The Bear Case vs The Bull Case for Megacap Tech

0:00Mindy Jensen:Mega cap stocks have been some of the biggest winners in the market, but does that make them a smart move for someone pursuing financial independence? With huge valuations, strong cash flows, and increasing concentration in major indexes, there's a compelling argument for both sides.

0:21Mindy Jensen:Hello, hello, hello, and welcome to the BiggerPocketsMoney podcast. My name is Mindy Jensen. He's Carl Jensen. And with me, as always, is my doesn't-have-many-single-stocks co-host, Scott Trench. Hey, I'm so excited to be here with you and Carl today to talk about this subject. And Carl and I had a great conversation about this at lunch the other day. So we're going to be hashing out some of those points here. And just to frame the debate at the highest level, 11 companies make up about 38 % of the S &P 500. Actually, nine companies make up about 38 % of the S &P 500 in the mega-cap tech complex.

0:53These companies are NVIDIA, Alphabet, Apple, Microsoft, Amazon, TSMC, which is not in the S &P 500, but is a major semiconductor manufacturing company, Broadcom, SpaceX, Meta, Tesla, and Oracle. And also call out that SpaceX is not yet in the S &P 500 following their recent IPO. But these 11 companies add up to about$32 trillion in market cap value,$33 trillion in market cap value. And$30 trillion of that is in the S &P 500 or about 40 % of the S &P 500. And these companies are richly valued. I think the aggregate multiple is well north of six to seven times of total sales for this aggregate. It gets worse if you exclude Apple, for example, as many people want to argue.

1:35And it gets even worse if you exclude the sellers like NVIDIA and TSMC. The question is, can these companies win? And what do you have to believe for them to win? I am skeptical of that. And I am so skeptical that I've made significant moves. I have paid taxes in order to do so to shift my wealth away from the S &P 500 index fund to an equal weight index fund and to factor tilts in small cap value, U.S. international. Carl and Mindy have done the exact opposite and I think are bullish on this AI world and the AI complex and think that there's many paths to winning here. And you have heavily invested your personal net worth in these stocks and continue to hold these positions as they've grown over the years.

2:16Is that the right way to frame it, Carl? Yeah, I'd say that's the right way to frame it. But one thing I will say is we are big fans of index investing. We have put more money into index funds than anything else. It's just that these stocks have done better than the index funds. We bought a lot of these a long time ago, like Tesla, I think we bought for$2 a share in 2012. Google was$85 a share in 2004. I think if you extrapolated that to now without the splits, it would be$15 ,000 a share, something around there. So we bought these companies a long time ago, but we do strongly believe in index investing, even though we are holding on to these stocks too.

2:49Let me frame my argument with one example of a company that I think you're really familiar with, which is Alphabet. Also, before I get to this, I want to call out, this is something that we, in the wake of the Money Guys podcast, you know, you guys were saying, hey, we're index fund investors. And they're like, are you? Well, yeah, I think you are. I think you just said it. You've put more of your money into index funds than you have into tech stocks. And many people invest in different things other than index funds with small percentages of their wealth. And that's exactly what you did. You just happened to hit these huge winners that you've let ride for many, many years.

3:22And that's why your wealth is so heavily skewed or concentrated in these tech stocks is because you hit the winners in the space. And somehow that makes you not an index fund investor. But if they'd gone to zero and you were 100 % index funds, then nobody would have a problem calling you an index fund investor. So I think that's kind of fun internet mental math or however, whatever the term is there. So yes, you guys are index fund investors in my book. You also have huge winners that you've hit. But here's my argument for Alphabet, right? So Alphabet trades at about 31 times trailing 12-month earnings.

3:53It's actually about 28 and a half. I'm citing a source from about a month ago when I wrote this. And you pointed out something here, Carl, yesterday when you were texting me. What was your reaction when I said that Google trades at 31 times? I was looking at the PE ratio. And specifically, I looked at it on like Yahoo Finance and Google Finance. And they have it at 17. And then you told me why you have a different number there. It looks like Alphabet's operating income for the last 12 months is$147 billion as of today's recording. Their market cap is$4.2 trillion. So$4.2 trillion divided by$147 billion is going to be 28.5 times.

4:31So it fluctuates day by day, but 31 is what I was using when I wrote the article. It's about 28.5 times core operating income. But their official price to earnings ratio is closer to 17, like you said. And the reason for that is because of these unrealized gains, like you mentioned, like their investments in Anthropic and SpaceX, although they've not been officially and fully disclosed what those are. There's almost it's almost certainly the investments in those two companies. So we have this issue where more than half of Alphabet's earnings, about half of them are coming from unrealized gains in companies that they're investing in in this technology ecosystem.

5:07That scares me because when Google invests$40 billion in Anthropic, and Anthropic is committing$200 billion over the next several years to Google's cloud revenue service, they're buying compute from Google that Google is building out. That's a pretty circular mechanism here. And somehow, some way, Google and Anthropic must generate enough revenue from somebody else in order to cover that. And so my issue is Google can win, Anthropic can win. There's cases for all of these. But can both of those two companies win together? And when you stack in the competition, the same thing's happening with Microsoft, where they're milling out hundreds of billions of dollars in compute for AI services, and they happen to be renting most of that out to OpenAI.

5:49Now we've got the same thing going on over there. We've got Amazon doing something very similar. We've got Meta entering the race with similar volumes of investment. However, their CapEx investment is primarily for internal use, although they keep flopping about whether they're going to rent it out or they're going to use it solely for their own build out. And then we've got Grok and SpaceX and X entering that fray, supposedly, with parts of their approach. We don't know what Tesla is going to be doing from a wildcard perspective. And we've got Apple sitting on the sidelines, apparently, investing the bare minimum in AI to continue offering core services.

6:23And then we've got the sellers, right? We've got NVIDIA, Broadcom. We've got TSMC Taiwan Semiconductor Manufacturing Company. And these guys are selling in various, you know, NVIDIA is typically selling to Apple, Microsoft, Amazon, and Meta, whereas Alphabet is one of the primary customers of TSMC and Broadcom here. And so we've got this interconnected chain, and they're all valued as if they're going to win. And my view is the combination is going to really struggle to make this work. It kind of reminds me of that quote from Harry Potter, neither can live while the other one exists or whatever that is.

6:56Do you remember that?

6:57Mindy Jensen:Neither can live while the other survives. Okay. I think this kind of applies to this. And one thing, Scott, to build on what you said, I was reading an article in the Wall Street Journal yesterday, and they pointed out that a lot of the capex that these companies have committed to, I think it was$3 trillion, isn't even accounted for. It's some kind of accounting trick where it's on the books. they have signed the contract, but it's in the future, so they don't have to formally record it now. So I think it's even worse than what you've said. You're Mr. Numbers, so you might have a better idea of what I'm talking about here.

7:26This is about one month old, but I wrote an example a little bit ago in preparation for this. And then, of course, Google releases their earnings call, so it's all updated. But price to earnings is a bad way to value Alphabet right now, not just for the reason we discussed, but also because of AI CapEx. For example, over the last 12 months, before their Q3 earnings call here, Alphabet had spent$110 billion on CapEx. About 60 % of that was on compute. They don't use GPUs. They use a device called a tensor processing unit. That's their bet on AI, basically, is that they're going to be cheaper, faster, safer, happier with these tensor processing units.

8:01And then about 40 % of that spend is on buildings, land, power, the data centers, and the other components that are needed to facilitate this compute. And so that is considered an investment. It does not hit the P &L, profit and loss statement, does not hit earnings in that calendar year. It begins to depreciate once it is placed in service. So we can spend the money now and we'll hit the expense part of the P &L later. In 2025, they spent$91 billion. Through the 12 months leading up to Q2, 2026, they had spent$110 billion. In 2026, we're estimating$180 to$190 billion in CapEx. And their CFO, when asked about what's 2027 is going to look like, said, we're going to significantly increase this.

8:45This is our famous quote. So all you can do is guess. So I put down$215 billion, but Wall Street varies because that's all we have, right, is this significantly increased remark from the CFO in terms of what they're going to invest. So at least through 2026 and 2027, they're going to be spending this money to invest in AI CapEx. And that means that their cash flow for the last 12 months is going to be$64 billion. And if you value them on the$4 trillion market cap, and again, these members move a little bit with market valuations day to day, but they were trading at about 71 times price to free cash flow.

9:20That's an extremely expensive valuation. In 2026 full year, they're spending effectively all their cash on this. So the numbers get very silly, right? We could say, oh, we're expecting net free cash flow of$8 billion in 2026. So they're trading at 568 or some crazy multiple of cash flow next year. And in 2027, you can basically put a zero or something very close to it again. And so what this brings is your bet on Google right now is you know they're going to generate no cash flow this year. You know they're going to generate no cash flow next year, at least if you believe what they are saying.

9:53And so you're betting on a ramp. You're betting on them to stop spending this capex and for it to generate real long-term cash flows. on a go forward basis. And there's real reason to believe that. However, we've got a triple cashflow excluding these investments in order for that to work. And we've got to really ramp it and stop spending as a precondition for that. That's what concerns me about Google, but I'm also not betting against them. There's a real possibility they can win, which I want to talk to you about. My broader issue is once you make that same story across five or six different companies that are building almost identical and directly competitive products here.

10:30And that where the suppliers like NVIDIA, their bull case is that Google continues to spend. And Google's bull case is that they stop spending one day and harvest those cash flows. How can they all win? That's my broader question with this analysis. Yeah, I'm not sure they can all win. But one thing that I want to back up and mention or talk about a little bit is one thing I always wondered when I started, When I did discover index funds, like J.L. Collins and Madfiantist, all these people praised them. And the question I had was, why does this go up and to the right? What causes that up movement over time?

11:03And so I actually remember asking J.L. and Madfiantist, and their answers were both the same. They said, I don't know. And then we were at the Berkshire Hathaway Conference, and Warren Buffett started talking about it. He said, well, there's two things that make this go up and to the right. And one of them is much more important than the other one. The first less important one is population gains. The second, much more important one are productivity gains. You think of historical things that have caused productivity gains. The tractor was a huge one. I think back around 1900, almost everyone was a farmer.

11:32The tractor came out and then now I think it's less than 5%. So that went down like crazy, the numbers, the steam engine, the industrial revolution. When you take a look at AI, well, to back up a second, the tractor, the steam engine, they revolutionized certain parts of the economy. AI has the potential to change everything. You read about the first jobs they're going to go, and they're probably white-collar jobs, like accountants or attorneys, things like that, because coders, that's what the computers can research and figure out online. But there's plenty of companies working on robots to take away the physical jobs, too.

12:07So my question back to you would be, what is Google betting on that they think they can spend hundreds and hundreds of billions of dollars in CapEx? They're betting on a future that's wildly different than the present. And I think the main thing that a lot of these people are afraid to say because they don't want the backlash, although I think Dario, the Anthropic guy, has said it, is that human jobs are going to go away if these companies, if their vision is fulfilled. And if that goes away, maybe some of these huge, huge multiples are justified because right now, who is paying for AI? And it's hard to tell.

12:41That's another important point. Like OpenAI and Anthropic are not public companies, so we don't know how much they're bringing in. but I'm sure it's not nearly enough to justify their spend. It's probably a fraction. But once you start taking human jobs out of the way, what is an AI agent worth that can replace a lawyer? I would say maybe hundreds of thousands a year, maybe millions, because that thing doesn't have to take breaks. It's never going to call in sick. It can work nonstop, and it's not going to make mistakes on and on. It might be able to do the job of 10 or 20 different people. So that's when you see the real money start to flow to AI.

13:13And I think that's what these tech companies, that's why they're betting hundreds of billions of dollars on. But to your point, I don't think they're all going to win either. And it's interesting, there might just be one, whoever gets to AGI first, recursive learning where the computer becomes hyper-intelligent. So I think what's interesting here is I would argue, I think you would agree, I'm the AI power user here. You use them lately and you've used them in the recent past here, but I have a subscription to Claude. I have a subscription to Grok. I have a subscription to ChatGPT, and we use Gemini to run basic operations here at BiggerPocketsMoney.

13:47I think my bill last month for AI in aggregate for personal use was$169.99. I'm using it all the time, right, for heavy coding for these applications. I'm using it to, you know, review very basic, not high stakes, basic legal documents. I'm using it to help me ideate content. I'm helping, you know, I'm using it to grade, you know, arguments and those kinds of things or preview them to various degrees. I can't imagine personally using more AI right now. And I know that the use case is an always-on AI bot. That, in practice, is much more of a pain in the rear than it is a value-producing play for me so far.

14:27Maybe that will change in the future. For now, I can spend very little. And what's more is every couple of months, one of the AI models leapfrogs the last one. And so if Claude, for example, if I got too dependent on Claude, I could bet you that within a few months, GPT is going to figure out something, catch up and flip the switch on it. And so that's the question I have is I completely agree that if you are going to build the bull case for the complex as a whole, okay, it's going to divert revenue from that goes to human labor right now in the white collar workforce in the near future, and maybe all jobs, including the physical world in the long-term future.

15:03But the question is cost, right? If you can do that, if accounting is now an AI bot, there's a race where the compute gets better and better and cheaper and cheaper at an exponential rate, given the investments that's going on. And does it become effectively free to provide those services from the AI? And so that's a threat there. But yes, I agree. I would definitely agree that betting on the complex right now is fundamentally in some form about that idea, that concept that you said, which is AI will substitute for a significant portion of today's current labor force. And we'll actually be able to realize revenue as it substitutes for that.

15:38I pay$100 a month to have my car drive me around, which is an AI system. You pay a couple hundred bucks to find your answers, but how much would someone pay to have an AI that could solve cancer, that could create a customized thing to produce antigens to destroy your tumors? I mean, something like that would be worth billions and billions of dollars. And I know things like that have already been done. Like the Mayo Clinic uses AI now for pancreatic cancer early detection and supposedly can detect it like three times earlier than a physician reading. And I wonder if these tech companies just keep it to themselves.

16:11If they know they have that kind of power, why would they solve to anyone? I think that that's a great segue here. I think it's what do you have to believe for this to work? So right now, these 11 companies, NVIDIA, Alphabet, that Apple, Microsoft, Amazon, Taiwan Semiconductor Manufacturing Company, Broadcom, SpaceX, Meta, Tesla, and I included Oracle because at the time I did this, they were worth a trillion dollars and were part of this complex. They've since imploded to 60 % of their value here. But those 11 companies, including Oracle, make up about$30 trillion in market cap. After net cash, That's about$29.5 trillion in market cap.

16:51And their combined free cash flow today for 2026, TTM, is$450 billion. So that means that they're trading at 65 times price to free cash flow at today's levels. If you want a 10 % return as an investor, and you give them reasonably generous assumptions here, you're going to need for them to compound that free cash flow total at 35 % a year for the next decade straight to$4.3 trillion in free cash flow as a complex in order to generate 10 % returns. If we're very generous and assume that after 2027, the hyperscalers will slow significantly their CapEx investments and NVIDIA and Broadcom and TSMC still find other ways to generate revenue, maybe new buyers, for example, and they collect their cash flows, then from there, they will need to compound free cash flow at 27 % per year for about a decade straight to deliver a 10 % return to investors.

17:55So I agree with you that there's pit places to win, but that magnitude of that bet for my model at biggerpocketsmoney.com slash mega cap, and I put all these inputs here for people to change, is to me fairly preposterous. And especially once you think about, hey, like, What does that mean on a planetary scale? And this is, you know, some people roll their eyes in the planetary scale argument, and I get it because, you know, you could have made that 10 years ago to some degree. But you're basically requiring at today's margins these companies to generate$17 trillion in top-line revenue to generate this$4.3 trillion in free cash flow.

18:30That's 17 % of the entire global corporate pool of revenue for these companies, and it's$2 ,000 bucks per person. If you just want to put that into countries with affluent consumers, then you're talking about$14 ,000 for everybody. $14 ,000 for you, Mindy, for you, Carl, for Claire, for Daphne, right? For all of those guys, you're going to need to be paying about$14 ,000 to some degree, and they're going to need to profit$3 ,500 for that. So this is by far the biggest revenue line item in every affluent human household. And so that's where I get tripped up is these numbers are so large that it's very difficult for me to believe that story from here to there based on the current valuations.

19:11And I also think that, remember, as a complex, they have to do this, some version of this. You can lighten some of the assumptions and make it slightly easier, but they have to do some version of this as a complex in order to get to that point. And my belief is that there will be some winners and some losers, of course. But if there's one winner or two winners, for example, I believe now you've got a secondary bet if you believe those Because is the American population or the worldwide population going to politically allow a single entity of that size to exist and harvest that revenue or cash flow?

19:43And that's going to be another challenge with this thing. I think there's actually a political bet embedded in the valuations of this complex in today's environment. So what's your thoughts on that? Those are very big numbers. I think you said$4.3 trillion. And what's the current U.S. GDP? It's like I think it's right around$40 trillion, something like that. So that's over 10 % of the current GDP. But again, that expanse of productivity goes crazy because of these AI agents. So yeah, US nominal gross domestic product was 32 trillion. So Carl, one thing I want to call out here is that's 4.3 trillion is the profits.

20:14That's one eighth of today's GDP that's applying for the profit. But to get there, they have to generate revenue. We'll have some expense. So I would actually say that the number that you're looking for here is this one here at$17 trillion in revenue for the complex. I think their expenses will go way down. I mean, software, Mark Andreessen once wrote that article, software is eating the world because Microsoft can write Windows once and sell a billion times. So as these models get better and better, they're going to have to invest less money, at least in the model generation aspect of it. Then everything goes to inference, which is where you ask a question and how you're actually dealing with the brain of the system.

20:53But I think that gets cheaper over time. That's the whole reason companies like Google, Mark Zuckerberg, again, SpaceX are talking about putting these data centers in space. That's for inference processing where you get free electricity and semi-free cooling. But yeah, those are very big numbers. And the two things I would say is not all these companies are going to make it. It's clear that some are going to be left behind, probably most of them, or they'll just revert to what they did before. Apple will continue to make hardware. SpaceX will make their money from launches, which isn't that much compared to what they're doing now for AI.

21:25But the other question I'd ask On the flip side, if you were a consumer, how much would it be worth to you to have a car that could drive itself to go pick people up from the airport, a robot that could take care of your lawn, do your laundry, maybe go pick stuff up for you, be a personal assistant? What about an AI assistant that takes care of your life, kind of like a virtual secretary? People pay a lot more for those kind of humans in their real life. So I'm not sure if it's worth$14 ,000 per person. I could definitely see$14 ,000 in a household. And I think that's what these companies are betting on.

21:56And they're looking at a future that we can't even imagine where AI and the products that come from it take over our lives. But then the corporate side of it, too, if the accounting companies can get rid of accountants, if drug discovery can go to AI. And instead of it taking a decade to come up with a new drug that may or may not pass muster when it comes to their phase two or phase three trial, AI comes up with it in a couple of days or a couple of weeks. I think the corporate side of it, the non-consumer side of it might be even more valuable. And that's where a lot of this money would come from.

22:26And again, if you get rid of jobs, that's a lot of people who aren't getting paid. And that's money that's going to flow to Google or Anthropic. These are very big ambitious goals that are probably a long time off. I saw SpaceX engineers say, no, yeah, our robots will be able to build. The Tesla bot will be able to build a house by around the year 2035. I do not believe that to be true. Although it will happen at some point. Maybe you could unclog the toilet. More sophisticated things than that. But yeah, yeah, Scott, there is a ton of money. And there's probably going to be a reckoning. We've already seen it a little bit.