META surges 11%, but free cash flow evaporates by 91%

NEXT PICK · Market Insights

META surges 11%, but free cash flow evaporates by 91%

Monday, September 21, 2026

On September 21, META surged from a previous closing price of $665.75 to $741.25, surging 11.34% in a single day, with trading volume expanding to 2.6 times the three-month daily average, marking the largest single-day gain since July 2025.

What ignited this big bullish candlestick was neither a financial report nor a merger or acquisition, but rather an AI assistant named Muse, which launched less than two weeks ago, surpassing ChatGPT, Claude, and Grok, and with Wells Fargo analysts pushing the target price from $640 to $796 in one go.

But just two months ago, on July 29, the same company had just collapsed from $8.55 billion in quarterly free cash flow to $784 million, a 91% year-on-year decline, and was slashed about 10% by the after-hours market.

The same company, in two months, went from "cash flow collapse" to "AI king's return." What real changes occurred in between, or did nothing happen, just a different story?

Latest price $741.25 ▲ +11.34%

Q2 free cash flow $784 million ▼ -91%

2026 capital expenditure $130–145 billion ▲ Guidance increased

3.6 billion people's attention, 97% of advertising dependence

Meta's revenue structure is almost dangerously simple: of the $60.8 billion in Q2 2026 revenue, $59.36 billion comes from advertising, accounting for about 97.6%. The remaining 2% is just over 2%, which belongs to Reality Labs, which produces Quest headsets, Ray-Ban Meta smart glasses, and Neural Band wristbands.

In other words, this company, repeatedly dubbed the "metaverse" and "AI giant," is essentially still a machine that wholesales human attention to advertisers, and this machine has been spinning faster recently than in recent years.

Its fuel is 3.6 billion people. As of June 2026, Meta's full suite of daily active users (DAPs) reached 3.6 billion, a 3% year-over-year increase; During the same period, ad impressions grew by 14%, and average ad price per ad rose by 12%—a rare combination of volume and price among mature ad platforms.

So where exactly is the moat? It's not in the algorithm, nor in the server, but in the literal meaning of the term "migration cost": you can uninstall Instagram, but you can't move your follower list; you can register a new social platform, but you can't take away ten years of family group chats on WhatsApp.

The advertiser's side is no different. Once a small and medium-sized merchant masters the placement model, accumulates conversion data, and develops creative intuition on Meta Business Suite, switching platforms means restarting these assets from scratch—this is not functional dependency, but data lock-in.

Now, Meta AI is embedded in all its apps, Muse is operating independently as an app, and with the upcoming integration of glasses on Meta Connect, this ecosystem is shifting from a "social product matrix" to a three-layer closed loop of "software—hardware—model." The problem is, the bigger the closed loop, the more money is burned.

A price-to-earnings ratio of 27.9 times isn't expensive, but is this EPS clean?

Based on a closing price of $741.25 and earnings per share of $26.56 over the past four quarters, META's price-to-earnings ratio is about 27.9 times; With a market cap of $1.89 trillion, it corresponds to about $228.2 billion in revenue over the past four quarters, with a price-to-sales ratio of about 8.3 times. Looking at these two numbers alone, they are not considered expensive among the seven major U.S. stock companies.

But this $26.56 EPS is not clean. In Q3 2025, Meta set aside a one-time non-cash income tax expense of $15.93 billion due to the implementation of the "Big and Beautiful Act," causing net profit to crash to $2.7 billion, an 83% year-over-year plunge; and in Q1 2026, the same bill delivered $8.03 billion in tax incentives, pushing quarterly net profit to $26.8 billion, or $10.44 per share.

On one side is huge tax write-downs, on the other is huge tax rebates—neither is money earned from core business. So here's the question: when the denominator itself is shaken back and forth by tax laws, how much reference value is left for the P/E ratio anchor?

What's even more worth watching is cash flow, which tells a different story. Free cash flow for the full year 2025 was $46.1 billion, down about 14.7% year-over-year; In the first half of 2026, the combined two quarters were only $13.18 billion—$12.4 billion in Q1 and $784 million in Q2.

The collapse is not on the operating side. In the second quarter, operating cash flow actually grew 25% year-on-year to $31.86 billion, but capital expenditure during the same period reached $31.08 billion, swallowing up about 98% of that. The more profitable the operating side, the heavier the infrastructure spending, leaving shareholders with only a fraction of free cash flow.

Based on a market value of $1.89 trillion and free cash flow for the full year of 2025, the free cash flow yield is only about 2.4%, and this figure is likely to be further compressed in 2026. The company still has $90.3 billion in cash and securities on hand, and $83.7 billion in long-term debt—enough ammunition, but no longer as calm as before.

Among the three discounted cash flow models presented in the report, the benchmark scenario is about $640 per share intrinsic value, the optimistic scenario is $900 to $950, and the pessimistic scenario is $400 to $450. It should be noted that the baseline scenario assumes $58.8 billion in free cash flow for the full year 2026, but only $13.18 billion actually came out in the first half—this assumption itself is quite aggressive.

There is also a widely circulated claim that needs to be corrected

The commonly cited "META's PEG is only 0.68" is based on a 43.69% revenue growth rate. Meta's full-year 2025 revenue is $200.97 billion, a year-over-year growth rate of 22.1%, not 43.69%. Recalculated at 22%, PEG is close to 1.3; if calculated at 28% year-over-year growth in Q2, it's also around 1.0—cheap, but not that cheap.

Today's 11% increase has pre-priced an AI assistant that hasn't generated a single cent in downloads into its $1.89 trillion market cap. Is this really a discount for the future, or a payment for emotions? For specific buy and sell levels and real-time signals, see the official account.

Stock prices during breakout periods, financial statements during investment periods

From the moving average structure, the market clearly voted in favor: the closing price at $741.25 was about 22% above the 50-day moving average at $607.38, and about 19% above the 200-day moving average at $624.16. The short-, medium, and long-term moving averages were in a bullish alignment, which is a typical breakout pattern.

Fundamentals are indeed accelerating. Year-on-year revenue growth is expected to rise from 15.0% in fiscal year 2024 to 22.1% in fiscal year 2025, then 33% in Q1 2026 and 28% in Q2 2026, marking four consecutive quarters of significant upward growth; The company's guidance for the third quarter is $61 to $64 billion, which continues to be maintained.

But the other half of the report looks like an investment period. Operating margin plunged from 43% in the same period last year to 31% in the second quarter, and total costs and expenses surged 55% year-over-year to $42.03 billion, including $2.4 billion in legal-related provisions and $1.18 billion in May layoffs in severance costs.

So META is currently in a state of tearing: the stock price is following a breakout phase, while the profit statement is following a heavy asset investment phase. When these two lines converge will be the biggest focus for this stock over the next 12 months.

No rival can replicate 3.6 billion daily active users, but every rival is competing for more time

Let's start with the conclusion

In terms of "user base × ad tech platform," Meta currently has no single competitor to compete head-on—3.6 billion daily active users combined with Meta Business Suite and WhatsApp Business's complete delivery toolchain, making it a unique asset worldwide.

But every flank is being exploited. Alphabet holds search and YouTube and naturally dominates user intent data, while YouTube Shorts continue to erode Instagram's Reels short video share; And in search advertising, the lucrative battleground, Meta simply has no qualification to challenge in the short term.

TikTok is the sharpest blade. With over 1.5 billion monthly active users worldwide and a recognized stronger recommendation engine, it continues to outperform Reels in both watch time and engagement rate, targeting not ad budgets but the attention of lower-tier young users.

Amazon's threat lies in closed loops. Its advertising business has grown into the world's third pole, and its complete data chain from search to order makes conversion attribution far more accurate than Meta's brand advertising model; What's worse, AWS's cloud revenue provides its AI infrastructure with organic cash flow—which is exactly what Meta currently lacks.

Microsoft poses no threat in advertising volume, and LinkedIn barely overlaps with Meta, but the two directly compete in AI computing power procurement: both compete for NVIDIA, AMD, and Broadcom's capacity, and whoever gets priority delivery wins time. Snap is more like a mirror, clever in AR filters and targeting young people, but the scale gap is too large to form real competition.

If I had to use a metaphor,

Meta is like a super port built at a river mouth—no one can replicate its throughput, but for every additional upstream dock, there is one less ship passing through. So the question is, when AI assistants start becoming new entrances, will the river at this port still be rerouted?

A penetration rate of 26%, with a denominator being redrawn by AI

By industry standards, the global digital advertising market is expected to reach about $740 billion in 2024 and is expected to surpass $1.1 trillion by 2027, with a compound annual growth rate of 11% to 14%—this is Meta's total reachable market.

Further down, social media ad spend plus embedded ads on e-commerce platforms together make up Meta's serviceable market, totaling about $330 billion. Based on about $196 billion in advertising revenue for fiscal year 2025, Meta's penetration rate in this pool is close to 59%, accounting for about 26.5% of the entire digital advertising market.

What does it mean for penetration to exceed half? It means the growth space for relying on capturing existing market share is narrowing, and real incremental growth must come from three directions: the commercialization of WhatsApp Business, a surge in volume of Threads ad products, and AI-optimized continuous increases in the average price per ad.

And AI is the most subtle—it can either enlarge the numerator (more precise targeting, higher clicks and conversions, more expensive CPM), or redraw the denominator. If in the future users increasingly obtain information through AI assistants rather than information feeds, the entire "impression × per unit price" advertising model itself will need to be reconsidered.

The advertising engine is accelerating, and it's a kind of acceleration where both volume and price increase

The logic of the bulls is not complicated

The core business not only did not decline, but with AI's support, it achieved the best slope in recent years, while the market's valuation still remains at the level of a "mature advertising stock."

▲ Bull Case

Free cash flow evaporated by 91%, while the stock price was cheering for a download ranking

The core of the bears isn't "the company is doing well," but "when will this debt be settled"—AI infrastructure money is actually spent, but the returns remain stuck in PPTs and download rankings.

▼ Bear Case

Three tough battles: Connect press conference, Q3 earnings report, and 2027 capital expenditure guidance

Over the next two months, META's stock price will be dominated by three types of events, the most recent occurring within 48 hours:

•	On September 23, Meta Connect 2026 will open, featuring Zuckerberg's keynote speech, which is expected to highlight Muse's integration with the new generation of cameraless smart glasses—a bullish direction, but today's gains have already priced in much of the expected gains
•	Q3 2026 financial report (expected late October): Guidance range is $61 to $64 billion, not whether revenue meets targets, but whether operating margin can stop the slide from 43% to 31%—the impact is neutral and highly volatile
•	Capital expenditure execution for full-year 2026 and guidance for 2027: Only $50.88 billion was spent in the first half. To meet the $130 to $145 billion guideline, the second half must accelerate dramatically, which will directly determine the depth of free cash flow collapse—resulting in a bearish impact
•	Muse's commercialization path disclosure: Currently, only downloads are available, no monetization plan. If subscription or ad monetization paths are presented at Connect or Q3 calls, it would be a true catalyst for revaluation—with a strong impact
•	Reality Labs' new hardware sales data: If AI glasses can reach scale, they will provide financial evidence for the "second growth curve" for the first time—with a strong impact but a long time window

From three years to three months, to twelve months of stagnation

2022 was META's worst year. Apple's ATT privacy policy severely damaged ad targeting efficiency, Reality Labs lost over $10 billion annually, the metaverse narrative was thoroughly cold-looked, the stock price retreated about 75% from its peak, and the price-to-earnings ratio once dropped to 10 to 15 times—at that time, the market doubted whether the social advertising business itself could still survive.

The "year of efficiency" in 2023 marked a self-rescue: laying off over 21,000 employees, cutting redundant spending, and ramping up AI infrastructure. The stock rebounded from a low of about $88 to around $350 by year-end, nearly tripling its value. By fiscal year 2024, revenue reached $164.5 billion, up 15.0% year-on-year, with net profit of $62.36 billion, and valuations returning to the 25-28 times range.

The real turning point will come in 2025. Although revenue accelerated to $200.97 billion, a year-on-year increase of 22.1%, net profit slightly declined to $60.5 billion, and free cash flow dropped from about $54.1 billion to $46.1 billion—the first clear divergence between growth and cash flow.

And so, today's strange situation has emerged

The stock price rose from about $560 at the start of the year to $741.25, a sharp rise over the past month, but even after twelve full months, the return rate remained negative. The market spent a year repeatedly oscillating between the two interpretations "AI is a moat" and "AI is a bottomless pit," with no definitive conclusion to this day.

⚠️ Risk Notice • The payback cycle for AI capital expenditure is uncertain, and there is no timeline for converting $130 to $145 billion in investment into revenue • Free cash flow remains under pressure, with only $13.18 billion in the first half of the year, significantly squeezing shareholder returns and buyback opportunities • Antitrust, data privacy, and algorithmic regulatory pressures have become routine, with $2.4 billion in legal-related fees already set aside in Q2 • Changes in tax laws have caused significant fluctuations in quarterly earnings, severely diluting the reference value of price-to-earnings ratios and earnings per share • Reality Labs has been operating at a loss for a long time, and there is a risk of delays at the commercialization turning point between AI glasses and Quest • The short-term rise is too fast and nearly 19% above the 200-day moving average, and earnings reports or press conferences that fall short of expectations can easily trigger a rapid pullback

🟡 Neutral The ad engine is indeed accelerating, but the cash flow balance will take two quarters to be fully calculated.

💬 Discussion Narratives should lead cash flow, so it's best not to chase highs; Post META to the official account for real-time buy and sell positions.

Data source

•	Source: NextPick real-time market snapshot + Meta Investor Relations official website quarterly earnings report and guidance + SEC disclosure documents + mainstream financial media coverage on the day.