The Rundown - Will Amazon, Google, & Uber Withstand the AI Market Backlash?

Episode Date: February 8, 2026

We break down the latest earnings from Amazon, Google, and Uber with Jason Helfstein, Head of Internet Research at Oppenheimer & Co. Helfstein explains the impact of how Google & Amazon's increased AI investments are shaping company performance and whether Big Tech can withstand growing scrutiny around AI returns.

Welcome back to the rundown, interview edition. Today, we are talking to Jason Helstein, the head of internet research at Oppenheimer. Jason covers many big tech stocks, including Amazon and Google, which both reported earnings this week. So in today's conversation, we talked about those two companies, why they sold off despite solid earnings, why investors are worried about the massive capex spending, and if any of these companies might cut back on spending. I really enjoyed Jason's perspective. I think you guys will as well. All right. Let's get into it.

I want to start with the big tech earnings, of course. You know, we had Google report this week, Amazon report this week, Microsoft last week. And there's been a similar narrative here where these big tech companies report solid numbers, but yet the stock sells off and a lot of it has to do with the heavy Cappex spending. So I want to turn it over to you. Is Cappex just not cool anymore, Jason? Is that what's going on? Like is the market just having sticker shock right now?

I mean, some of it is a sticker shop. The absolute numbers are just so much higher than what people expect it. I mean, you're talking about doubling cap-x on a base of, you know, 100 million. Like, 200 is like the new black, basically. But the problem is that the payoff is two, maybe three years down the road. And that's what investors are struggling with. It's not that, hey, you know, if Google says we're going to spend money, we can get a return on it via search or via a,

you know, a cloud, it's when, right? And then Amazon, it's even further when. So really, that is the issue. And we can talk about some companies, you can make an argument that the payoff comes sooner than later. But really, I think it's the duration risk. And by the time the duration happens, by the time we get there, will the world need this much AI capacity? And I think that's, that's like one of the questions that people are asking themselves. So why is that the concern now? because this was always the case even last year. And I feel like over the summer, you know, last year these tech companies could announce the biggest KAPX number in the stock would rally.

In fact, the bigger, the better. But now all of a sudden is the market just kind of like, well, wait a minute. Why is the market waking up to like the realities and like wanting an ROI now? Why is there, why do you think, what do you think is driving the sentiment shift? It's because the numbers are getting so big. Like when you're using all of your free cash flow for two years, like literally Amazon will use all of their free cash that they generate from the business, right, very profitable business for the next few years to fund the payoff in 27 and 28, that is eye-opening, right?

So now, you know, they have the cash to do it, right? They don't really need to be dependent on the outside financing markets, whereas, like, you're seeing it, other companies who aren't dependent on the outside financing markets are kind of running its issues. So I think it was, it went from, okay, you have a whole lot of cash, you're going to spend a good chunk of that and like half your free, you know, half your free cash flow, like you're going to basically spend all of your free cash flow and dig into your cash reserves pretty significantly for something that we won't necessarily see the payoff

for, you know, two years plus. Well, let's start with Amazon since you brought them up. So I think they, um, they're reporting $200 billion in capex, which is higher than all the other Mac 7 companies. and their cash from last year was like, what, 120, $130 billion. So they're going to have to actually maybe borrow money to make up that gap between the cash they generate from their operations and how much they're expected to spend in 2026. Is that what's driving your concerns as well for Amazon?

Amazon, we're still outperform. We did take our price side I get down. But that was mostly like, you know, our current price target on the stock right now is 260. and, you know, kind of, you know, under $200, that's kind of plenty of upside. So, yeah, it's kind of unrelated. I guess what I'll say is yes. Like, so when you add the higher capax, you have to depreciate it, right? So you're like, okay, we're going to spend $200 billion next year.

That $200 billion, let's say you have to take a sixth of that, your expenses every year for six years. So that depresses your margins until you start to see the positive revenue from that. But that revenue doesn't show up for, you know, they would say 18 to 30. 36 months. I would like to think of more of 18 to 24, but the point is that is like far in the future of when investors like to go, hey, what's a company worth on next year's earnings, right? You have to start to say like, well, what's it worth on 27 or 2028 to factor that in? So that's really what it is. And I'll say like when I look at where the street estimates came or came out this morning as far as the institutional analyst, you know, they're looking for something like 20%

AWS growth the next two years, we're at 30%, right? And like, to me, this is just math. Like, you go, this is how much you're spending. This is how many gigawatts you get. This is a revenue per gigawatt. Yeah, like the price, you know, the pricing could get weaker, the market gets the market better. But like within a range, you get the numbers. And it's almost like the the analysts are just struggling to, like, they're afraid of letting their numbers go up. And then the, the by side, the institutional buyers are like, hey, like, we generally have a with these like kind of spending cycles but like what you guys don't agree with us so listen I think this is the kind of thing where everyone's going to look at everybody's numbers over the next few weeks

and be like wait a second if there are folks that are like kind of comfortable and my clients are saying yeah like yeah if they spend 200 billion dollars like why wouldn't they grow 30 percent albeit it's going to take time to get there that you'll see numbers go up so I think Amazon could self-correct I think is a stock pretty easily within the next few weeks because the core retail business is doing great, double-digit retail, the margins are fine. You know, listen, on the near-term AWS, they accelerated four points. They gained more dollar share sequentially than Azure the last two quarters, which was not the case for the prior three, four quarters. And then they did gain basically the similar dollar share as Google Cloud, which we'll talk about because that was

actually quite bullish for Google. But yeah, we think Amazon as a stock should be fine. Just I think people need more time than overnight to process the numbers because it's quite complicated. You know what my theory is on this? So you're right. Like I was surprised to see the reaction in the stock today because of like the acceleration in AWS. I'm like, wow, the fact that they're accelerating growing at the fastest rate in three years, very impressive.

I feel like it might have to do with Andy Jassy. I feel like maybe the street just isn't as confident with Andy Jassy leading the ship. He doesn't have the same level of aura as like Bezos. obviously, or even some of the other guys, Sundar at Google and Satya at Microsoft. Maybe that's silly, but I really think there's something about Jassy. He's like an unproven leader in this situation, and they just, the street might not trust him. So I'm going to kind of disagree in that. Bezos was not on the earnings call for many years.

Actually, prior to Jassy, kind of becoming CEO, was just the CFO who did the earnings call. And you could say it was a little more like just kind of question. in an answer and like very technical we ask model questions or this this is whereas now they're kind of it's more let's put a story out there and then we'll take a limited amount of questions and you know it so so yeah I would say look jassy is the person who created a ws so like which you know he created one of the greatest businesses to ever exist so I think you can't take that away from him and when he became CEO they were already overspending building out the fulfillment centers kind of in the middle of COVID when they kind of got false signals that, like,

they thought the, the commerce business was going to be much bigger because the COVID pull forward when, in fact, like, oh, people will go actually back to buying certain things in stores and et cetera. So, yeah, like, I actually, I think it's not that. And I will say on, you know, as far as, you know, Sundar at Google, there's a period of time people were like calling for his head. So that's true. Listen, it's easy to blame his CEO. You go through these ways. I think at the end of the day, Jassy will vindicate himself. you know look I think he's I think he's being legitimate like he's literally on the call last night look

this 200 million is 200 billion isn't like you know like blue sky like 200% we literally can see the exact return on investment now he's not telling the street what it is right and it probably wouldn't be a bad idea for amazon to host like an analyst day and literally like lay out like here's the five year plan here's a return on investment now maybe they don't want to competitors to see it it's been many years

Oh, they don't know. No, I think it's wrong. I think they absolutely do know. Now, look, they don't know, like, does it take 18 months to get a data center out? Like, maybe it used to. Now it's 24 months. Maybe the word is going to go to 36 just because of all the backlogs. Maybe pricing comes down.

So there's timing. I don't think they know quarter to quarter. But I think if you're looking at over the next four to five years, I, you know, unless, like, everybody's, like, way wrong on the adoption of AI, I think they actually know with pretty high certainty. I'm going to be watching that for sure. But yeah, I just found that interesting where Amazon was the worst performing Mac 7 stock over the last 12 months. And I'm just trying to make sense of that. And that was a silly theory that I came up with. Well, I will say, listen.

They did. Here's the rub for Jazzy. He was late to recognizing the importance of AI. So Google was always first, right? They were like, they created the transformer model. They were always like at the leading edge of machine learning. You know, Microsoft like smartly listened. to, you know, Sam Alton.

He's like, you need more capacity for me. They're like, okay. Like, now Microsoft could say, like, they saw it too, but like I'm going to kind of believe that like, ultimately Microsoft had a deal with Open AI and they kind of trusted his signal they were providing. And Amazon kind of just didn't see it. And so they had this like very successful,

we'll call it Legacy Cloud AWS business that wasn't necessarily based on GPUs and AI. And it's doing great. And yes, competition like had been increasing. and they were kind of doing their best to fend off Azure and fend off Google Cloud. But then AI hit, and literally they could not move fast enough or spend the money fast enough to catch up. But now, you can make an argument like you could see them catching up. And so they've kind of made up for like some of their past mistakes.

Well, let's pivot from Amazon to talk about a company that seems to be the leader now when it comes to AI, which is Google. You recently upgraded the price target for Google after their earnings. And so I want to start there. What do you think makes Google's AI story different than Amazon's story? Because very similar, I mean, obviously Google's cloud is going faster, which we're going to talk about. But how do you see those two companies be different? Sure. So, I mean, Google basically is seeing the quickest return on investment because they basically have, we'll call it two.

And I'll call it two and ultimately it'll be three different businesses, right? So you have a search business that relies on compute to do a better job answering your search questions and matching the advertising. Then you have the Google Cloud business that provides the same compute that businesses can buy that Amazon, AWS, and Microsoft as yourself, right? And then the third stool is going to be AI, right? It's going to be Gemini. And the idea of it, that's both a model that enterprise customers can use and then ultimately a consumer tool and we'll talk about it. but they should be in the pole position to have, like, the best consumer assistant. Everyone got all freaked out and excited about Claudebot.

Now we're calling it something else. Open claw, open claw, right? Because we don't want to, you know, we want to violate an Anthropics trademark there. But ultimately, Google is alphabet is in the best position to ultimately provide that to the consumer in kind of an easy and safe way. So, but what that allows them to do is move their compute from their different businesses. Right. Whereas, you know, effectively, Amazon has two businesses, but you can argue like how, you know, compute intensive is retail. It used to be very, you know, relative it was in the past more, but now it's a very small percent of their compute, right? Whereas like meta has one business, right? But Google kind of has two, kind of going to three and then potentially going to four, right? Like if you separate Gemini between consumer and enterprise. And so as a result of that, again, you saw on a dollar basis, GCP, Google Cloud, actually generated the same increase sequential in dollars as Amazon did, which is pretty amazing, and more than Microsoft, right? So you could say that, you know,

okay, Google Cloud is not taking share from Microsoft in Cloud. On top of that, you had search accelerate, and then they're, you know, what they're spending relative to the free cash they generate is, like, not as impactful, right? They still have positive. free cash, like we're still forecasting in the next two years, that they're going to generate, you know, positive free cash for share, whereas the other companies are going to eat up all of their free cash to fund. So, yeah, so you can make an argument that right now, Google right now is kind of in the best near-term position. That's a very interesting point that you made, though, where like the more and more compute that Google brings on, all the investment that Google makes,

saying that they can deploy it across, like their entire business, whether that's for YouTube, whether that's for search, whether that's for cloud. And like they can see a return on that investment way quicker because of that instead of having to wait for, you know, demand or whatever. Well, you could argue they've been constantly building too, right? Like they, you know, again, the biggest rep to Amazon is they just almost, they weren't leaning into building centers fast enough and they kind of got it. It just took them a year to kind of fix it.

whereas Google never really stopped. And then again, their business is fungible. Like they can kind of maybe you move search, compute away from countries of the world that maybe don't monetize particularly well because you want to allocate it to Gemini corporate customers who are willing to pay big bucks right now. That's a great point.

Now, speaking of search, like you mentioned, search was up, was it 17%, which is wild. Like search is accelerating, right? And I mean, not too long ago, nine months ago, 12 months ago, everyone's like, oh, search is code, Google's code, they're dead. Chad GPT is going to take all the search volumes. The fact that search is accelerating is like mind blowing.

Do you ever see that slowing down? Do you still see the search business to be a threat or to be under threat at all in the mid to long term? I mean, look, the search business needs to become the, you know, chatbotogenic business, right? Like, it's a transition. So the idea is that, you know, okay, most of the revenue, you have 70% almost all of it is like coming from what we consider traditional search, right? You go to Google, you ask a question, there's a blue link, right? More and more people are then, like, they're looking at AI overviews, you're just starting to see some ads, and then over time, and then it's AI mode, right? Ultimately, search becomes AI mode, right?

Like, at some point, you go to Google and literally it is AI mode, right? And then it's like, okay, how they'll, like, you'll find video or different things. We'll see. But like, so the question is how successfully can they manage that transition? So from blue link search to AI mode. And then Gemini will, you know, we'll see like, maybe they're going to call AI mode Gemini at some point. You'll rebrand that. And then it'll be like, okay, I could have Gemini's an app on my phone versus Gemini's search.

But like you will literally just, it'll be, instead of saying Google it, you'll Gemini it. Right? Like can they make that transition? So that's from an ad standpoint that that that that that's the key over the next few years. Well, I mean, I think they're doing a good job of like showing better results in this in the AI overview. I use it all the time now. But it but that can they make it as profitable right? Like the search business is the greatest. Oh, I think there's no question they can make it as profitable. Can you talk more about that? Because they've told you right. They've said that users who use AI mode they ask their look you know, they put in longer questions right. So like I can tune your answer better. And then they ask follow-up questions. So, like, if you're,

especially if you're thinking like searching for, like, a product or a vacation or a service, like, ultimately, if they can match you with the best service at the bottom, that click that brings you to service is that much more valuable to the advertiser because you have a higher chance of converting. And therefore, they can charge them more. And they've already trained advertisers to think about their business this way, right? Like, that's how performance advertising. So, yeah, I think it's and I mean they've given you stats right they basically talked about the point is that they said on their earnings call that everything around AI is driving more search engagement and that's positive yeah I mean that's true and so I mean I think I think it was that but also like the growth in Gemini right like the growth in Gemini just is just insane what 750 million monthly users now that they're still behind chat GPT but yeah yeah we all look at at the market share, right, and they've been gaining share against ChatGBT, GBT, and I think that's probably, like, say from a near-turn stock perspective, that's probably the biggest,

like, factor. That, you know, if ChatGBTDB launches a new model, which then they put into OpenA launch a new model, they put into ChatGBT, and then they start to show up more of a positive inflection in users, user share. That will hurt Google. But, I mean, do keep in mind on Android, those numbers are heavily influenced by Android because, you know, they do a good job of like getting you to open up Android Gemini on Android more. We're like, it would be very interesting like you look at the data on Apple, right, where you have to like more proactively download Gemini onto like Apple products, which they're seeing good traction.

But that to me is more indicative of the success. That's a good point. But also like Google has the advantage compared to chat GPT where like they have this money printer ad business, right? And so I wonder how long they're going to keep Gemini ad-free to continue to take market share from ChatGPT because ChatGPT has said they're bringing ads to their platform. So then maybe Google will kind of delay their ad strategy on Gemini to try to continue to gain market share. Well, I think that's right. I mean, because they don't need to, right? Like, it's first they're not even fully penetrator with ads on AI overviews.

Then it's get ads to AI mode. Then maybe you move the whole thing to AI mode, right? And then at some point, yeah, it'll all be Gemini. But the point is, like, there's no rush. They don't need the money. Again, 17% growth. They're doing just fine. Whereas chat, G, be, you know, open AI.

Like, I think investors want to see, like, where does the revenue come from, right? And so. I want to, I want to zoom out a little bit and talk more big picture. So we've seen, we've seen these companies accelerate CAPEX. The market reaction hasn't been so great this time. Do you think it's possible some of these big tech, uh, hypers scaleers scale? scale back their K-X at all. Maybe Q2, Q3, they come out and say, you know what?

We're going to cut 10%, 20% than we expect it. You don't think that's going to happen at all? No, it can't. I mean, like, they have to let it. They need to let it. Well, the only reason the numbers, remember, they got annual and we guess that they quarterly impact. So it's entirely possible supply chain delays,

like, you know, maybe the quarterly build is slower because, like, you didn't get a data center to bring on, right? Like, you couldn't get the power. etc so it's entirely possible that like maybe the cadence but like i i don't see any of these company the only company who could maybe do an about face would be meta whereas if they launched whatever the newest version of their frontier model is and if it kind of falls flat and everyone's going to be like why do you need to spend this much money if you can't produce like a frontier model why don't you just use claude or use javanagh or you know you know etc so they would be the only

ones in my opinion that Cole pulled back the rest of the companies because they are, again, are running data center or hyperscaler businesses to sell tokens. Like, they're not going to pull back in the next world. They're not going to get, the CFOs, CEOs aren't going to get a little queasy from the stock price correcting 15, 20, 20, 20. They have, like, there is a spreadsheet. And each of these companies, they were literally like for every dollar we put in CAPEX. This is what the profit is in two to three years.

Like within like a very high degree of certainty. Now, again, the only reason that won't be there. is if three years from now, the amount of usage of AI compute ends up being way less than what everyone forecast when everyone's just talking about there's an absolute shortage of capacity right now. Okay, gotcha. I mean, that's, I think that to me is like the biggest, everyone says that. Like, yeah, there's no way they're going to scale back CAPEX. And I want to believe that as well. But I mean, I just, you know, people can do weird things, especially when like the market start turning. So it'll be interesting to see. see what happens. Do you have a couple minutes for an Uber question? Sure. Now we can, we can have it. Okay. So you cover Uber as well. They recently reported and you downgraded the stock, right? Or did you cut price target? Price target. You reduce the price target like $15.

So you cut the price target. But you still, yeah, yeah, you still have an outperform rating on it. I'm looking at it right here. So I just want to kind of talk through Uber, you know, a lot of people think that they're going to be disrupted by the future of robotaxies and all in, and, you know, with Waymo expanding like crazy. So I just want to turn it over to you. Are you concerned about that when it comes to Uber being disrupted? Or I guess I'll simply put, what is it about Uber's business that makes you the most nervous? And what are you most excited about?

Sure. So the reason we cut our price to our target was we had to take our profit estimates down. So the reality is like, you know, the business is not running as profitable as we had previously thought. Most of it has to do with them leaning into Uber one membership. growth and just other promotions or drive usage. And so, I mean, it did pay off. Uber. They had 46 million Uber members.

Uber 1 up 55%. I think that's probably faster than any other subscription models probably growing right now. So, yeah, Uber 1, completely working. And then the way they generally get people to sign on is they largely target you through food delivery. And once they get you to like sign off. That's how they got me.

You can get like deals on rides, free upgrades, cap surging, et cetera, et et cetera, and then they're trying to change user behavior. But, yeah, so that was specifically the numbers. But look, on Robotaxy, we call that a terminal risk. Like, it's far in the future, but everyone acknowledges Robotaxy will happen. The big debate will be, do you have one or two companies who run away with the technology? Does it become commoditized? And you're going to have 10, 20 companies, every automaker will be able to produce a self-driving car at some point in future.

and therefore if you can aggregate all of that supply to the consumer, a Uber, you end up doing well. And like what we already know today is that in markets where Waymo is on Uber, it's way more profitable like on a per car than if Waymo just tries to use it itself. Because it can't kind of fill all the capacity. And even as simple as like, okay, like would you have it do food deliveries or something? Like if there's no demand for rides. and okay, but like, you know, and again, they're DoorDash, they could go work with them and try, I think they are, right? But like the idea of like if all the technologies in one place, does Uber ultimately pull that off seamlessly better than anybody else? That's kind of the debate.

And I think, you know, Uber has said they will be in 15 cities globally with Robotaxy. Again, with, you know, they're obviously not partnering with Tesla, but with Waymo and then kind of call it, you know, multiple other parties. I think they just announced this morning. We ride in the Middle East. They were going live with, I think, was 1,500 cars, I think, over 1,000. I don't remember the exact number. So it's coming. So I think Uber is a kind of fact where if you're patient, you will get paid.

The stock is, you know, inexpensive on current valuation. The business is generating out of cash flow. It's growing very nice. They basically said U.S. growth will accelerate mostly because they're getting a reprieve on insurance rates in California. you and they can then pass that on by lower prices. And they know when you lower the price for Uber rides, more people use it. But yeah, like, look, we fully expect you're going to get more positive announcements at a Waymo, more positive announcements at a Tesla. You'll probably see some foreign companies try to bring cars here as well.

But ultimately, we think robot taxi will end up being a commodity. Like, it's not going to be controlled by one or two companies. It's going to be, like, a ubiquitous technology. it's just taking the later entrance longer because they don't have the learnings. I mean, look at it. Waymore started first. They right now have the perceived safest product, right? Tesla was like second.

They narrow the perceived to have the second, right? And so just give it time. I think Uber is just a stock where you give a time you'll make money. Yeah. And if it does become a commodity, then Uber has. I mean, they control the demand. Everyone's going to get on there. And then you're not going to go to check four apps to get a ride.

Like if you're leaving a sporting event. There's something that's busy. And you're not going to go check four apps and then price check. You're just not going to do that, right? You're going to ultimately go to whatever app that you trust is giving you the fastest, cheapest ride. I also feel like, like, Robotaxies feel in perfectly when it comes to like, you know, usually when it comes to ride demand, there's always a surge, whether it's a surge,

you know, in a certain time of day or at a concert or sports event. So like Uber kind of can fill in that surge by with the Robotaxies. And like you said, that probably results in the highest. profitability of these robotaxies like they're seeing in the markets where Waymo is on Uber, right? They're probably seeing the highest return. So it's a very interesting company to watch. And I look forward to your research on it. So, Jason, thank you again so much for coming on talking.

And, you know, hopefully get a little bit break now now that some of the major earnings. Midcap next week. It's not over. The quarter earnings season lasts about six weeks. That's true. So I guess we get happy. We're not all within four or five days. But, yeah, the earnings is not over.

So more to call. I know, I love it. And I appreciate your commentary on it. Thanks again for coming on. Sure, thanks. How good one. Well, all right, guys, hope you enjoyed that conversation with Jason Helvstein. You know, my favorite part was how Jason broke down how Google is likely to see a quicker ROI on their AI investment because of all the businesses that they have.

I also like this take on Uber and what happens to them in a future of Robotaxies. Let me know the comments on Spotify and YouTube, what you thought about today's conversation. And while you're at it, consider giving us a five-star rating on Spotify. Spotify, YouTube, Apple, wherever you listen to your podcast, all that engagement really does help us out. And it helps other people find the show. Thank you guys so much for listening, watching, and commenting. Shout out to Mike and Connor for all the work behind the scenes. And we'll see you guys back here tomorrow.

Frozen lasagna, medium power, 15 minutes. Sounds like Ojo time. Let's play. Feel the fun with Playojo. The online casino with all the latest slot and live. Casino games. What you win is yours to keep with no wagering requirements. Instant payouts and no minimum

withdraws. Hey, I just won. Woohoo. Do the fun. Play Ojo. Honey, forget about the lasagna. Let's celebrate. 19 plus Ontario only. Please play responsibly. Concern about your gambling or that of someone close to you. Call 16-531-2600 or visit conxonnetsoot.ca. The Madamy Holmes bike for brain health supporting Baycrest's returns on May 31st for its fifth anniversary with a new start and finish at the Aga Khan

Museum. Join thousands of cyclists as we take over the DVP and Gardner Expressway in support of dementia research and brain health. Riders of all abilities are welcome, and both regular bikes and e-bikes can participate. Bring your friends, family, or corporate team, and make an impact. Register today at fightforbrainhealth.ca.