Meta is certainly the most volatile of the Magnificent 7 stocks. A brutal 75% drop in 2022 was followed by an equally astonishing 750% rally. After that kind of rally, it seems implausible for a stock to still trade at attractive levels.

Yet, Meta trades at a forward P/E of just 21 — a discount to the broader S&P 500! And while there’s much to debate about Meta, it should be clear to all of us that it’s a better business than the average S&P 500 company.

I won’t sit here and pretend that, after its most recent 20% drawdown, Meta is as good an opportunity as it was in 2022. Nevertheless, I think it would be a mistake not to look at it.

On that note, Meta told the market that it’ll invest about $100 billion in AI next year and $600 billion in the following years. Considering that Meta’s LLM Llama isn't really competitive with Gemini or ChatGPT, the market didn’t immediately see how this could turn into a profitable investment.

But Meta’s AI efforts go way beyond Llama.

Let’s dive in!

— Daniel

Applications Open: Join The Intrinsic Value Community

You might have noticed that we publish our research for free and with very limited ads. We do that because this is what Shawn and I would’ve wanted when we started our investing journey. Valuable, high-quality research that’s not hidden behind a paywall.

The other thing we would’ve wanted is the ability to talk to the people behind the research and to find like-minded investors to learn from and with. So, somewhat selfishly, we decided to build that group ourselves, using the audience around this newsletter and our podcast to do it.

By now, you probably know about The Intrinsic Value Community. It’s a trusted peer group that gives you feedback, ideas, and support, and it’s also where Shawn and I spend a good bit of time every day.

Every few months, we let readers like you know that our Intrinsic Value Community is accepting applications for the next cohort of members.

We’d love for you to consider joining and see for yourself just how powerful a peer group of like-minded investors can be in 2026, whether you’re a beginner or a pro.

Visit the link below to learn more and apply!

Meta – Yet Another $100 Billion Bet

A big part of today’s pitch will be Meta's AI story. However, I initially got interested in Meta after Clay Finck pitched it as his best idea for 2026 in a joint episode on our sister podcast, We Study Billionaires.

The growth rates this company still achieves are close to unbelievable. No other big tech name, aside from Nvidia, prints numbers like these. And we’re talking about a company that already has 3.5 billion users across its main apps. That’s more than 40% of the world’s population. How can a company of that scale still grow its top line at more than 20%?

It’s a bit embarrassing that I’m surprised by this. As an investor and as someone who uses Meta apps every day, I should have had a more intuitive grasp of what the numbers look like. But I think the reason I’m surprised is precisely because I’m a user.

From a user perspective, ad space on Facebook has looked “fully utilized” for years, and Instagram already feels like it has an ad cadence that can’t possibly be pushed further without ruining the product. But if user growth isn’t going to be the main driver either, then where exactly is growth supposed to come from?

The Family of Apps – Benefitting from AI

One answer is more effective ads. In our podcast, I told a (perhaps too) lengthy story about my vintage watch hobby and how Instagram's ads were so good that I moved from 20 minutes of screen time a day to 40 just because I kept scrolling through ads…It certainly proves how well they’re working.

So Meta doesn’t necessarily need to expand ad inventory; it’s more about improving conversion, which increases pricing power. Meta’s ads do not have a fixed price. They work with an auction mechanism. So, prices depend on what advertisers are willing to pay.

In Clay’s pitch, he described Instagram Reels as a roughly $50 billion business already, which is wild considering it only launched in late 2020. When introduced, Reels was a defensive move against TikTok. Meta needed a native short-form video format that could keep people inside Instagram. There was a well-founded fear that Meta could lose an entire generation to TikTok’s short-form video concept.

It’s somewhat underappreciated how successful Reels has been. Not only did it fight off TikTok (both more or less co-exist now), but it also created a significant ad opportunity. Short-form videos are pretty addictive. They significantly increase the time spent on Instagram, which creates more opportunities to increase impressions. And more impressions, when paired with a strong auction, translate into more revenue.

Because of the auction model, the availability of new ad space makes it seem like ad pricing power would decline. But that’s only a short-term phenomenon driven by the excess demand.

When a lot of new inventory opens up, prices drop at first because the auctions for those placements are less competitive. At the lower price, the inventory becomes more attractive, more advertisers enter, bids rise, and the platform gradually learns how to deliver better performance for those ads. By now, Reels is about as profitable as in-feed or story ads.

Beyond that, the stark drop in advertising growth in 2022 was also due to a combination of a weak overall digital ad market amid economic uncertainty (inflation, recession fears), increased competition from TikTok, and the lasting impact of Apple's privacy changes (App Tracking Transparency), which made ad targeting harder.

It’s questionable, though, how long the recent 20%+ growth rates can be sustained. The longer-term trend in impression growth is slowing. Worldwide impression growth was around 30% in 2023, dropped as low as 6% in 2024, and is now back in the mid-teens. But at this scale, I wouldn’t expect double-digit growth to be sustainable for too long.

But this is where Meta’s AI story begins. Meta has offset that slowdown with pricing growth. And although most AI headlines have been about LLMs, ad pricing and impressions are where Meta benefited most from AI. According to Meta, end-to-end AI-powered advertising tools have achieved an annual run rate exceeding $60 billion.

After Apple’s privacy changes reduced cross-app tracking, Meta had less direct signal about what people were likely to buy. The immediate consequence was that ads became less efficient, conversion rates fell, and advertisers pushed back on pricing.

Meta’s response was to lean harder into AI-driven prediction inside its own ecosystem, leveraging the massive amounts of data Meta has. Few apps have better insights into consumer behavior and interests than Meta. Over time, those models improved at predicting who is likely to take action, even with less explicit tracking, thereby improving advertiser ROI.

And when ROI improves, it shows up as higher bids in the auction, because advertisers can justify paying more per impression or per conversion.

If you look at Meta’s AI plans from this angle, they seem to make a lot of sense. Invest heavily in computing, use it to improve the ad network, and sell computing that isn’t needed to other companies.

The problem with this argument is that it revolves entirely around Meta’s Family of Apps and the ads business. In reality, that’s not all you buy, and that’s also not all AI computing and technology are needed for.

Let me introduce Reality Labs.

Reality Labs – Zuckerberg’s $100 Billion Grave

Reality Labs is the part of Meta that Wall Street truly detests. It’s where Meta houses everything related to virtual, augmented, and mixed reality. It’s the segment that carried the entire “Metaverse” narrative, and it’s also the segment that became the symbol of 2022, when investors looked at Meta and started asking whether Mark Zuckerberg had lost his way in a vision only he could see.

The scale of the spending is hard to overstate. Since Meta began reporting Reality Labs as a standalone segment (over the past 20 quarters), the accumulated losses exceed $70 billion. Chances are, the overall costs already surpassed $100 billion!

Shawn joked to me that this could be a record number of losses for a single business unit, and honestly, it might be. That said, there are good reasons why Mark Zuckerberg doubled down on this again and again.

Meta’s core problem is that it’s sitting on top of other people’s operating systems. When Apple changed its privacy rules, Meta couldn’t do anything about it. And while Google keeps (successfully) expanding into all sorts of industries and business models, Meta has not enjoyed the same success. A big part of that is the lack of controlling any form of consumer hardware.

I believe that Zuckerberg saw an escape hatch in AR glasses. I know this sounds crazy today, but given his investments in this technology and the progress Meta undoubtedly made, I believe this is/was the plan. In interviews, he says he can see a world in which we have to pick up the phone less and less. He does not yet say it, but the end goal would be not to pick it up at all.

I’m impressed by the progress made in recent years, but I consider this vision to be unlikely. Shawn pointed out that most great inventions, like the iPhone, become a hit right away. I could think of some that took a while to change the world, and yet AR and VR glasses have certainly taken their time to do so...We are 10 years into the journey.

Perhaps AI can be the missing piece. It makes the interface much more intuitive. I urge you to watch the AndroidXR demo below. Quite fascinating!

By the way, if you take the chart of the Reality Labs losses and compare it to the operating income of the ads business, well, then it doesn’t look that frightening anymore.

And yet, if you were to do the math (I did so in the podcast), I think it’s safe to say that Reality Labs is unlikely to ever drive satisfying returns to investors. Even Mark Zuckerberg recently announced that he will cut 1500 jobs (10% of the workforce) in the Reality Labs segment and shift the internal focus from the Metaverse to AR/AI glasses.

The market liked this news. However, I don’t think it will change much in the long-term. As long as Zuckerberg believes AR/AI glasses are the way to go, and he seems pretty convinced about that, he will keep spending.

What looks like savings at first are likely just investments now flowing into the “AI bucket,” which ends up in the hardware too. This does not have to be a bad thing; however, I wouldn’t model lower capex going forward. That assumption would’ve been wrong basically every single year of the last decade.

And we already know that Meta will invest about $100 billion this year, most of it in AI. And this will be a hit to earnings in the years to come. Meta’s CapEx to Depreciation and Ammortization Ratio is close to 4x right now. When CapEx runs well above depreciation, the asset base is still expanding, meaning the depreciation hit from this build-out hasn’t fully arrived yet.

The last time we saw this ratio reach today’s levels was during the Metaverse debacle. Unsurprisingly, return on Invested Capital (ROIC) tends to drop on these investment cycles as well. It takes time until the benefits of these investments, namely better ad targeting, higher engagement, and potentially new hardware driven by AI, show up in profits.

And to be honest, no one knows how profitable they will turn out to be. I think there’s some margin of safety here, since excess compute can be sold to other companies that need computing power. And yet I can’t shake the feeling that Meta struggles with investments outside of its core business – social media.

Just going over the headlines of Meta’s AI journey leaves a bitter aftertaste. First, you lose the LLM race with Llama, then you sign some of the largest deals in history to poach AI talent from OpenAI and other players. That doesn’t really work. Then you buy ScaleAI for $14 billion to get Alexandr Wang on board. Only for Wang and AI godfather Yann LeCun to get into disagreements strong enough for LeCun to leave Meta. It all seems like Meta doesn’t really know what it wants. Or at least it’s not able to get there.

Monetizing the Sleeping Giant – WhatsApp

But let’s take the chance and talk about another high-potential asset of Meta. The most-used app on my phone: WhatsApp. I know our American readers won’t have too many touchpoints with this app, but it is the number one messaging app globally, so yeah, it’s quite big.

Meta bought WhatsApp for roughly $20 billion in 2014. Back then, it only had 200 million users. Today, it’s over 3 billion. But over all these years, it has served more as a multiplier for the ecosystem than a money machine.

It’s bringing in $15 billion right now. Two-thirds of that comes from WhatsApp Business, where companies can create verified business profiles, interact with customers, and use messaging as a customer support or commerce channel. This is often done in combination with Facebook or Instagram, where businesses place a button that lets interested users chat on WhatsApp.

I’m a bit skeptical about the potential of monetizing WhatsApp, though. While WhatsApp Business is already widely used in India and parts of South America, it’s a pure-play messaging app in Europe. And to generate meaningful profits, WhatsApp needs to leverage its monetization potential in Europe, where ARPUs (Average Revenue per User) are much higher.

And just from personal experience, I don’t like using WhatsApp Business. WhatsApp is where I have my personal chats with friends and family. I’m not keen to have a dozen business chats there. And the few times I used it, I ended up talking to a chatbot that wasn’t helpful at all.

I know that the experience in India and elsewhere can be different. Due to lower labor costs, it can make sense in those countries to hire someone to handle these chats. That’s not yet the case in most European countries.

Again, AI could be a game-changer here if it can improve the conversation while reducing costs. Who knows, perhaps WhatsApp can become a sort of CRM tool. If we look back at WhatsApp in 5 years and it’s a huge profit driver, it’s likely to come from WhatsApp Business.

While there’s no other business in the world that understands social media ads as well as Meta, and it’s almost a sin to bet against their success on this, I don’t see WhatsApp becoming a good platform for ads.

Snapchat started putting ads in the chat section of the app, and, in my humble opinion, it felt like a huge privacy breach and was extremely annoying. I doubt Meta will make the same mistake. Currently, ads are planned for the status section of WhatsApp. WhatsApp Status is essentially the equivalent of Instagram Stories: short-lived photo and video updates that disappear after 24 hours.

Meta claims that WhatsApp Status has roughly 1.5 billion daily active users and describes it as the “most viewed story product in the world.” If that’s true, ads could be a huge success here. I’m skeptical, though, as I don’t know anyone who uses that feature. I guess it’s different in international markets (let me know if you use Status in the comments!). But again, Meta is a company that brings in $80 billion of operating profits. If you can’t monetize the European WhatsApp user, it’s hard to imagine ARPU is high enough to make a difference. At least in the short-term.

Valuation and Investment Decision

Meta is an unusually dynamic company, making it hard to pin down not only how much it will invest in the coming years, but also what kinds of returns those investments can realistically generate. We know 2026 will likely bring roughly $100 billion in spending. Zuckerberg has even thrown out a figure of up to $600 billion for total AI data center investment over the next few years.

I’m hesitant to anchor on that figure, because it didn’t sound like a carefully calibrated capital allocation plan as much as an emotional signal of intent. He wants to make it seem like no one will (and can) outspend Meta.

And at the risk of repeating myself, I’m skeptical. Meta has materially improved its ad system with AI over the last couple of years without needing anything close to a hyperscaler-style buildout. Now the company is spending like one.

Microsoft, Google, and Amazon run massive cloud businesses where selling compute is the product. They can justify multi-hundred-billion-dollar infrastructure cycles because there’s a direct revenue line attached to them. Meta doesn’t have that. So I’m not confident enough to judge whether a number like $600 billion could realistically build a “Meta cloud” over time, or whether “we’ll sell any excess compute” is mostly an easy justification that sounds great on stage, while reality is much more complex.

As Shawn rightfully argued, Apple’s route might not be the worst. Instead of spending hundreds of billions of dollars on AI themselves, they acknowledge that they are a hardware company first. Pioneering AI is not their field of expertise. Perhaps Meta should realize it is a social media company first.

Anyway, with that backdrop, let’s get into the valuation. The main idea I wanted to test with my model is whether Meta still looks cheap if revenue growth slows into the low teens and operating margins in 2030 are roughly where they are today.

The heavy capex wave we’re seeing now will eventually flow through to depreciation, putting structural pressure on reported margins. And if there isn’t a clear payoff within the next three to four years, I think this is a realistic outcome.

It’s important to note that I don’t necessarily believe all of Meta’s efforts will yield low returns, but the last few years have not yet proven that Meta can monetize AI efforts beyond its Family of Apps. And that part of the business certainly doesn’t need hundred-billion dollar investments.

So in my model, operating income still grows from roughly $80 billion in 2025 to more than $150 billion by 2030. For free cash flow, I assume a conservative 70% conversion from operating income in the first few years, given elevated AI capex, followed by 80% by the end of the decade. This takes FCF from roughly $45 billion to around $110 billion over the forecast period.

I’m aware that this newsletter sounded skeptical about Meta at times. That’s not because I don’t think it’s a great business. The ads business is phenomenal. But I see everything outside of social media as expensive call options. I haven’t yet seen enough to put trust in their success.

Yet, as you can see in my model, I think Meta is attractive at today’s prices and likely to return double-digit returns going forward. But with little cash remaining in our portfolio that we aim to hold for market downturns, Meta would need to not only clear our hurdle rate but also beat our existing positions.

Comparing profiles, Google is the closest competitor. Shawn said that he would be more inclined to add to Google than to establish a position in Meta. I agree that Google is a better business. If you listened to the podcast, you’ll see that I even consider Google to lead the AR/AI glasses race. One of Meta’s key products of recent years.

But I’m too much of a value investor to ignore that Google is trading at one of its highest levels in the last decade, while Meta is valued below the average S&P 500 company.

So, given our 13% position in Google, I’d rather add to Meta at these levels (than Google). That said, I don’t see any positions I’d like to trim or sell for Meta either.

That’s a long-winded way of saying Meta is at the top of our watchlist but hasn’t made the cut yet. Next week, Shawn and I will have another portfolio review call. Meta will likely be a topic there, too.

For more on Meta, you can listen to our podcast here.

More updates on our Intrinsic Value Portfolio below 👇

Weekly Update: The Intrinsic Value Portfolio

Notes

• Adobe can’t catch a break. The stock has now dipped below $300 and trades at roughly 14–15x forward free cash flow. For a company still growing its top line at a double-digit rate, and compounding per-share even faster thanks to a buyback yield around 9%, that valuation looks like a bargain in our view.

  • The market clearly doesn’t agree. The prevailing narrative is that growth is about to slow sharply, or even stall. Shawn and I don’t see it that way. Both of us added to our positions in our personal portfolios. The harder question is whether we should do the same in our Intrinsic Value Portfolio.

  • Adobe is the position we’ve lost the most money on in the portfolio, and we’ve already doubled down on it a few times. We have to ask ourselves whether price action alone is enough to continue adding or whether we need to see some sort of catalyst.

  • To be clear, we still strongly believe in Adobe, but we want to avoid not learning from mistakes. And doubling down a third time on a stock that went against us until now has to be carefully considered. There’s a fine line between conviction and ignorance. I don’t want to find myself on the wrong side of it. Shawn and I will be meeting sometime in the coming weeks to discuss the Adobe position in particular, and we may or may not decide to either add to it or hold it steady.

Quote of the Day

"Effective capital allocation requires a certain temperament. To be successful, you have to think like an investor, dispassionately and probabilistically, with a certain coolness.”

— Michael Mauboussin

What Else We’re Into

📺 WATCH: Shopify Founder Tobi Lütke on how he thinks about Business and Founding

🎧 LISTEN: Stig Brodersen: Reflecting on Money, Life, and Happiness

📖 READ: Mark Leonard’s Constellation Shareholder Letters 

You can also read our archive of past Intrinsic Value breakdowns, in case you’ve missed any, here — we’ve covered companies ranging from Alphabet to Airbnb, AutoZone, Nintendo, John Deere, Coupang, and more!

See you next time!

Enjoy reading this newsletter? Forward it to a friend.

Was this newsletter forwarded to you? Sign up here.

Join the waitlist for our Intrinsic Value Community of investors

Shawn & Daniel use Fiscal.ai for every company they research — use their referral link to get started with a 15% discount!

Use the promo code STOCKS15 at checkout for 15% off our popular course “How To Get Started With Stocks.”

Follow us on Twitter.

Read our full archive of Intrinsic Value Breakdowns here

Keep an eye on your inbox for our newsletters on Sundays. If you have any feedback for us, simply respond to this email or message [email protected].