And they get a really crappy deal now with Microsoft Cloud and using Azure. I mean, the agreement...
pre-project Stargate was that they had to run all of their compute on Microsoft Azure. That has now changed. Microsoft has right of first refusal, but how do you go and negotiate with AWS, Amazon's AWS or Google Cloud when they know that Microsoft has right of first refusal? Why would they lower the price for you just for Microsoft to come in and take that? So they're in a tricky position. And that's why I think that's part of the reason
they're kind of pushing their custom chips there's a lot of reasons to do so strategic reduce their dependence on nvidia but i think part of it is they want to go build data centers with oracle with project stargate and with the tens of billions of dollars that masasan is giving them right and a couple of interesting things there to me is that first and foremost that microsoft part of the reason it got hit so hard after its results and it's kind of stuck in the mud here is that the deceleration that they saw in azure
is only likely to be further magnified, especially if they keep spending the way that they are, tens of billions of dollars. They kind of suggested it's going to be $80 billion. And so you go from being capacity constrained, we heard that again and again over the last kind of year or so, to now maybe having the capacity, right? And now, especially if you have open AI that's broadening out somewhat more,
away from them. And if you think about it, how some of these companies are going to compete on the data center side is on price. You know what I mean? So you're going to see market share shifts, but you're going to see decelerations with some folks like Microsoft who might not be in the catbird seat anymore. And
The other thing that was really important, I think I heard you say this, is that Sam Altman getting in there in Stargate and kind of doing an end around Elon Musk, right? And then if you think about Larry Ellison doing an end around on Microsoft, it just seems like there's going to be a lot more competition.
especially now that you have SoftBank in there throwing tens of billions around, that you have this talk of Stargate, which is kind of BS. It's going to be a half a trillion dollars committed and deployed over the next four years. But talk to me a little bit about the psychology of Satya Nadella saving Sam Altman's job, Sam Altman and Elon Musk bitter enemies.
Larry Ellison, where their cloud business, OCI, is just a small percentage, really growing fast off a low base. It just seems like all of a sudden things got a little haywire at a time where a lot of folks, like you said, are questioning return on investment and the commoditization of these models and what are the use cases to be monetized in the near term.
And then you add in Masayoshi-san with his 300-year vision and his billions and billions of dollars. It's just like, bring out the popcorn. There's so much going on. There's so many intertwined relationships and places for falling outs and complications.
In my mind, there's sort of two races, two major races in AI going on right now. There is the race between the hyperscalers to host more sort of inference level compute for the enterprises that want to use these cloud platforms to, you know, integrate
AI into their own organizations and become more efficient, right? And that's where, you know, Microsoft is competing with Amazon and Google. Google is in a really interesting place because they're also developing the models, right? And they have their own custom chips already, which are making things, you know, more efficient, they say, the TPUs. And so,
The cost is coming down and they're all competing for this business and who can do it the most efficiently. And that's still going to cost billions and billions of dollars in CapEx to host all of that activity. But there's another race going on that I would say is between more the foundational model builders, the open AI, the Googles, the Anthropics, the Llamas, the Mistrals.
to keep pushing the frontier, right? To them, like distillation is almost a dirty word because it takes what they've already done and essentially copies it. Even though, you know, many people argue that there's a lot of innovation on top of that. That's not just deep sea, that's researchers at university of Berkeley, that's Stanford university of Washington. They're all doing these models even cheaper, even more efficiently, even quicker. So that second race
is something that, let's say, Masayoshi-san and Sam Altman are spearheading. They have to push the frontier. They have to make the breakthrough. And I like to call this AGI at all costs. That's what they're trying to do. This is elusive. You're not sure you're going to get there, but they're happy to spend billions and billions of dollars because that's ultimately who's going to win. That's how...
Open AI won in the first place with ChatGPT. It created this product that people started to use that broke it into the mainstream. So the first to reach AGI is going to reap a lot of those rewards. And you hear Sam Altman talking about that more and more.
Yeah, well, I'm scared of AGI, to be honest with you. And I know that- I don't even know what AGI is anymore. See, this is why I hate using these words, because there's so many different definitions of it also. Yeah, I mean, listen, I think it's interesting that R1 is a reasoning model. I mean, that's like one of the first steps that a lot of folks think is like, if a machine or a model can reason, then it can get to a point where if it can do all of these tasks, right, that humans do, and then they add on the reasoning functionality,
You know, I always take the over in these sorts of technologies. I know that we're in a period where things just seem to be like, you know, just moving much faster than they did in past tech cycles. If you just go back to like 09, the convergence of, you know, cloud and mobile and social and, you know, like to see mass enterprise adoption of those things, it just took longer than people expected. Then you go back, you know, to late 90s and, you know,
know how the internet was going to transform all these different industries when there was really only a couple killer apps at the time it was basically email and web search you know what i mean and so um you know so i take the over but again the technology that's being deployed to do this the amount of capital that's doing the global aspect of it
I get it. And to me, I think there's also this thing that didn't really exist as far as the internet is concerned. It's like this, you know, to your point about AGI at all costs. I mean, at some point the winners of this and it becomes sovereign AI. It's, it's,
really could be a tipping point for the way we see geopolitics going forward. And I know that sounds maybe a little pie in the sky, but think of a company like Anduril. It's just like they are raising money hand over fist. Their valuation keeps skipping higher. This is a defense tech company. And the future wars are going to be fought with these sorts of technologies. It's not going to be like the idea that China has a billion...
one or three or whatever people and they could just throw millions and millions of people at a war. That's not happening. You know what I mean? We're not having land wars like that anymore. It's going to be technological.
Hi.
Hi, everyone. It's Guy. You've probably heard me mention Current before on the pod, but we can't stop talking about how easy managing all your money is with their app. It helps you grow your savings, you can build your credit, and it works for the entire family. Plus, with their new Paycheck Advance feature, you can get up to $500 before payday when you switch your payroll and qualify. No credit checks or mandatory fees required.
It's super easy to sign up. Just go to Current.com slash OK. O-K-A-Y. That's Current.com OK to get the app.
For more information.
I want to just kind of like take some stock here, literally take some stock. So Microsoft got hit hard on Azure growth rate that was literally like in line. You know, the same thing happened with Google. The same thing really happened with Amazon. You know, their growth rate is slower, but they're coming off of, you know, a much higher base and the like here. So the way I see these companies and let's just throw them at it in there. So is it like bifurcated between the companies who are doing this spend and the companies
They're hosting, but they're also using the technology themselves. So let's put Meta up there. You know, they've been monetizing internally. They are not like renting out their compute. But then you have Amazon, I think, is also in the bucket that could be kind of benefiting from this spend. And then Google eventually, for some reason, is not getting that sort of pass. So talk to me a little bit about that differentiation.
- Between the hyperscalers? - Yeah, like the ones that are being perceived to be using the build out, using the tech, you know what I mean, with their own business models. - Right, I see. And like Google's really at the forefront there, right? But I think what you're alluding to is they just don't get enough credit for it.
And part of that is their own problem, right? They have the Transformers paper. Why didn't they have the chat GPT moment? They actually had the deep seek moment too, but they didn't get any credit for it. They actually had deep research before OpenAI did, weeks before, and they just don't get the credit. And part of that, and I cover Google very closely, is they just have...
issues with their messaging. And I don't know if you tried, one of the sort of great things what DeepSeek did in the reasoning was it showed the chain of thought. Did you try that out, Dan? Yeah, yeah, yeah, yeah. It was so like charming and delightful as a user because you're actually seeing like what, how AI is thinking. And it was so fun. And actually I just found out the other day that Google had that weeks before. Yeah.
I just, but it was in an experimental mode. It was difficult to find. And then when I started playing around with it, it wasn't as fun. It was a lot more like technical, like, okay, this is what the user wants. Not like, oh, I think the user wants this. Let me think about a joke. Let me do this and that. And that was sort of, so when I was covering Google earnings too, I said, what's, what's the difference here?
Sundar Pichai, the CEO, has said in an internal meeting that he wants Gemini adopted by half a billion people by the end of 2025. It's kind of amazing that it's not there given Google's distribution. OpenAI, JatGBT is somewhere at like 300 million users. And when you look at Gemini in the App Store, it's consistently not even in the top 10. So what is technological breakthrough without adoption?
It's not much. And that's sort of Google's problem two years ago. And it's still its problem. It's doing amazing things on the technology front, but just not really getting people to adopt it and use it. Well, it has become a really crowded sort of situation. I don't know if you saw the OpenAI ad on the Super Bowl.
last night. It was interesting because I did not do a great job of explaining what it could do to the lay people out there. Cause you and I've talked about this a lot over the last six months or, or so it's like, we're right in the middle of this. And I know, you know, there's plenty of people in Silicon Valley been using these tools for a couple of years as soon as they came out. Right.
A lot of people on Wall Street are starting to adopt them. I'm hearing like really interesting use cases for analysts and PMs. And, you know, one of our great partners, FactSet, they launched something, I want to say six or nine months ago, it's called Transcript AI. So this would be for you if you use FactSet and you're sitting there, you know, on the desk right there where you are at one market and
all these earnings are kind of flying around. You could take these transcripts and attach it to this API. And basically, it's summarizing these transcripts in real time. And there's just a lot of really interesting use cases. But again, these are for professionals. And I don't even think they have widespread enterprise adoption just yet. These are like early movers. You know what I mean? Does that make some sense to you? And the rest of the people haven't gotten there yet.
And even the whole idea of naming all these different models, like why do we have 01, 03, and Gemini Flash, Flash 2.0, Flash Mini, there's a naming problem. And I'm so glad you brought up the Super Bowl commercial because it's been a big topic of debate in our newsroom this morning. Most of us didn't like it, but I know a lot of people did. And I agree, it was sort of like a missed opportunity. Tell people what
what ChatGPT can do in your everyday life. And like, I obviously, I use it a lot during the day for when I'm processing information, but I use it at home to like,
tell stories to my kids before bedtime. I don't have to think of it as much anymore. I can give them a few parameters. And I have to say, like my daughter loves the stories that we do through chat GPT. You can use it for gardening, right? When do I cut this? What kind of flower is this? There were so many use cases that would touch, you know, on the consumer needs that I
felt like they didn't show. Yeah. And it's interesting, like two years on, I think the big focus, remember, was just hallucinations. How can you trust this? Yeah. You know, like, and now here we are, people like you and me are using it on a daily basis and we do trust it for whatever that's worth. You know, one of the things I think is interesting about, you just mentioned all these different models and how confusing they are. Last night I was at a Super Bowl party hosted by a friend of mine who's like a tech
founder here in New York City, a bunch of his guys and gals who work with him there. We were chatting about a bunch of this stuff. And a lot of them are just using these two, like the general GPT and then whatever the $200 a month was. And
It's just really confusing. Like, and you know, the same thing with Gemini, you know, and somebody's, and I was like, well guys, why for the lower end version, why wouldn't you just use perplexity and get choose any of those models for 20 bucks rather than paying 20 to Gemini, 20 to GPT. But a lot of it depends. You know, if you are a Google mail person,
Gmail, I just sounded like just aged myself, a Gmail user. And they have seven properties with over a billion users. Sooner or later, Sundar is right. They will get to that half a billion, and then they're going to just get embedded into all those different products and services.
Right now, it's still very confusing. And I was just interested to hear how these people are using it in their day-to-day life. We know that, you know, engineers coding, you know, is one of the very good early use cases of these sorts of things. But it is confusing. And I think Bloomberg had an article out today talking about that very thing. And so there's a marketing issue here. And so you might want to bet on the folks who are really good at messaging too.
Right. And that's why I'm just surprised by the Super Bowl commercial. They had a chance to say, here's how you use it in your everyday life. All right. So really quickly, just to kind of put a kind of a bow on this is like I think about Google. I think they have a lot of potential. I think in the near term, they're going to be in the penalty box. Microsoft is clearly in the penalty box. And I think it probably stays there. This is I'm talking about the stocks here. I think Meta is the one that everyone, at least in the Vestor class, has really coalesced around. And then Amazon's like right behind that.
Well, can I say there's really good reason to be excited about Meta. We talked about distillation, but the other major sort of
I don't know if it's a breakthrough, but like thing that DeepSeek brought to the forefront is the value of open source models. And Meta, Mark Zuckerberg has long been talking about this and making a bet. And they're just so well positioned in this world where open source models are being used and integrated and appreciated. I mean, they're cheaper. Why would an organization pay for open AI models?
APIs that are closed versus something that's open, that's free to them. I mean, there's there's it's complicated. They have to lay a lot of groundwork to do that. But that's ultimately where we may be heading and where a lot of the folks I talk to say we're heading is this new open source order. Yeah. But by the way, no one uses Meta AI unless you have those tricks. It's not good. Like you and I do. All right.
Real quickly, I'd love to get one-minute take on Uber here. So just going back to, I think it was October, Uber broke out to a new all-time high the day after Tesla's robo-taxi event. A lot of folks went away from that, not particularly impressed. They didn't see...
You know, they were not taking the under on when these fleets were going to be deployed. And, you know, what does that mean for Uber? Well, they still have drivers. They don't have their own self-driving sort of capabilities anymore. They partnered with Waymo to kind of get in the robo taxi game. You know, and the stock since then, you know, since people came around to some of the initiatives as it related to Uber,
You know, RoboTaxi or more deregulation given Elon Musk's proximity to Trump. So getting rid of some of the regulation that's holding them back on full self-driving and then comes RoboTaxi. So the stock is up a lot, 20% over the last few trading days. Bill Ackman of Pershing Squared took a 30 million share stake, by the way. He thinks it's undervalued. He loves Dara, you know, like...
He's probably taking the over on robo taxi too. You've covered this stock since pre IPO thoughts on a activist taking a role. I mean, it makes a lot of sense. There's always been this bull case around Uber that robo taxis are not going to, you know, scale fast enough. You're still going to need some kind of hybrid and Uber with the
Massive distribution is so well positioned to just be that platform for both robo-taxis and human drivers. It's never going to be either or. You're going to have a world in which you have both. I don't know. I'm like...
I live in San Francisco. I take Waymos all the time. I haven't used Tesla FSD in a while, but I've heard people say incredible things about the latest version, say it's really, really good. My own brother says that he exclusively relies on it. When he gets in his car, he just uses, turns it on and it brings him to where he needs to go. I can't imagine myself living in a world where I choose things.
a human Uber driver over a Waymo. And I don't know how quickly that happens, but it feels to me like that is a tough proposition for Uber in the long run, but maybe there's still a lot of time until we get there. But again, living in San Francisco,
I've incorporated Waymo into my daily life. I send my kids in Waymos, which is something I wouldn't think about doing in an Uber. So, you know, San Francisco is a very different market than New York and it's going to take
I don't know how long for you to see Waymos at the rate that I see them. Yeah, no, I love Waymo. I used it a bunch when I was there in September and I would not choose to have an Uber, you know, with a driver anytime soon. And just really quickly, the last thing I was going to hit and, you know, Lyft is up about 5% today. They made an announcement that they had partnered with
Mobileye. They're going to have self-driving cars in Dallas at some point in 2026. And the only thing I'll say is people have kind of left this name for dead. It has a $5 billion enterprise value. Just think of that. It's got a really good little balance sheet there. And if they have any success in one of these markets where Waymo or RoboTaxi is not there, I mean, this thing is a double and then a double again. And I just mean that from the standpoint of just sentiment also.
And I think that's a bull case for both Uber and Lyft is that if you have lots of competitors, if there's a world in which this isn't just Waymo and Tesla, but you have Mobileye, you have Chinese competitors, which are testing on the roads here, then you may need a platform to be able to reach all of them. And that's where I think Lyft continues to keep its smaller rival status, niche rival status.
Yeah, which has been the case, I guess, since its existence. So we're rooting for you, Lyft. Listen, Deirdre Bosa, she is the host of CNBC's Tech Check. You made me smarter here today. I appreciate you joining us and I hope we'll do it again really soon. Yes. Thank you so much for having me. Thanks.