Meta Platforms – Entering an AI Arms Race It Has No Business Fighting

Meta’s world-class ad machine remains intact, but its costly AI strategy threatens long-term returns.

That advertising and social media giant Meta Platforms operates one of the most remarkable business models in the world is no news. The company has mastered the art of monetizing global attention at scale, and its core Family of Apps continues to execute at an exceptional level.

But it’s been almost two years since I last covered Meta on InvestInsights, and a lot has happened since, much of it fundamentally reshaping the investment thesis.

Despite its operational strength, Meta’s stock has struggled. Shares are down roughly 19% over the past month and are up only 1.5% year-to-date, underperforming the S&P 500 by about 10 percentage points. And critically, the recent sell-off is not without justification. Beneath Meta’s brilliant advertising engine lies a far more complicated reality: a generational spending spree is now underway, one that brings significant execution risk, uncertain demand, and enormous financial implications.

Meta is betting heavily (and early) on long-term demand for AI and data centers to avoid falling behind its largest peers. This investment cycle is expected to meaningfully pressure margins, free cash flow, and ultimately the company’s medium-term earnings power, all while offering little visibility on when, or even if, these investments will pay off. In short, the Meta outlook is no longer the straightforward compounding story it once was.

Therefore, today, I want to update the Meta investment thesis by going over its most recent quarterly report to assess its performance and financials, before I address Meta’s outlook amid recent AI investments in much greater detail and update my medium-term projections and fair value estimate.

This is my Meta thesis update – let’s delve in!

---

Welcome to InvestInsights — an independent equity research publication rooted in long-term, buy-and-hold investing, publishing actionable stock/equity research reports weekly!

• 📈 You’re reading my latest [FREE] stock analysis. If you like this analysis, make sure to like & subscribe to receive much more like this!

• If you’d like full portfolio access (16% CAGR since 2022) & additional exclusive analyses, consider becoming a paid subscriber!

---

Meta delivered brilliant growth in Q3, but the P&L reveals Red Flags

I want to start by reviewing Meta’s Q3 results to assess its financials and performance. The company released its Q3 earnings on October 29, which weren’t particularly well received by investors, as shares lost over 10% in the following trading session. However, I would argue that the advertising giant delivered a very impressive Q3 performance, highlighting once again that it operates one of the most remarkable business models.

Although I also can’t deny that the report started to show some red flags.

Starting at the top, Meta reported total Q3 revenue of $51.2 billion, beating consensus estimates by a solid $1.83 billion and reflecting YoY growth of 26%, which is really exceptional – Meta has beaten the revenue consensus for every quarter for over 3 consecutive years now.

However, I am even more impressed by the level of growth Meta delivered in Q3. This accelerated further from 22% in Q2 and 16% in Q1, despite lapping 19% growth last year and working from a massive revenue base – yet this was the best growth in 1.5 years.

Meta accelerating and delivering mid-twenties YoY growth is exceptional.

Of course, the far majority of these revenues continue to be generated by Meta’s Family of Apps, which includes Facebook, Instagram, WhatsApp, and Threads. Revenue from these apps totaled $50.8 billion in Q3, up 26% YoY. Of that amount, $50.1 billion was pure advertising revenue. This scale is extraordinary, not just in absolute terms but because Meta is already the second-largest advertising platform on the planet.

Growth at this level, off a base this large, underscores something essential about the company: its business model is one of the most effective machines for monetizing user attention ever built, and it continues to compound.

Its engagement loops are self-reinforcing, as more content drives more usage, more usage drives more ad inventory, and more ad inventory attracts more businesses. Critically, usage only continues to improve!

Meta is still able to grow its Family of Apps user base, somehow. In Q3, this reached an estimated 3.54 billion unique people who use at least one of its apps, up another 8% YoY. Instagram hit 3 billion daily actives in the quarter, and Threads also continues to grow, hitting 150 million daily actives.

Remarkably, Meta continues to grow its Family of Apps unique users at a mid-to-high single-digit rate. And, crucially, not only does the number of users increase, but engagement does too, driven by Meta’s best-in-class content recommendation systems that continue to improve, keeping users engaged on its platform for longer and longer.

In Q3, this resulted in a 5% increase in time spent on Facebook, a 10% increase in time spent on Threads, and an acceleration in Instagram engagement growth. In the U.S., time spent on Facebook and Instagram even grew by double digits, driven by “continued video strength as well as healthy growth in non-video time on Facebook,” to quote management. This is all driven by consistent improvements in the company’s recommendation systems, which are already best-in-class.

Reels, in particular, continue to be a driver of engagement growth, with video time spent on Instagram up a whopping 30% YoY, bringing the advertising ARR for reels to over $50 billion alone, and with still room to grow.

So, not only does Meta somehow continue to grow its massive user base at a reasonable rate, but it is also rapidly growing user engagement on its platforms through better recommendations, leading to substantial gains in time spent.

All these numbers continue to trend in the right direction and show incredible traction.

Crucially, Meta also continues to be extremely capable in monetizing this growing engagement.

Q3 ad impressions were up 14% YoY, highlighting that the higher time spent on its platform and larger user base allow it to show more ads without pushing users away, avoiding oversaturation. Growth here was particularly strong in the Asia-Pacific region, with impressions up 23% YoY.

As highlighted below, these improvements in underlying dynamics allow for accelerating growth in impressions.

On top of more impressions, price per ad has also continued to grow at a healthy rate, up 10% YoY in Q3. The reason is simple: demand is high, and Meta’s pricing increases accordingly. Additionally, ad performance and advertiser ROI improve, justifying higher pricing. This led to double-digit price growth in Europe and North America.

Ultimately, this combination of Meta's exceptional innovation and execution, along with strong demand trends, led to a 26% increase in advertising revenue.

Simply put, Meta continues to execute really well within its Family of Apps, and underlying numbers reflect this!

Besides advertising, Meta also generated $690 million in other revenues from its Family of Apps, a 59% YoY increase, primarily driven by WhatsApp’s paid messaging revenue and Meta Verified subscriptions.

Additionally, Meta generated $470 million in revenue from its Reality Labs, up 74% YoY. This was driven by retail partners stocking up on Quest headsets ahead of the holiday season, as well as a healthy contribution from AI glasses.

On that note, let’s move to the bottom line. And whereas I was very impressed by Meta’s top-line performance and underlying dynamics in its Family of Apps, the P&L raises questions – even some emerging red flags.

For starters, Meta reported total Q3 expenses of $30.7 billion, up 32% YoY and outgrowing revenue. Furthermore, this is a 20 percentage point acceleration in expense growth compared to Q2, driven primarily by exorbitant growth in employee compensation from the hiring of AI talent and rapid increases in infrastructure operating costs associated with higher data center capacity.

In other words, what we see right here is a massive acceleration in cost growth due to Meta’s investments in data centers and AI, which should tell you this isn’t a one-off or temporary bump; these costs are here to stay and likely to accelerate further in quarters and years ahead – these are structural increases in cost.

That should be a concern.

For reference, R&D costs in Q3 were up 35% YoY. They grew to 30% of revenue, up 200 bps YoY, and G&A costs were up 88% YoY, to 7% of revenue, driven by an 8% higher employee count and higher average compensation, as Meta isn’t backing down from overpaying for top-tier talent. Some of this growth in expenses was offset by flat marketing and sales costs YoY.

On top of this, Meta’s reality labs losses also continue to mount, registering another $4.43 billion loss from operations in Q3, bringing cumulative losses to over $73 billion.

Ultimately, this led to an operating income of $20.5 billion in Q3, representing a 40% operating margin, down roughly 300 bps YoY, as costs outpaced revenue.

Moving further down the line, Meta also faced an exceptional 87% tax rate in Q3, although this was a one-time occurrence. To quote management;

“Our tax rate for the quarter was 87%, which was unfavorably impacted by a onetime noncash reduction in deferred tax assets that we no longer anticipate using under new U.S. tax law. Our tax rate would have been 14%, excluding this charge. Although the transition to the new U.S. tax law resulted in an accounting charge in the third quarter, we continue to expect we will recognize significant cash tax savings for the remainder of the current year and future years under the new law, and this quarter’s charge reflects the total expected impact from the transition to the new U.S. tax law.”

Nevertheless, this resulted in a net income of only $2.7 billion, or an EPS of $1.05. However, if we exclude this one-off, net income would be $18.6 billion at a 36.3% net income margin, down 240 bps YoY. This translates to an adjusted EPS of $7.25, which would beat the $6.71 consensus by a substantial margin.

Moving to CapEx, this was elevated to $19.4 billion in Q3, up 111% YoY, driven by heavy investments in servers, data centers, and network infrastructure, as Meta is all-in on the hyperscaler investment boom.

As shown below, Meta’s capex has exploded since Q4 2024, rising from a low-twenties percentage of revenue in 2024 to a high-thirties percentage in the latest quarter. Indeed, Meta is now investing 38% of revenue into CapEx. Those are extreme levels, as the trend is still on the rise. In fact, Meta is guiding to substantially faster CapEx growth in 2026 than in 2025.

Obviously, these far higher CapEx investments are dragging on FCF. Even as Meta grew revenue by a very impressive 26%, higher operating costs and exploding CapEx limited FCF to “just” $10.6 billion, down 32% YoY, reflecting a 21% FCF margin, down 17 percentage points YoY.

Yes, Meta’s cash flows are dropping like a brick, and this should worsen further in the quarters and years ahead.

That is a massive concern.

On a positive note, Meta’s Q3 FCF did fully cover its $3.2 billion in repurchases and $1.3 billion in dividends. Additionally, it ended the quarter with a healthy balance sheet containing $44.4 billion in available cash and $28.8 billion in debt.

Ultimately, I am of the opinion that Meta delivered excellent top-line Q3 results, showcasing strong operational performance as its Family of Apps continues to perform well. Meta's core business remains absolutely brilliant.

However, Meta’s P&L revealed emerging red flags, with rapidly escalating structural costs and high CapEx already pressuring profits and cash flows, as Meta is fully committed to AI.

And while I like Meta’s core business, it’s these AI commitments that concern me, raising serious doubts over its outlook for the several years ahead.

On that note, let’s delve into this outlook and break down Meta’s medium-term future.

---

---

Meta’s AI ambition is destroying the investment thesis entirely.

The reason for my concern over Meta’s medium-to long-term outlook, and the reason for my skepticism toward the business as revealed by the title of this analysis, is simple: the company is entering one of the most aggressive and prolonged investment cycles in its history at a time when the benefits of those investments remain uncertain, and the risks are unusually high.

You see, in recent months, it has become increasingly evident that Meta is fully committed to participating in the massive AI infrastructure cycle, now in its early stages, to become one of the leaders in the age of AI computing, alongside Amazon, Microsoft, Google, and Oracle. Meta wants to position itself as one of the world’s AI leaders – reinventing itself as a major AI ecosystem and compute leader – which means rebuilding its entire technical backbone and investing heavily in data center capacity to deploy and train its own models.

Of course, the potential upside of this ambition is obvious. Meta wants to join the LLM party with its own top-notch models, to be deployed everywhere. The company has a 3.5 billion-user base through which it can monetize such services or agents, creating a massive revenue opportunity if it can deliver competing-quality LLMs – arguably.

However, I can’t help but feel like this is much more of a defensive move from Meta than an offensive one. Meta is the standout in the high-spending group, the only company attempting AI at this scale without the natural monetization pathways of a cloud platform like AWS, Azure, GCP, or OCI. That alone forces an uncomfortable question: why build AI infrastructure of this magnitude when it isn’t inherently aligned with Meta’s core business model? It seems to start on the back foot.

To me, the answer is that Meta fears the cost of not participating even more than the cost of participating. It sees a future where falling behind in AI could weaken the relevance of its apps, its recommendation systems, and ultimately its advertising engine, or one where it becomes dependent on the AI models and computing of its big tech peers, similar to how it is already dependent on the operating systems of Apple, Google, and Microsoft to run its apps.

In this sense, Meta is building an AI supercomputer empire less because it extends its advantage, and more because opting out could risk strategic irrelevance. And when investment becomes primarily defensive, history suggests it is far more prone to value destruction than value creation.

On top of that, Meta seems to fear missing out on the next big computing opportunity - this looks like a strategic pivot at least in part driven by FOMO.

And that brings me to the cost of this endeavor: the financial implications and execution risks here are immense, complicating Meta’s medium-term outlook and any investment thesis.

Meta is now in the early stages of a generational spending spree that reshapes its entire financial profile. This shift could expose the business to meaningful pressure on margins, returns on invested capital, and, critically, investor patience – the result is a much weaker medium-term profit outlook and little visibility on when these investments will pay off.

Meta recently committed $600 billion in CapEx over the next three years to build out AI infrastructure in the U.S., which amounts to roughly:

• 2026 CapEx of $140 billion.

• 2027 CapEx of $200 billion

• 2028 CapEx of $260 billion

This aligns with management’s 2026 commentary, stating ”CapEx dollar growth will be notably larger in 2026 than 2025.” For reference, this compares to an expected $114 billion in net cash provided by operating activities in 2025 and an expected CapEx of $70 billion.

These are insane financial commitments.

And on top of these CapEx expenses, infrastructure operation costs will also balloon. As we saw in Meta’s Q3 results, this puts significant pressure on margins, with these costs far outpacing revenue growth from the Family of Apps.

So, the eventual costs of these efforts will be massive. These investments will most certainly continue to drive up operating costs, leading to lower operating margins, and I expect the addition of capex to eventually result in negative FCF, based on management’s current spending plans. This will likely lead to growing debt as well, to fund its expansion plans, with FCF turning negative. We can already see this happening, with Meta recently announcing a bond offering worth up to $30 billion.

In other words, I expect depressed cash flow and growing debt levels through at least 2028, but likely into the next decade – I expect the amount of debt needed to fund its plans to be significant.

Meta is participating in an AI spending race it has no place in.

And then we still have the question of whether Meta can monetize these investments and, if so, in what time frame. Honestly, management’s strategy on this front is shocking to me. Here is a quote from Mark Zuckerberg:

“think that it’s the right strategy to aggressively front-load building capacity so that way we’re prepared for the most optimistic cases.”

I absolutely hate that strategy. You are telling me you are going to spend $600 billion in CapEx alone over the next three years, entirely depressing cash flows and earnings, as well as worsening your balance sheet, and you don’t actually know when these investments will pay off, if at all.

Again, seems like a lack of strategy driven by FOMO. That is a massive red flag to me, no matter how good your core business is.

“That way, if superintelligence arrives sooner, we will be ideally positioned for a generational paradigm shift in many large opportunities. If it takes longer, then we’ll use the extra compute to accelerate our core business, which continues to be able to profitably use much more compute than we’ve been able to throw at it. And we’re seeing very high demand for additional compute, both internally and externally. And in the worst case, we were just slow building new infrastructure for some period while we grow into what we build.”

This mindset is mind-boggling to me.

What makes the situation even more precarious is the lack of a proven monetization path for AI that can justify this level of spending in the first place.

For one, even if Meta succeeds in building powerful models, there is no guarantee it will win in distribution. Consumers increasingly encounter AI through OS-level interfaces, search platforms, and productivity environments, areas firmly controlled by Apple, Google, and Microsoft. If these companies own the primary gateways for AI queries and agent-based interactions, Meta’s own AI surfaces could become secondary layers at best. That would sharply limit the revenue potential of its agents and diminish the payoff from the massive GPU and data-center build-out.

I simply don’t see it as the winner in the AI agents battle, significantly capping its revenue potential.

This leads to a broader, more troubling possibility: Meta may struggle to monetize these investments at all, or may do so only on a much longer timeline than the company appears to assume.

Another critical bear-case concern is the growing doubt that real-world AI demand will ever justify this scale of spending in general.

Across industries, generative AI adoption remains limited and inconsistent. Most deployments are experimental; productivity gains are uneven; and integration issues (hallucinations, compliance risks, data governance challenges, and workflow disruption) continue to slow meaningful adoption.

The economic benefits of AI remain far behind the hype.

This creates a major timing risk. Meta is investing as if AI usage is about to inflect sharply upward, yet the actual adoption curve could be slower, flatter, or less transformative than expected. If enterprises do not scale AI usage materially in the next few years, Meta could be left with specialized, rapidly depreciating infrastructure that sits underutilized.

Because AI-optimized hardware quickly becomes obsolete, any period of excess capacity directly depresses returns and prolongs the drag on free cash flow.

I think my skepticism here is pretty clear. The risk of poor returns on invested capital, depressed earnings, prolonged, structurally weak free cash flow, and a balance sheet burdened by years of spending with little incremental value created is extremely high here, in my opinion.

If AI adoption fails to accelerate meaningfully or even takes longer than Meta is betting on, the company risks being left with an oversized, capital-intensive infrastructure footprint that produces far less economic return than its cost implies.

Long story short, Meta is in an AI infrastructure race it has no place in, and the risk-reward, in my view, is skewed heavily toward the negative.

The result is a complicated, mostly unattractive investment thesis, as well as a clouded, worsening medium-term outlook.

No, I am not a fan. This is increasingly looking like a situation that I would actively avoid investing in at practically any price.

On that note, let’s jump to Meta’s medium-term outlook!

---

Want more out of your subscription? Even more content like this weekly?

Consider InvestInsights PRO - $7.50/month ($70/annually)

This gets you:

• A guaranteed 6+ stock analyses every month (of which at least 2-4 are paid-exclusive).

• Full insight into my own portfolio, including allocation, transactions, watchlist, performance, and relevant thoughts (16% return CAGR since January 2022).

• Instant transaction alerts anytime I buy or sell any shares (Fully transparent).

• A complete overview of all my target prices and ratings (online available).

• Access to the InvestInsights Hub, containing all of the above in a single online sheet, updated instantly!

---

Outlook & Valuation

Starting with guidance, management now expects Q4 revenue to be in the range of $56 billion to $59 billion, suggesting YoY growth of 19% or a somewhat slower pace than recent quarters. This assumes continued strong ad revenue momentum, offset by lower Reality Labs revenue due to a tough YoY comparison. Do keep in mind that Meta management tends to be cautious with its guidance.

For 2025, this translates to an expectation of $198.6 billion in revenue, assuming the midpoint of Q4 guidance. Furthermore, management guides FY25 expenses in the range of $116 billion to $118 billion and CapEx in the range of $70 billion to $72 billion, both at the higher end of prior guidance, driven by increased AI-related investments.

As for my own projections, I am assuming Q4 revenue to be at the high end of the guided range, growing 21.4% YoY to $199.7 billion, mainly driven by upside to management’s advertising guidance. Meanwhile, I expect EPS to grow by just over 10% in 2025, as margins slip due to cost pressure and management scales back buybacks to preserve its balance sheet.

Looking ahead to next year, management provided some commentary on costs, noting an intention to invest aggressively in its own infrastructure and to contract with third-party cloud providers. This leads to the expectation that CapEx dollar growth “will be notably larger in 2026 than 2025,” suggesting a 2026 CapEx of well over $100 billion. Similarly, operating expenses will grow at a significantly faster rate in 2026 compared to 2025, mainly due to rapid increases in infrastructure and employee compensation costs, similar to Q3 but ramping up quickly.

Given this commentary on 2026, I believe it is safe to assume that operating expenses will outpace revenue, putting pressure on profits. I now expect Meta to deliver 17% revenue growth in 2026, with advertising growth normalizing. However, EPS will grow considerably slower, driven by pressure on margins due to the aforementioned cost growth in relation to Meta’s AI efforts.

Through 2028, I expect this trend to persist, with revenue growth remaining very strong in the mid-teens but slowing as advertising conditions normalize and newer verticals such as reels mature. Still, this is remarkable growth.

Yet, I also expect pressure on profits and cash flows to increase in 2027 and 2028, as I don’t expect any monetization from AI investments, while its spending plans suggest mounting investments and, consequently, higher costs. As a result, I expect EPS growth to fall well into the single digits, while FCF is likely to turn negative in 2027 and 2028, assuming management remains committed to its spending plans. However, we might see Zuckerberg pull back on some commitments as shareholder pressure might grow.

For now, the graph below reflects my forecasts.

Moving to valuation, the one positive here is that the recent sell-off has made Meta shares less expensive. With shares flat YTD and down 19% over the last month alone, some of this worsening outlook has been priced in, though arguably not enough.

At a current share price of $598, Meta shares still trade at roughly 23x this year’s earnings and 20.5x next year’s earnings, which I find far from cheap for a highly cyclical advertising business expected to grow earnings at a high-single digit CAGR through 2028, at best.

For reference, if we adjust this for forward growth projections, we are looking at a PEG of almost 3x, almost double its 5-year average, and more than double the sector median.

Even after the correction, Meta still trades at a valuation that embeds a level of confidence in its long-term trajectory that I simply don’t share.

A ~20x forward earnings multiple might look reasonable for a dominant, asset-light software business with stable margins, but that is not the Meta we are dealing with anymore. Meta is transitioning into a capital-intensive AI-infrastructure company with structurally lower free cash flow, rising debt, and a decade-long investment cycle that materially weakens its near-term profitability. Yet the market continues to value it as though its historical margin profile remains intact.

And when we look at valuation through the lens of free cash flow, a far more relevant metric for a business embarking on an unprecedented capex cycle, the picture becomes even less attractive. Free cash flow is set to decline meaningfully in the coming years, with a realistic possibility of turning negative by 2027 or 2028. Valuing Meta on FCF makes the current multiple look stretched by any reasonable standard, especially for a business facing this level of uncertainty.

In my view, the current valuation only makes sense if you believe Meta will emerge from this capex cycle with higher long-term margins, a durable AI advantage, and a clear monetization pathway that justifies its spending. I see very little evidence of that today.

Instead, we have falling free cash flow, mounting financial commitments, deteriorating visibility, and a massive investment program built on uncertain returns and a strategy that seems built on FOMO and hope.

Unless Meta can reveal a much clearer investment strategy and a way to monetize these investments, or changes the direction it’s heading in and adjusts spending plans accordingly, I don’t think Meta is a compelling investment opportunity at practically any price, given the heightened risk and uncertainty.

These new endeavors fully overshadow Meta’s brilliant core business, and the result is a business that simply doesn’t align with my investment framework for a good buy-and-hold.

Per illustration, based on an 18x 2027 exit multiple, I calculate an end-of-2027 target price of $566 and a current-day fair value of $450, down another 25% from today’s share price.

Meta Platforms is a clear avoid in my book.

Rating: Avoid - Accumulate below $450
FY27 Target Price: $566
Implied CAGR from the current price: -

---

Please leave a like if you enjoyed this post and found it valuable!

---

---

Disclosure: I/we may or may not have a beneficial long position in any of the securities discussed in this post, either through stock ownership, options, or other derivatives. This article expresses our own opinions, and we are not receiving any sort of compensation for it.

No recommendation or advice is being given as to whether any investment is suitable for a particular investor. The information provided in this analysis is for educational and informational purposes only. It is not intended as and should not be considered investment advice or a recommendation to buy or sell any security.

Investing in stocks and securities involves risks, and past performance is not indicative of future results. Readers are advised to conduct their own research before making any investment decisions.