META Stock Just Got More Dangerous

Meta has become one of the strangest stocks in the market. We can see year to date the company's down around 10%. Yet it's growing revenue at almost 30%. His advertising machine is still getting stronger and more than 3 and a half billion people while they're using its app every single day. Yet at one point this year, Meta, as we can see, it's fallen 25% from its highs. And even after recovering, as we just saw, it's still down year to date. It's a remarkable disconnect for a business producing numbers like these. But this week, another problem landed on Zuckerberg's desk. Med's now been ordered to pay hundreds of millions of dollars more following a major child safety case in New Mexico. and the fine itself. That's not what worries me because investors are increasingly asking two completely separate questions at exactly the same time. How much regulatory risk is sitting underneath Meta's social media empire? And how much money is Zuckerberg prepared to spend trying to turn Meta into an AI company? And today I'm going to try and answer the second question in a completely different way. Instead of building one normal valuation for Meta, I've built three completely separate DCF models. One of them assumes Zuckerberg's AI spending doesn't deliver the returns investors are expecting. The other model, well, that's in fact my current base case. And another model, well, that assumes Zuckerberg is ultimately right and AI becomes another enormous growth engine. And here's the crazy part. One of those models says Meta is overvalued today with another which puts the stock worth more than $1,000 per share. So, the difference between these outcomes is enormous. And by the end, I'll show you which future I think the current $592 share price is actually pricing in and which one I personally think is the most realistic. Now, let's start with the news. A New Mexico judge has ordered Meta to pay an additional 567 million, bringing total penalties from this particular case to around 942 million. Now, look, for a company the size of Meta, that's painful, but it's certainly not existential. Meta, in fact, generated more than 60 billion of revenue in their latest quarter alone. So, if this simply were a story around writing a $567 million check, well, I probably wouldn't even be making this episode. The more important part is what the ruling could eventually mean for the product itself. And we've increasingly seen people describe these lawsuits as social media's potential big tobacco moment. And whether you agree with that comparison or not, from an investment perspective, the concern is straightforward. The biggest risk isn't necessarily the fines. is regulation that starts affecting how Meta's platforms generates engagement. Things like notifications, age verification, recommendations autoplay infinite scrolling, all of the little product decisions that determines how long somebody remains inside Instagram or Facebook and therefore how many advertisements Meta gets an opportunity to show them. And this is why I thought this particular explanation of the judgment was much more important than simply looking at the dollar amount. Listen specifically to what could potentially be at risk beyond this one case. >> And the judge did not mandate a total algorithm overhaul. But the judge did say one reason he's not mandating major tweaks to the likes of autoplay or infinite scroll is that they are so widely used. So those features could be at risk. in the other cases which also include YouTube, Tik Tok and Snap. And there are a range of other lawsuits in the works with the next one a trial later this month in which Meta will defend itself against a coalition of state attorneys general accusing it of designing addictive features. There are another set of trials happening in October against YouTube and Tik Tok as well. Now Meta reported $2.4 billion in legal expenses that quarter which contributed to his profit decline. That speaks to the massive cost and potential distraction of all of these lawsuits. >> And that's the distinction I think shareholders need to make. A billion dollar judgment is manageable. But if these cases eventually force structural changes to the mechanics that drive engagement, well, that has the potential to affect the economics of the underlining Meta business itself. Now, look, at the same time, I also don't want to exaggerate this. Meta is appealing these rulings and not every case is going against the company. A separate US teen addiction lawsuit was withdrawn just days before the trial. So this is not me saying litigation destroys the meta thesis. What I'm ultimately saying is that regulatory risk deserves a bigger discount than perhaps investors were giving it 2 years ago. And unfortunately for Meta this is arriving at exactly the same time as Zuckerberg is making the biggest capital allocation bet of his entire career. I mean you can see that nervousness in the stock med is down roughly 10% year to date despite the S&P 500 itself being positive and the comparison with Alphabet well it's particularly striking. Google's up 13% year to date. Met is down around 10. same broad AI theme, very different market verdict. And that the worst point of the draw down, MED is down around 26% from its highs. And after the latest earnings report, well, in fact, the stock dropped around 10% again. As investors, they're focused less on the strength of the core business and more on the size of the spending itself. And that distinction, that's crucial. The market isn't looking at Facebook and suddenly saying, "Nobody uses this anymore." It's saying how much of the cash this business generates is Zuckerberg going to reinvest before shareholders ultimately get to see this again. And look, analysts themselves have been asking exactly the same question. JP Morgan, all of these in fact happening within just the last few days. They cut their target to $640. They cited concerns around AI monetization. We had Guggenheim. They in fact went down to $700 specifically pointing towards capex concerns. We've got another analyst here $650 pointing out to higher cost. We've got Scotia Bank $600. And others remain much more constructive. Bank of America, well that still sits fairly high, $810. Rosenlot, they've got it around $883. And that's almost the perfect illustration of the disagreement around Meta today. Everybody can see the same extraordinary underlining business. They're simply assigning very different probabilities to whether Zuckerberg generates a return on the money that he's spending. And this is where the bearish story becomes much more complicated because fundamentally Meta's latest quarter was extraordinary. Revenue reached just shy of 61 billion, up 28% year-over-year. And look at the trajectory here. quarterly revenues climbed from around 30 billion several years ago to more than 60 billion today. So Med is not some mature business struggling to find growth. I mean their advertising business alone, well that generated 59.4 billion in the quarter. That's up around 27% year-over-year. Meta's ad business isn't merely surviving AI disruption. Right now it's accelerating. And on a training basis, while Meta remains one of the two enormous global advertising machines alongside Google, its advertising revenues continue climbing rapidly even as the industry itself becomes increasingly competitive. And the recent growth comparison, that's also impressive. Meta's trailing advertising growth in this data set is around 27.7% well ahead of Google, Amazon, and Snap. Now expectations naturally slow from here, but the current performance is anything but weak. And then we move on to engagement. Family daily active people. Well, that reached 3.6 billion. That's almost half the global population using at least one Meta app every single day. And despite how enormous the user base already is, that number is still gradually growing. But also understand this, Meta isn't relying purely on adding users. It's monetizing them better. Average revenue per person rose to $16.86. That's up meaningfully when we look at the same quarter last year at $13.65. Worldwide ad impressions, well, that grew 14% year-over-year. It tells us Meta is creating more ad inventory. And the average price paid per ad, that increased around 12%. Think about the combination. More users, more impressions, higher pricing. That's an incredibly powerful advertising flywheel. And most importantly, the growth isn't isolated to one geography. US and Canada, they remain the monetization powerhouse, but Europe, Asia-Pacific, and the rest of the world, they all continue contributing materially. The family of apps, well, they generated more than 60 billion of revenue. Royalty Lab will only 431 million. It tells you everything you need to know about Meta today. The core advertising empire is funding almost everything else. And the family of apps, they produced 23.4 billion of operating income in the quarter. Royalty Labs, well, that lost around 4.6 billion. So, when people say Zuckerberg has a history of spending aggressively on long-term projects, well, honestly, they're not wrong. Yet, despite all of that, trailing operating profit remains close to 87 billion, is still one of the most profitable businesses on planet Earth. Which is why today's debate isn't whether Meta is a good company, it's what happens if an exceptional company starts spending exceptional amounts of money. And here's really where things get uncomfortable. Meta spent 31 billion on capital expenditure in Q2 alone. A year earlier, well, that was a lot lower. We're talking 17 billion is an increase of more than 14 billion in 12 months. And the acceleration is extraordinary. For years, quarterly capex mostly sat in the singledigit billions. Then it moved into the teens, then above 20 billion, and now around 30 billion in just one quarter alone. Through only the first half of 2026, Med has already spent around 51 billion. Last year, over the same period, we're talking 31 billion. And management's fullear target, well, that's enormous. And when we look at it from this perspective, let's measure it against revenue. Meta's projected capex burden is around 60% of sales on this comparison, higher than Microsoft, higher than Google, more than double Amazon's ratio, and only Oracle in fact sits higher. And this, look, isn't uniquely a meta phenomenon. Microsoft, Alphabet, and Amazon have all dramatically increase spending. But that's exactly why investors have suddenly become obsessed with free cash flow because ultimately somebody has to pay for this AI buildout. Now, Meta, they generated around $ 32 billion of operating cash flow in Q2. And then you've got property and equipment, which consumed around 30 billion. Finance lease payments, they took another billion. What was left? Well, we're talking 784 million. Now, for context, Meta generated 10.9 billion of free cash flow in Q2 2024, 8.5 billion in Q2 2025, and now just 784 million. The underlying business hasn't collapsed. The cash is simply being redeployed. And this is where the AI story becomes much more interesting because some of the caution is no longer coming purely from outside analysts. These were reportedly comments from Zuckerberg himself during an internal town hall. Reuters saying that Zuckerberg said that the 2026 reorganization of the company wasn't as clean as it could have been and also saying this internal town hall that AI agent development over the last four months has not accelerated in the way we expected. This is a continuation of some commentary we've heard from him in the past. Um he said that in this town hall uh according to this report in Reuters that top executives were worried in early 2026 that we weren't going to move fast enough and that the bets the company made on its reorganization have not come to fruition just yet. >> And those words matter not because they prove Meta's AI strategy has failed. They don't. But when a company is spending more than a hundred billion dollars in a year, investors understandably want evidence that execution is accelerating alongside the spending. Right now, that evidence is still mixed. You can see the AI race flowing through expenses. Well, research and development, well, that jumped to 36% of revenue in Q2. That's a dramatic increase from the mid to high 20s only a short time ago. Now, look, Meta still earns extraordinary margins. Gross margin in fact sits at 82%. EBIT margin sits at 38% and net margin sits almost at 30%. But free cash flow margin that has in fact compressed sharply. Historically medic produced free cash flow margins comfortably above 30%. The latest trading figure here shown is closer to 18 and this is precisely why today's DCF is so difficult. The same broad effect is visible across the hyperscalers. Free cash flow is being squeezed as the AI infrastructure buildout accelerates. The market is currently making an enormous collective bet. So, the cash flow line eventually snaps back upwards. And the commitment doesn't stop at the equipment Meta's already purchased. The chart here we can see Meta with roughly 279 billion of unccommenced lease commitments. doesn't mean Meta writes a 279 billion check tomorrow, but it does show how substantial the long-term infrastructure commitment has now become. There is however here an important counterargument. Meta says demand for compute itself is extremely strong. Management's even discussed receiving offers for compute capacity at multiples of what Meta paid for it. So excess infrastructure, it wouldn't necessarily be worthless. And it gives Meta another possible path towards return on investment, use the infrastructure internally, sell intelligence and enterprise services, or potentially monetize surplus compute. But here's the key phrase in Meta's own comments. We don't have a more definitive timeline to share right now. And markets, they hate uncertainty around a hundred billion dollar investments. And that brings us to the question I think matter far more than whether Meta pays another billion dollars in legal settlements. What return is the company going to generate on all of this AI infrastructure? Because if the answer is not much, today's free cash flow compression becomes structural. If the answer is a lot, today's spending could look brilliant 5 years from now. And this analyst explained the investment debate perfectly. ignore the quartertoquarter noise and listen to how directly free cash flow connects to return on invested capital. >> Um I just want to ask you about cash flow. I mean the expectation it was going to come in negative for the quarter. It came in at 784 million with an M which is uh incredibly low for compared to what they had normally been posting in quarters past. Are you concerned at all about that or is that just something you can look through? I I mean I I think the the the free cash flow story for Meta and for all the other hyperscalers goes handinhand with the ROIC story and investors will be willing to accept the the pullback in free cash flow if we're getting negative. Uh obviously to only to certain point um but but it's it's the free cash will pull back, the capback spend uh debt and equity raises. uh investors will be willing to be patient if they see the story on the ROIC side coming through as well. >> And that's exactly why I decided not to do my normal valuation today. Instead, let's value three completely different versions of Meta, same discount rate, same terminal assumption, same 10, 12, 14% sensitivity framework. The only major variable change is what happens to free cash flow as this AI investment matures. First, the bare case. And importantly, this isn't a scenario where Facebook disappears, Instagram doesn't collapse. Advertisers don't leave. The core business can remain excellent. The problem is simply that the incremental AI investment fails to earn an adequate return. So, I still allow Meta's core advertising engine to grow, but the spending remains much heavier than investors expected. Infrastructure depreciation builds, operating costs stay elevated, and new AI revenue arrives too slowly to offset it. That gives me 2026 free cash flow around 2 billion, 2027 at 17 billion, 2835 billion, 2029.50 and around 2030 at 65 billion. And keeping my normal growth rate framework exactly the same at 10% I get around $471 at 12% 501 and 14% 532. Now using that middle case here, the intrinsic value comes to $51 per share. That's around 15% downside, no margin of safety, in fact an 18% premium today. So if Zuckerberg ultimately fails to turn this infrastructure into meaningfully higher cash generation, I don't think Meta is cheap today. And that's important because it means investors buying at $592. They're already making some assumption that AI works. Then we move on to the base case, and this is broadly where I currently sit. I assume the near-term free cash flow pressure is real. 2026 remains ugly, but over time the core advertising machine continues compounding. Kabik's growth eventually moderates and AI improves monetization across Meta's existing platforms. And remember, we already have signs of this. More impressions, higher prices, higher revenue per user. AI does not need to create a completely separate 100 billion business to be economically valuable. If it improves the efficiency of an advertising platform already generating almost 60 billion every quarter, the incremental value can be huge. My base case assumptions are 2 billion in 26, 20 billion in 27, 50 billion in 28 and 75 in 2029 with 95 billion in 2030. Now this produces values of around $690 at 10%, $734 at 12 and 780 at 14%. So using the central case here, we get $734 against the share price today. That's around 24% upside and we can see a 19% margin of safety. Not spectacularly cheap, but enough upside that I still view the stock positively if my underlining assumptions are right today. And then we get to the most interesting possibility. What if Zuckerberg is right, not partially right? What if this huge infrastructure bill creates an entirely new layer of monetization on top of the advertising empire? And this is why I don't think investors should dismiss the bullish scenario. Only a few months ago, the route to AI monetization looked extremely vague. Now, we're beginning to see actual possibilities emerge. I think this has been the biggest concern from investors over the certainly the last six months is hey you're spending aggressively you're making these big commitments you're doing the things on paper that AI companies should be doing but what is the plan to make all that back and so uh that was a real concern a few months ago when they reported first quarter earnings since then we've seen them come out with a consumer subscription for their AI chatbot they're charging for AI agents uh they're building you know setting the framework internally for a potential cloud enterprise business maybe they're even going to sell some of the compute that they've accumulate ated. So they're really, you know, setting up all of these various businesses around AI that didn't exist just 2 3 4 months ago. And if you're an investor, you have to be saying at the very least, okay, we know where the company plans to get some of that investment back, which was not necessarily uh as clear at the beginning of the year. And that's the bull case. AI improves ad targeting, AI improves engagement, business messaging scales, and meta AI, well, it becomes monetizable. developer access becomes a genuine business. Compute becomes another revenue source and most importantly eventually the rate of infrastructure spending stops increasing faster than the cash those assets generate. Now under the scenario my free cash flow rises from 2 billion26 to 23 billion to 62 and then to 100 billion in 2029 ultimately 135 in 2030. Same methodology same growth rates and suddenly the valuation changes dramatically. 10% well it gives me around $977 12% $1,040 and 14% around $1,100. So if Zuckerberg is genuinely creating the next major profit engine inside Meta the stock doesn't look moderately undervalued. It looks massively undervalued from today's price. The medium case implies a roughly 76% upside with investors getting a 43% margin of safety. And that is why today's debate is so difficult. At first glance, Meta actually looks fairly inexpensive. The stock trades around 18.6 times Ford earnings on this measure. Its 5-year average is around 22.1 times. So investors are already applying some discount to Meta's history. And you can see when we look at the blue tunnel from Simply Safe Dividends, which highlights intrinsic fair price, it's sitting below even the bottom end that indicates a potential undervaluation signal. But always important to point out Meta as we can see going back over the last 10 years has spent many many periods trading in fact what I'd call a severely undervalued level. And when we look on the non-GAAP measure while the same story appears is sitting around 19 times forward earnings roughly 14% in fact below its own 5year average. So I wouldn't simply call the stock cheap and move on because we can see the earnings multiple isn't the only thing changing. Operating margins have started coming down from recent peaks. Cash conversion has weakened substantially and the company's capital requirements, they're very different from the meta investors owned only 2 or 3 years ago. The fascinating part, well, is that revenue continues climbing while valuation multiples compress. That's usually where interesting opportunities begin, but only if the cash ultimately follows the revenue. And Wall Street will capture this uncertainty almost perfectly. Their average price target $757, the low-end $580, the higherend $1,000. The range looks remarkably similar to what are three DCFs just produced. So where exactly do I land? Because I don't think the $51 bare case is the most likely outcome. Meta's core business, well, it's simply performing too well for me to assume the AI investment produces virtually no meaningful return. 3.6 6 billion daily users, revenue per person rising, ad volumes rising, pricing rising. Those aren't the numbers of a business being disrupted. They're the numbers of an advertising platform that is still getting more powerful. But I also don't think I'm ready to underwrite $1,040 as my central case. We still don't have enough evidence that Meta can create a major standalone AI business. The infrastructure commitment is huge. cash flows collapse in the near term and execution has clearly not been completely smooth. The bull case is absolutely possible and if meta begin showing real revenue from AI subscriptions agents enterprise products or compute, I'd happily revisit those assumptions. But today, I'm not paying for that outcome in advance. So, my current view remains closest to the $734 base case at around $592. I think Meta offers a reasonable margin of safety. Not the kind of obvious bargain we had during previous Meta crashes, but still enough upside for me to remain constructive. And crucially, this week's legal ruling doesn't make me suddenly bearish. The billion dollar headline is not the thesis breaker. What I'm watching is whether these cases eventually interfere with the engagement mechanics underpinning Meta's advertising economics. and on AI while I'm watching something even simpler. Does free cash flow begin recovering? Because Zuckerber can spend 130 billion, he can spend 145 billion, he can spend even more next year, but eventually those assets need to generate a return. If they don't, well, that's the bearish case here. And hence why we get the $51 intrinsic value. If things go broadly to plan, well, $734, that's the base case. And if Zuckerberg is right, more than $1,000, that is the range investors are buying into today. And for me, that's what makes Meta one of the most fascinating stocks in the entire market right now. Let me know in the comments which of those three scenarios you think is most realistic. Does Meta's AI spending ultimately prove to be another expensive distraction? Does the core business simply grow through it? Or does Zuckerberg end up building something far larger than the market currently expects? And don't forget, as always, we release a weekly newsletter covering severely undervalued stocks, as well as what's been in the market over the last few days. So, you can click below, sign up, and read straight away. More importantly, though, have a great day. I'll see you all on the next one.