In this podcast, Motley Fool analyst Sanmeet Deo and host Mary Long discuss:
- Poor results from Tesla's automotive segment.
- Whether Elon Musk turning more attention to the company can revive it.
- Half-marathons, and the future of humanoids.
Then, Motley Fool analyst Asit Sharma joins Mary for a look at AMD and how the chip company is different from its biggest competitor.
Where to invest $1,000 right now? Our analyst team just revealed what they believe are the 10 best stocks to buy right now. Learn More »
To catch full episodes of all The Motley Fool's free podcasts, check out our podcast center. When you're ready to invest, check out this top 10 list of stocks to buy.
A full transcript is below.
Don’t miss this second chance at a potentially lucrative opportunity
Ever feel like you missed the boat in buying the most successful stocks? Then you’ll want to hear this.
On rare occasions, our expert team of analysts issues a “Double Down” stock recommendation for companies that they think are about to pop. If you’re worried you’ve already missed your chance to invest, now is the best time to buy before it’s too late. And the numbers speak for themselves:
- Nvidia: if you invested $1,000 when we doubled down in 2009, you’d have $287,877!*
- Apple: if you invested $1,000 when we doubled down in 2008, you’d have $39,678!*
- Netflix: if you invested $1,000 when we doubled down in 2004, you’d have $594,046!*
Right now, we’re issuing “Double Down” alerts for three incredible companies, available when you join Stock Advisor, and there may not be another chance like this anytime soon.
See the 3 stocks »
*Stock Advisor returns as of April 28, 2025
This video was recorded on April 23, 2025
Mary Long: The robots are coming, but maybe not very quickly. You're listening to Motley Fool Money. I'm Mary Long, joined on this fine Wednesday morning by Sanmeet Deo. Sanmeet, great to see you. How are you doing?
Sanmeet Deo: Hey, nice to see you.
Mary Long: We've got one story that's going to be our single story today because there's a lot to talk about in this report, none other than Tesla. Dropped earnings yesterday after the bell. Lots of anticipation with this one. Obviously, it's a large company. It's a company led by a controversial leader. Let's put it at that. Coming into this report, we had Wedbush analyst Dan Ives. He's a longtime Tesla Bull, and he told NBC that this report is a ''Moment of truth for Tesla.'' We're going to dive into the details of this report in a second. As I said, it's our single, our sole story today. But let's start with the big picture idea here. You own Tesla? What truth was revealed in this report?
Sanmeet Deo: I think the truth is that the automotive segment is hitting the brakes. The one number that can symbolize everything that's happening for their segment this quarter was the 2.1% operating margin, which was significantly lower than last year's 5.5%. The whole story is really captured in that margin number. It's lower average selling prices for vehicles, lower delivery volumes, volume time price, lower revenues, and higher R&D expenses. That margin has significantly is lower than what they've had in really over the past few quarters, so very concerning in that sense. Now, energy and storage, and services came in very strong. That was great, but they're much smaller part of their revenue. The question is have they taken the eye off the ball? Is competition hitting them harder than the market suspects? The one other truth I'll say is that the market liked Musk's comment, which I think we're going to talk about, about reducing his time with DOGE and getting back to focusing on Tesla.
Mary Long: Let's hone in on that piece, because in spite of that comment seems to be what is causing this rise in Tesla stock that we're seeing this morning. We also saw a rise after hours yesterday, pre-market this morning, and that's only continued throughout today. But again, you just walked through the earnings. There were some glimmers there in other segments, but this automotive segment, as you said, largely was hitting the brakes. It seems that the surge is largely attributable to Musk's comment that he'll be taking time away from DOGE and returning to Tesla as soon as May. I've got a question on this, but is that what you two attribute this jump to, or do you think, maybe, there's something else going on?
Sanmeet Deo: No, absolutely. This quarter, if you look at it, with automotive segments being probably 80, 85% of their revenues, this is a bad quarter for them. Their vehicle deliveries were disappointment. That we already knew. That was already reported. Profitability came in a lot lower than expected. I would anticipated on a report like this, the stock would be down. But given that Musk made a statement that he's going to focus back on Tesla, that's something that has been an overhang on the stock. Also, the market is very big today off of relief rally. That, too, is helping their stock bounce.
Mary Long: This time allocation comment is an interesting one to me because I can see, obviously, why Musk returning to Tesla could be a boost for the company. But I also wonder how much of Tesla's miss here, in this quarter, is attributable to the time that he's not spending at Tesla versus how much of it is attributable to the political associations that he's tied himself to. How do you think about that? How much of this miss, especially in the automotive segment, do you say, hey, this is a problem that's due to Musk not actually being at the helm, and that will be solved by his return to the helm or actually, this is a problem that's attributable to the political associations that Musk has made for himself.
Sanmeet Deo: I think a decent amount could be alleviated with Musk spending more time at Tesla. He's known to have a tight reign and high attention to detail when he's focused on the company. I've heard reports from people that work at Tesla that he's very detail-oriented. He's very in the weeds when it comes to the company, but that's if his attention is there. His attention has not been there, so that's that eye off the ball part where he's not allocating the time needed to really guide that ship. Some of it, too, is increasing competition, cheaper cars from China, causing some effects there. I think there's been a lot of talk about brand degradation. Tesla is a brand. It's successful a lot due to its brand. Musk political associations have rubbed people the wrong way. People may not like his associations, how much time he's spending. It has taken a hit to the brand. That's pretty noticeable in the numbers as well.
Mary Long: We talked about this morning about how Wall Street is buying up Tesla. They like this comment from Musk. But if you look at insider activity at the company, it seems that over the past 12 months, Tesla insiders have been doing the opposite. Over the past 12 months, Tesla insiders have sold 28 times and bought zero times. You like to pay attention to insider activity here at The Fool. What do you make of this? Is this a red flag, yellow flag, or something that you can put an asterisk by and justify somehow?
Sanmeet Deo: I think there's no flag on the play, honestly. All these sales were part of a planned or pre-arranged stock option exercise strategy. I like to look at open market buys and open market sales when it comes to insider buying or transactions, and none of the ones I saw were really open market sales. Although there was one open market sale in the past six months from Elon's brother, Kimbal Musk, for 75,000 shares totaling $25.6 million. Maybe he's buying a new house? I don't know. That in itself could possibly be a yellow flag, but all the others I'm not too concerned about.
Mary Long: If he's buying a new house with $25.5 million, I want to see that house. [laughs]
Sanmeet Deo: Absolutely.
Mary Long: Tesla's all-time high was last hit on December 17th when it closed at nearly $480. Today, it's closer to 250, again, it's moving up, so that might change by the end of the day, but that's where it is right around the time we're recording. Breakfast News, which is our daily newsletter here at The Fool. It gives a rundown of daily market happenings. They asked readers this morning in the newsletter when they think Tesla will return to its all-time high, if ever. I'll pose that question to you before we dive into more of the details of this report. When do you think Tesla will hit its all-time high again, if ever? How do you think it gets there?
Sanmeet Deo: I think it's going to hit the all-time on April 23rd, 2030. Now, I'm kidding. [laughs] I think it could be at least five plus years or so, something like that. Usually, when we see these huge massive market corrections, what I've noticed is whether it be the market or certain stocks, they hit highs, they correct heavily, and then it takes a long time to hit that all-time high again at some point. That's assuming the businesses continue to succeed and do well. In order for them to get there, the automotive segment needs to gain its growth momentum, and we're going to talk a little bit about that later, too, about how they could do that. Some positive traction on the fully autonomous driving humanoids, which we'll talk about, too. That could really boost the enthusiasm for the future prospects of the company and the business, and the stock. If they can start making more traction rather than empty promises, then it could hit all-time high again.
Mary Long: The large weak spot in this report was the automotive segment. We were told during the earnings call that, ''Given economic uncertainty, resulting from changing trade policy, more affordable options are as critical as ever.'' The idea of a more affordable Tesla has been teased for a while now, though plans have remained ambiguous, elusive. Growing this segment back and gaining traction here again, a clear path to that seems to be, if you can make this affordable option a reality, that would be a great way to, again, revive this automotive segment. How do you see that playing in? Again, I've mentioned that these plans for an affordable Tesla remain ambiguous. What would you like to see that plan and practice for a more affordable Tesla actually look like?
Sanmeet Deo: I think that affordability is absolutely critical to Tesla's automotive thesis related to their electric vehicles because they're getting heavy competition from Chinese makers. Like I said before, they're producing very cheap cars. Now, whether those cars are just as cheap here in the United States versus those in their home countries is something to wonder. When I think of affordability, when it comes to cars, I think the Gold Standard Hondas and Toyotas. Those are the most affordable that are out there. You see them all over, and they're for the masses. If Tesla can create a car for the masses, I think they need to get it to around $20,000 price point because you have Hondas and Toyotas at their lower base models at around $23,000, $25,000. I think that if they can get to that price point, make it profitable, it could be huge for the automotive segment. Then you'll start seeing Teslas literally everywhere, not just for the high-end. I think, though, they need to create a clear product roadmap and what is going to look like for them to get there, because Musk has the tendency to overpromise and underdeliver, and they need to flip that script and really make it plausible that they can achieve this mass market.
Mary Long: If Tesla can develop this more affordable option, that's one way to revive its automotive segment. But if we see vehicle deliveries truly begin to flatten out, as seems to be happening in this report, what does that mean for the Tesla growth story?
Sanmeet Deo: It's going to be challenging because, again, automotives vehicles are about 80 something percent of their revenues. If that just starts to flatten, that's a majority of their business that's flattening. While energy and storage and services, and all these other great pie-in-the-sky autonomous and humanoids are great. They're not a huge core part of their business. This is vehicles are a core part of their business. If they can't make it work, their business will struggle. Now, that's not to say that autonomous and humanoids can come out of nowhere at some point down the road and make up for all those losses. That could happen, but that's still a very aggressive and far-out into the horizon prospect.
Mary Long: Then let's focus on where those other business segments are today. The energy generation and storage segment of Tesla has seen nice steady growth over the past several quarters. This is an area of the business that actually saw notable revenue gains this quarter. Where is that growth coming from?
Sanmeet Deo: They're getting a significant increase in demand for both residential power walls, grid scale, the megapac battery solutions because of things like renewable energy adoption, growing need for grid stabilization, resilience, rising energy costs. They're in a sweet spot of the market where their demand for their products are high.
Mary Long: We had this whole conversation at the top about what it means that Musk is away from Tesla, what it might mean if he returns. What's interesting to me is that we're seeing this growth in the energy segment while Musk is away, running DOGE. Is that a bright spot? Does that mean that, hey, the energy side of the business can actually effectively run itself?
Sanmeet Deo: I think the overall operations, the day-to-day, can probably do a decent job, like we've seen, because of how they've performed on a day-to-day basis without Musk, but the overall vision, strategic direction of the company. I've always thought of Musk and Tesla, you can't get into the mind of Musk, really. But he has some grand vision of how things are going to all piece together when it comes to autonomous and cars and energy, humanoids, all this stuff is going to it's probably altogether in his mind. He's probably having a hard time delivering the message to all of us. That whole vision is needed, and I think him providing that and focusing on that is going to help guide things. Day to day, they can probably do well, but whether they can scale to another level without him, I don't know if that can happen.
Mary Long: A piece of that vision that's long been teased is this idea of the robotaxi and the Cybercab. Musk said on the call that ''We remain on track for the pilot launch of robotaxi in Austin by June.'' June is right around the corner, so that feels very soon. It will be interesting to see if that is indeed something that the company can deliver on. But notably, Musk also says, the purpose-built robotaxi product Cybercab, is scheduled for volume production starting 2026. We throw these terms around a lot, robotaxi, Cybercab. What actually is the difference between the two products, and how do they work together?
Sanmeet Deo: I'm glad you're asking, because that's critical, and I always confused myself before I actually looked into it. The robotaxis, basically, they're going to utilize existing Tesla models, primarily the Model Y, to run the fully self-driving mode. Cybercabs are going to be specifically built cars for the robotaxi service. Its whole purpose is to be used as an autonomous taxi service. The robotaxi service it's a pilot program in Austin. They're going to collect data. They're going to get that experience out there, see how it operates. Then the dedicated cyber calves will come out, start being produced around 2026, which then they'll roll out at some point. On a side note, I was in Phoenix a few weeks ago, and I saw Wemos all over the place, and they're pretty wild. It's very futuristic. That's all I can really say about them.
Mary Long: I'm sure. We'll close out by touching on what I think is your favorite piece of this company, which is the humanoids. Again, this is something that we're still seeing ramp up in production, still in development largely. Musk said, though, on the earnings call that he expects to have thousands of Optimus robots working in Tesla factories by the end of the year, and that he expects to scale Optimus faster than any product he thinks in history, to get millions of units per year as soon as possible. He later clarified that timeline and said that perhaps the company could reach a million bots per year in less than five years, maybe four. Why is Optimus allegedly so easy to scale/do you buy this timeline fully, or is this another example of Musk overpromising and potentially underdelivering?
Sanmeet Deo: [laughs] I love humanoids. I'm a humanoid fan. But I think it's potentially easy to scale because of Tesla's expertise in manufacturing, AI, vertical integration from developing all these, at least the EVs and the software that they build. They can scale it. Whether this timeline is believable, I'm not buying it, because I think that, again, Musk has a tendency to overpromise and underdeliver. Could it happen in 10-20 years? Possibly. Is it going to happen in the next couple of years? Not so sure about Optimus.
Mary Long: Whenever this does happen, Musk has called out that he thinks that the humanoid robots can bring in $10 trillion in revenue for Tesla. What does your analysis say? Regardless of when these robots are actually delivered at the scale that Musk is talking about, do you see the same possibilities in terms of revenue that he does?
Sanmeet Deo: Ten trillion does sound like a wild number, probably very unachievable. But if you take a step back and think about it, there's about 128 million households in the United States. Maybe you assume each one purchases at least one. They've been rumored to be about 30,000 once they've brought the cost down to a reasonable amount. That right there is about three trillion in revenues for households, commercial side of humanoid production, could it hit seven trillion? Possibly, you have them in factories, even in businesses, even in restaurants, you have in different places. Then you have to factor in maybe parts or pairs servicing all the revenues that you get from that, as well, because it's not just going to be sales of these humanoids. It's going to be all the other ancillary stuff, too. Ten trillion wild sounds wild. Maybe not that wild.
Mary Long: Over the weekend, humanoid robots raced against actual people in a half-marathon in Beijing. The point of this isn't for humanoids to outrun humans. I saw this all over the news, and I was like, wait, really, what is the point here? It seems to me maybe more like a publicity stunt or just a test case to see, hey, how capable are these robots actually doing human actions? Well, close out on a fun question. Optimus was not in this race, but had it been, how do you think it would have stacked up?
Sanmeet Deo: I think it would have been terrible. If you've ever seen them walk, they walk really slow and really measured, and I don't even know if they can run, honestly. That was funny. The first thing I thought of when I saw that race was, why are humans trying to create something better than us? Why do we do that?
Mary Long: Well, and it misses the point of why people run marathons or half marathons in the first place. We all know that we can't hit the fastest time in the world, but we're about striving to be better and for self-improvement.
Sanmeet Deo: Exactly. I wouldn't be as impressed if a human breaks the fastest speed record versus a human doing it.
Mary Long: For sure. Sanmeet Deo, always a pleasure. Thanks for coming onto the show and for giving us a bit more insight into Tesla Earnings today.
Sanmeet Deo: Thanks, Mary.
Matt: Hello. My name is Matt.
McKinley: I am McKinley. We are the Father-Son team that brings you history dispatches.
Matt: History Dispatches is a short daily history show where we talk about topics from all over the world and all throughout history. We talk about people, places, events, and even objects.
McKinley: While anything is fair game, we have a soft spot for the weird, the wacky, and the obscure things you may have never even heard of.
Matt: Do you have any examples?
McKinley: How about Waj tech, the bear who rose to the rank of corporal in the Polish Army, or the Great Emu War? Or how about the biggest treasure taken in the history of piracy?
Matt: That sounds cool, but do you have a story about the head of Oliver Cromwell? What about the ancient Library of Alexandria? A story about the first woman to climb Mount Everest would be cool.
McKinley: Well, we got those as well. Every weekday, there's something new and fun.
Matt: Sweet. How do I get this trove of goodness?
McKinley: All you have to do is go to historydispatches.com or just look for History Dispatches in your favorite podcast app.
Mary Long: NVIDIA is not the only chip company in town. Up next, Asit Sharma joins me for a look at Advanced Micro Devices, or AMD, the company that's trying to give NVIDIA a run for its money. Asit, despite the fact that semiconductor stocks are largely cyclical, it feels like they've been in the news all the time over the past few years. One of the names that's often in the news nearly every day is NVIDIA, but a competitor that gets a little bit less time in the spotlight is AMD. The news, the media love NVIDIA, you love AMD. Maybe, help us set the table here. These are both chip stocks, but how are they different? What is AMD doing that's different than what NVIDIA is doing?
Asit Sharma: Mary, yes, to full disclosure. I do own both companies. I own AMD, I own NVIDIA. I have both recommended NVIDIA and AMD, and services I personally run, and services that I work on. I like both companies. AMD is a little bit different because it's more of a diversified player in the semiconductor ecosystem. Although I'm sure there's someone out there who's listening right now and saying, wait, NVIDIA is becoming incredibly diversified. It's branching into so many areas. But the traditional way we look at how semiconductor companies operate I think you give AMD the edge in diversification. It plays in the chip market, so it makes chips for servers. It makes embedded chips that go into industrial Internet of Things devices. It makes chips for gaming, GPUs for gaming, like NVIDIA does. It also makes GPU accelerators, which is where all the attention is focused. I didn't know you were saying. It seems like if these are such cyclical companies, why are they always in the news cycle? But I think we're going to be talking about such companies for quite a while.
Mary Long: Help me understand this. AMD's biggest customers include Microsoft, Meta, Alphabet, Sony, Oracle, big names. We know a bunch of them. The list also continues beyond those big names. NVIDIA does not publicly disclose its customer list, but it is widely believed that its biggest customers are get ready, Microsoft, Meta, Alphabet, Amazon. There's a lot of overlap between those customer lists. You mentioned that AMD is more diversified. But if I'm the person who works at one of these tech companies and I'm in charge of buying up AI chips, what's getting me to buy AMD chips rather than solely purchasing from NVIDIA?
Asit Sharma: Mary, first, we're going to make a distinction here because you said AI chips, and that signals to me that you want to talk about GPU accelerators, the kinds of chips that are used for artificial intelligence, specifically generative AI that help us use large language models and are being applied to so many different industries. If you are, let's say, a hyperscaler, like an Amazon Web Services, or a Microsoft as your or an Oracle, why would you want to buy these chips? Number 1, it decreases your sole reliance on NVIDIA, which has been leading the charge and really developing the strongest, most powerful chips for the last three years since GenAi exploded onto the scene. But also, there's a growing argument within these companies that we want to be able to offer the ability to use generative AI at a lower cost to our customers and for our own bottom lines. Hence, Oracle just placed an order for about 30,000 MI 355x. I think that's the name of the accelerator, a series from AMD, which is an order worth billions of dollars. This was disclosed just a few weeks ago in Oracle's earnings conference call. Because one of the reasons is that total cost of ownership over time for Oracle is going to be less versus buying similar GPUs from NVIDIA. There's some case where you want to buy NVIDIA's GPUs to offer that power and performance. But there's lots of places in the generative AI world for inference, the outputs of these models and for some training purposes, too, where AMD's chips are just as good for lower cost.
You name NVIDIA as being the player with the strongest, fastest chips. The general consensus is that, OK, AMD creates chips that can compete pretty well with NVIDIA, but they still have to catch up with NVIDIA from a technological standpoint. As retail investors, how can we understand the intricacies of the differences between these technologies? What does that path of catching up to NVIDIA from a technological standpoint actually look like for AMD from the outside? I think in some ways, it's becoming a little bit easier to understand than it used to be. There's one very visible thing that I think so many listeners may have heard of. One of the things that makes NVIDIA's products great not only is the GPU hardware, but it's the software libraries, collectively known as CUDA, that you get when you buy NVIDIA GPUs. Some of these come with the purchase for these big hyperscalers and even, academic institutions. Some of those have higher costs associated with them, but they make those GPUs really powerful, and that's been an edge that NVIDIA has had for a long time. Now, that's a closed system. It's NVIDIA'S own.
AMD has chosen to go another route with ROCM. This is their open-source version of accelerator libraries, which they basically invite the world to come and help improve that. That's getting better and better. One of the things that we need to see out of AMD is not just being able to be within spitting distance on the GPU side, but to have its software libraries become more powerful. To bring down that total cost of ownership, but also just to make their GPUs function at a level that NVIDIA's do. Now, there's another big picture thing for folks to watch in the coming years. NVIDIA is so ahead of the race because it's now moved on from supplying these great GPUs to supplying rack-scale systems. You and I were talking about Vera Rubin a couple of weeks ago. All these crazy names that NVIDIA has that are poetic. What this simply means is that instead of buying GPUs and making them operate, companies that are hyperscale companies or think even enterprise businesses now can connect those GPUs on racks and have those GPUs communicate with each other and become this integrated unit of computation that's much more powerful than just buying them piecemeal and throwing them up on a server rack. Rack Scale means interconnecting a lot of these GPUs. NVIDIA has the connection technology, which is ENV Link, and I have talked about.
They also have these great GPUs. We're going to Asterisk this bit because I know you're going to ask me about an acquisition that AMD made that answer is part of this question, but AMD needs to develop Rack Scale systems to really compete with NVIDIA. These are the two things. Get better in software and migrate to Rack Scale systems. I think between those two things, it can really have a competitive offering. I think we've moved beyond the day where we're always looking at the specs. Like, how fast is this GPU? What's the performance of it? What's the workload? How's it performed vis-à-vis benchmarks? AMD is getting closer and closer on the benchmarks. Now it becomes a question of software and making those GPs talk together.
Mary Long: If you're an outside investor, one of the ways that you might measure AMD's progress in those two areas is not just listening to what the company actually says, but also keeping tabs on their R&D numbers. Between 2021 and 2022, AMD nearly doubled its R&D spend. It's continued to tick up in the years since, but at a slower pace than that interval. We haven't yet seen that payoff in AMD's margins. Operating margin was north of 22% in 2021. It's under 8% in 2024. For comparison, NVIDIA's operating margin was nearly 62.5% for fiscal 2025. You're seeing foundation being laid by AMD to try to catch up with NVIDIA. When and how will investors be able to tell whether those R&D investments are actually paying off for the company?
Asit Sharma: One of the things I want to point out before I answer that question, Mary, is that NVIDIA's operating margin of 62.5% is an unfair comparison, not just to AMD, but to any major company. This is probably the first or second-highest operating margin in the S&P 500. If you think about the biggest and baddest US companies, this is NVIDIA riding a wave of demand in which it's exercising a lot of pricing power. Historically, NVIDIA's operating margin is healthy because it's more of a GPU-dominant business. It can range between 15 and 30% in good times. But it's a company that also has negative operating margins at the bottom of the cycle, we've seen that too, out of NVIDIA. Right now, it's taking that advantage and exercising the fact that its products are so in demand. As an NVIDIA investor, I watch that as one of the things that's going to come back down to earth, what should be an operating margin for AMD in a good part of the cycle? To me, it should be somewhere around 20% above. You mentioned it was hitting that in 2021. Again, it's a more diversified business. It has different paths to win.
I think we're seeing that R&D investment paying off this year in 2025. We should see operating margin move up to around 10% this year. The cadence looks like it's going to hit somewhere between 12 and 14% in 2026 and should hit around 20% in 2027. Only now we're seeing the investment in that R&D payoff, but that was a lot of quick-turn investment where AMD pivoted to the accelerator space because they saw the opportunity. Recall something that Lisa Su did when she first took over at AMD in October of 2014, which is to say, guys, we're going to innovate. We're not going to worry too much about the outside world, and we're going to make great products. It took two or three years for those investments to pay off, but it became look a leader in the chip space. This year, it displaced Intel for CPU coverage in data centers. I think as these years play out, the next three years, we're going to see that operating margin climb all the way up to 20% by 2027.
Mary Long: You teased out news about this recent acquisition that AMD has pursued and followed through on. If you're one way to play catch-up in this chip race is to build things in-house, another way to grow your company might be to acquire businesses that are doing work that you're already doing or that you haven't yet touched. AMD earlier in August of last year, announced that they would be pursuing an acquisition of ZT Systems. They're a service maker. AMD shelled out nearly $5 billion for that company, paid about 75% of that price tag in cash. What does ZT Systems do, and how is that going to expand AMD's potential?
Asit Sharma: ZT Systems is a designer of server systems, rack systems that I was just referring to in data centers. It not only designs them, but it manufactures them. AMD, yeah, shell out that $5 billion. Interestingly enough, Mary, it's going to actually sell off the manufacturing portion of ZT Systems because at its heart, AMD is a design company. They design chips. They don't really manufacture them. TSMC is one of its partner companies that actually manufactures chips. It's going to do the same thing here. That will help it also keep from maybe competing with some of its own suppliers. But I like this a lot because it lets AMD take a technology of its own, which is called Infinity Fabric, and basically replicate what NVIDIA is doing with its rack-scale systems. We should see in a system that's called the MI 4,000 sometime in 2026, AMD's first real convincing answer to NVIDIA's dominance. My thesis all along is that AMD doesn't have to displace NVIDIA, just needs a few billions off the top. NVIDIA is rolling with tens of billions of dollars of GPU revenue every quarter. Just give AMD a few billions of that, and this company is going to see a great boost to its margins in free cash flow. Free cash flow, I should mention, is going to more than double this year, even after AMD has announced a tariff hit from export controls on a lower-level chip it was designing for the Chinese market. It's still going to double its free cash flow this year, and it's on its way to a triple probably by 2028 in terms of free cash flow.
Mary Long: Throughout this entire conversation, we've been making the comparison between NVIDIA and AMD and you just pulled out some numbers stating that, AMD's free cash flow is going to double, potentially triple relatively soon. If you look back from where we are now over the past year, whereas NVIDIA shares are up nearly 30% in that year-long time frame, shares of AMD have fallen over 40% in the same time period. What gives? It sounds like you're laying out a very compelling case for AMD and its growth path forward. Why does the current share price not seem to reflect that?
Asit Sharma: I think the market's concerns are legitimate. The market is saying, look, if you are so great, AMD, then why didn't we see you explode in GPU sales in the first year after you said you were also going to play in this business? They did get off to a slow start out of the gate. There are questions about execution. Companies want to know if AMD really can provide that cost advantage. The other thing, I think, that poses a cloud over AMD is just this comparison. I've argued all along that AMD doesn't need this business to succeed as a company, but the market sees it very much as a race between the two most capable makers of GPUs, and NVIDIA is today been so far ahead that I think it suffers from that comparison. There's execution risk, and there's also this, I think, slightly unfair comparison that AMD suffers under, but that's actually a good place to be. AMD loved that position when it was just a shadow beneath Intel and took over that business. I'm not trying to forecast that it's going to take over NVIDIA's business.
Again, I love both companies but I do think there's room in a company that now seems relatively cheap versus its future potential for it to grab some of that market share. I think the order that Oracle made that I mentioned at the beginning of this conversation is one of the first indications that the cost proposition is making sense to companies that don't want to keep spending indefinitely year after year at the pace that NVIDIA is rolling out as innovations. Remember, you and I were chatting about NVIDIA trying to have a new better product every 12 months. That's great until people's appetite and capital propensity starts to really push up against this. I liken it to people who have sort of free money and can keep buying the latest either car or stereo equipment, and then suddenly, when that money is tight, you start to really love what you've got. I like this vehicle. Sometimes those people turn into, and I have friends like this from trading out cars and leases to, I'm going to drive this car into the ground. I paid it off. I get that. AMD can really benefit from a world in which some of these hyperscalers are like, hey, I want to run some of these GPUs into the ground.
Mary Long: As always, people on the program may have interest in the stocks they talk about, and the Motley Fool may have formal recommendations for or against, so don't buy or sell stocks based solely on what you hear. All personal finance content follows Motley Fool editorial standards and are not approved by advertisers. With the Motley Fool Money team, I'm Mary Long. Thanks for listening. We'll see you tomorrow.
Suzanne Frey, an executive at Alphabet, is a member of The Motley Fool's board of directors. Randi Zuckerberg, a former director of market development and spokeswoman for Facebook and sister to Meta Platforms CEO Mark Zuckerberg, is a member of The Motley Fool's board of directors. John Mackey, former CEO of Whole Foods Market, an Amazon subsidiary, is a member of The Motley Fool's board of directors. Asit Sharma has positions in Advanced Micro Devices, Amazon, Microsoft, Nvidia, and Oracle. Mary Long has no position in any of the stocks mentioned. Sanmeet Deo has positions in Alphabet, Amazon, and Tesla. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Meta Platforms, Microsoft, Nvidia, Oracle, and Tesla. The Motley Fool recommends the following options: long January 2026 $395 calls on Microsoft and short January 2026 $405 calls on Microsoft. The Motley Fool has a disclosure policy.
Login
