Meta: $60 Billion in Sales, Less Than $1 Billion Left

Meta just posted the highest revenue in its history and almost none of it survived.

In the second quarter of 2026, Meta revenue reached an all-time high of $60.8 billion. The Meta business model continued printing money at a scale almost no company has ever matched. Yet Meta AI spending exploded so violently that free cash flow collapsed to just $784 million a rounding error against $60.8 billion in sales. This is not a typo. It is not an accounting footnote. It is the headline number from Meta’s own earnings release dated July 29, 2026.

Read that again. Meta generated $31.86 billion in cash from operations in a single quarter. After paying for property, equipment, and finance leases, less than one billion dollars remained.

A company that spent most of the last decade as one of the most reliable cash machines ever built has deliberately turned itself into a capital-intensive infrastructure business in public, on purpose, while its own investors flinched.

The question is no longer whether Meta is big. The only question that matters is this: where did the money come from, where is it going now, and does the paper trail prove the bet is actually working?

Where the Financial Story Really Started

The origin story most people know is the Harvard dorm room in 2004. The financial story starts later, and it starts with a question that nearly broke the company: how do you charge anyone for something you give away free?

For its first several years, Facebook had users but not a monetization engine. The answer it eventually landed on auction-based advertising, priced by attention and sharpened by behavioral data was not invented by Meta. Google got there first. What Meta built was a version tuned for a different asset: not what people typed into a search box, but who they were, who they knew, and what they lingered on.

That distinction is the foundation of the entire Meta business model, and it explains almost every strategic decision that followed. Meta does not sell software. It does not, for the most part, sell hardware profitably. It sells access to attention, and it sells the targeting and measurement tools that make that attention worth paying for. Everything else the company has ever done has been either a way to capture more attention or a way to protect the machinery that converts attention into revenue.

Two early acquisitions show how well management understood this. In 2012, Facebook agreed to buy Instagram for a headline price of about $1 billion. By the time the deal closed that October, Facebook’s post-IPO share price had fallen so far that the final consideration came in at roughly $715 million in cash and stock, according to contemporaneous reporting. Two years later, the company announced the WhatsApp acquisition at a widely reported $19 billion; because much of the consideration was in stock and the share price rose before closing, the final booked value came in around $21.8 billion. Also in 2014, Meta bought Oculus VR, the deal that would eventually seed Reality Labs.

Here is the part worth sitting with. Instagram bought for well under a billion dollars became one of the most economically important assets in the modern internet. WhatsApp cost roughly thirty times as much and, more than a decade later, still has no publicly disclosed standalone revenue line. Meta does not break out per-app financials in its SEC filings, which means any confident claim about what Instagram or WhatsApp “earns” is an estimate, not a verified fact. What the public record does establish is that both were defensive as much as offensive: they removed two credible challengers from the board.

The First Real Breakthrough Was Not a Product It Was a Pivot

The single most consequential financial decision in Meta’s history may be the mobile transition. When Facebook went public in May 2012, its prospectus flagged the company’s failure to meaningfully monetize mobile as a material risk. The stock was punished for it. Management responded by rebuilding the ad product around the mobile news feed rather than trying to port desktop banners onto small screens.

Before: a desktop advertising business with a structural threat to its future. After: a mobile-first ad system that grew into the dominant share of revenue. The lesson is not “mobile was important” everyone knew that. The lesson is that Meta was willing to cannibalize a working product to chase a format it did not yet know how to monetize. That pattern repeats.

The financial results speak for themselves. Full-year 2024 revenue was $164.5 billion. Full-year 2025 revenue was $200.97 billion, up 22% year over year, with income from operations of $83.28 billion and an operating margin of 41%, per Meta’s fourth-quarter and full-year 2025 results. Net income for 2025 was $60.46 billion.

Notice something odd in those numbers. Revenue grew 22% and operating income grew 20%, but net income fell 3% versus 2024. Why? Taxes. Meta disclosed that its full-year 2025 effective tax rate was 30%, and explicitly stated that absent a valuation allowance charge tied to the implementation of the One Big Beautiful Bill Act during the third quarter of 2025, the rate would have been 13 percentage-point-adjusted down to roughly 13%. That is a textbook example of why revenue, operating income, and net income are three different things that should never be used interchangeably.

How the Income Machine Actually Works

Strip away the branding and Meta’s economics reduce to a simple loop with four moving parts.

Users supply attention and behavioral signal at zero monetary cost. Ranking and recommendation systems decide what each person sees, which determines how much attention is captured and how many ad slots can be sold. Advertisers bid in an auction for those slots. Measurement tools tell advertisers whether the money worked, which determines whether they come back and bid higher next quarter.

Two disclosed metrics tell you which lever is doing the work in any given period: ad impressions delivered and average price per ad. For full-year 2025, impressions rose 12% and average price per ad rose 9%. In the second quarter of 2026, impressions rose 14% and price per ad rose 12%. Price growth outpacing volume growth is meaningful it suggests advertisers were willing to pay more per unit of attention, which is generally a sign that targeting and conversion tools were delivering, not just that inventory expanded.

Meanwhile, Family daily active people reached 3.60 billion on average for June 2026, up 3% year over year. Compare that to the 7% year-over-year growth reported for December 2025 and you see the arithmetic reality of a company that has already reached a large share of the connected world. Meta cannot grow revenue much longer by adding people. It has to grow revenue per person.

That single constraint explains the AI spending better than any executive quote could.

That puts Meta in the same broad category as other attention platforms, where revenue is easier to see than the full standalone economics. YouTube’s revenue story shows how much can remain undisclosed even when a platform operates at global scale.

The Money Machine Gets Interrupted: The 2022 Shock

Every good financial story has a year when the assumptions break. For Meta, it was 2022.

Apple’s App Tracking Transparency framework, rolled out from 2021, required apps to ask users for permission before tracking them across other apps and websites. Most users said no. In February 2022, Meta told investors the change would reduce its 2022 sales by roughly $10 billion. That figure is a company estimate, not an independently audited number, and it should be treated as such but the direction was undeniable. The targeting signal that made the ad auction so valuable had been partially cut off by a company Meta did not control.

At the same time, TikTok was absorbing short-form video attention, and the newly renamed Meta was pouring money into Reality Labs. The stock fell roughly 64% over calendar 2022, according to market data.

The response is what makes this a case study rather than a cautionary tale. Management declared a “year of efficiency” in 2023, cut headcount and projects, and simultaneously rebuilt its ad targeting using machine learning to infer what it could no longer directly observe. In other words: it replaced lost data with better math. Revenue and margins recovered sharply over the following years.

The lesson buried here is uncomfortable but real. Meta’s greatest structural vulnerability was never a competitor. It was platform dependency the fact that its distribution runs through operating systems owned by other companies. That risk has not gone away.

Where the Money Is Going Now

Which brings us back to the number that opened this piece.

In the fourth quarter and full year 2025 release, Meta guided 2026 capital expenditures, including principal payments on finance leases, to $115–135 billion. By the first-quarter 2026 report in April, that range had been raised to $125–145 billion. By the second-quarter report in July 2026, it was narrowed to $130–145 billion. Full-year 2026 total expenses were guided to $165–169 billion. For context, actual full-year 2025 capital expenditures were $72.22 billion.

So the company roughly doubled its capital spending in a single year, and revised the number upward twice in six months.

You can see the effect on the balance sheet directly. Property and equipment, net, stood at $225.72 billion as of June 30, 2026, up from $176.40 billion at the end of 2025 nearly $50 billion of hard assets added in six months. Total assets reached $449.96 billion. Long-term debt climbed to $83.66 billion from $58.74 billion over the same period.

And you can see it on the income statement. Second-quarter 2026 research and development expense was $21.66 billion, up from $12.94 billion a year earlier roughly 36% of revenue, against 28.5% for full-year 2025. Total costs and expenses rose 55% year over year while revenue rose 28%. The result: operating income fell 8% to $18.78 billion, operating margin compressed from 43% to 31%, and net income fell 14% to $15.85 billion despite record revenue.

Two one-off items contributed: $2.40 billion of charges related to legal proceedings and $1.18 billion of severance connected to a May 2026 headcount reduction. Meta reported headcount of 75,472 as of June 30, 2026, including approximately 8,000 employees affected by that reduction who would largely disappear from the count by the third quarter. Press reporting placed the cut at about 10% of the workforce.

Strip out the one-timers and the picture is still the same in direction, if not degree: this is a company converting current profit into future capacity, and asking shareholders to wait.

How Meta Is Paying For It

Here is where the financing story gets genuinely interesting, because Meta did not fund this out of the cash pile alone.

In October 2025, Meta and funds managed by Blue Owl Capital entered a joint venture to develop and own the Hyperion data center campus in Richland Parish, Louisiana, a transaction reported at approximately $27 billion. By July 2026, CNBC reported Meta’s total Louisiana data center commitment had reached roughly $50 billion. Days after the Blue Owl announcement, on October 30, 2025, Meta sold $30 billion of corporate bonds its largest offering ever after reportedly attracting around $125 billion in orders, per Bloomberg and Reuters.

Read those two transactions together and a strategy emerges. The joint venture structure moves a portion of the development off Meta’s own balance sheet while preserving access to the capacity. The bond sale locks in long-term debt financing at a moment of extraordinary investor appetite. Cash, cash equivalents, and marketable securities actually rose to $90.26 billion as of June 30, 2026, from $81.59 billion at the end of 2025 not because operations threw off spare cash, but because the company raised it.

This is a meaningful shift in financial character. For years, Meta was a nearly debt-free business returning capital to shareholders $26.26 billion of buybacks and $5.32 billion of dividends in 2025 alone. In the fourth quarter of 2025, share repurchases were nil. In the second quarter of 2026, the company paid $1.35 billion in dividends and the release discloses no buyback figure in the highlights. Capital that once flowed to shareholders is now flowing into concrete, silicon, and power.

The Investment Nobody Can Score Yet: Reality Labs

Before AI, there was the metaverse. And the metaverse is the clearest documented example of Meta’s willingness to lose enormous sums for a long time.

Reality Labs recorded a full-year 2025 operating loss of $19.19 billion, compared with $17.73 billion in 2024, according to segment reporting summarized by industry press. In the fourth quarter of 2025 alone, the unit lost $6.02 billion on $955 million of revenue. CNBC’s running tally put cumulative Reality Labs operating losses since late 2020 at over $80 billion by the first quarter of 2026.

Eighty billion dollars. That is more than the total lifetime revenue of most companies in the S&P 500.

What did it buy? Honestly, the public evidence is mixed and should be described that way. It bought a hardware manufacturing capability, a research base in displays and sensors, and arguably most valuable the smart glasses product line, which press coverage has treated far more favorably than the headset business. It did not buy a self-sustaining commercial segment. Meta’s own 2026 guidance stated it expected Reality Labs operating losses to remain similar to 2025 levels.

Was the 2021 rebrand and the metaverse pivot a mistake? Judged only by segment results, it looks like one of the largest voluntary capital destructions in corporate history. Judged fairly, the picture is murkier: Meta was trying to escape the platform dependency that Apple had just demonstrated it could exploit. Owning the next device category was a rational response to a real, documented risk. The execution and the timing are what remain open to challenge and a negative outcome does not automatically prove the original reasoning was irrational.

There is also a colder reading available. Reality Labs normalized a culture of enormous, sustained, loss-making investment inside a public company. The shareholders who tolerated $80 billion of metaverse losses are the same shareholders now being asked to tolerate $130–145 billion of annual AI capital expenditure. The first bet may have purchased permission for the second.

The AI Bet, and What the Evidence Actually Shows

In June 2025, Meta invested approximately $14.3 billion for a 49% non-voting stake in Scale AI and recruited its chief executive to lead a new organization, Meta Superintelligence Labs. It is a striking deal structure: a non-voting minority stake, which is a common way to obtain talent, technology access, and commercial alignment while reducing the chance of a merger review. Reported coverage later that year suggested the working relationship with Scale was more complicated than the announcement implied a reminder that announced deals and delivered outcomes are different things.

What can be verified about the AI investment’s returns so far? Less than the enthusiasm suggests, and more than the skeptics allow.

On the supportive side: second-quarter 2026 revenue grew 28% year over year, an acceleration from the 22% full-year 2025 rate, at a scale where acceleration is genuinely difficult. Average price per ad grew 12%. Meta’s management has consistently attributed improvements in ad ranking, targeting, and creative tools to AI systems. That attribution is a company claim, but the pricing data is at least consistent with it.

That gap between visible growth and measurable return is not unique to Meta. Our analysis of AI video pricing and unit economics shows why fast adoption and ambitious revenue claims can coexist with opaque, compute-heavy economics.

On the skeptical side: none of Meta’s disclosures isolate AI-attributable revenue. There is no reported line item that lets an outside analyst calculate a return on the AI capital expenditure. Widely circulated figures for Meta AI assistant users vary considerably across sources and methodologies and should be treated as estimates rather than verified facts. Depreciation on nearly $226 billion of property and equipment will weigh on margins for years regardless of whether the AI products succeed.

The honest analytical position is this: the spending is verified, the revenue acceleration is verified, and the causal link between them is inferred rather than proven.

The Legal Bill Nobody Priced In

While the AI story dominated headlines, a second financial story was building in courtrooms and it may prove more consequential to near-term earnings.

Meta won the big one. In November 2025, U.S. District Judge James Boasberg ruled that the FTC had not demonstrated that Meta holds a monopoly in social media, ending the existential threat of a forced divestiture of Instagram and WhatsApp. Had it gone the other way, the entire structure described in this article would have been dismantled.

But the youth-safety litigation went differently. In March 2026, a New Mexico jury found Meta violated state consumer protection law, with reported penalties of $375 million, and a judge subsequently ordered payments reported at $567 million along with mandated safety measures. In July 2026, Reuters reported that Meta stated in a court filing that four states were seeking $1.4 trillion in penalties ahead of an August 2026 youth-safety trial. That figure is a plaintiff demand, not a judgment, and demands of that magnitude are frequently reduced dramatically or rejected entirely but the company’s own guidance language is notably direct: it warned of youth-related trials that “may ultimately result in a material loss.”

The $2.40 billion legal charge in the second quarter of 2026 was large enough that Meta explicitly raised the lower end of its full-year expense guidance to accommodate it.

In Europe, the picture is one of managed friction rather than crisis. The European Commission fined Meta €200 million in April 2025 over its pay-or-consent advertising model under the Digital Markets Act. By December 2025, Reuters reported the Commission had signaled acceptance of Meta’s revised Less Personalized Ads offering. Meta confirmed in its full-year 2025 release that it had “aligned with the European Commission on further changes.”

Regulation, in short, has become a recurring operating expense with unpredictable spikes not a one-time event.

Who Actually Controls the Money

None of this can be understood without the governance structure, which is the quietest and most important fact in the whole story.

Meta operates a dual-class share structure. Class B shares carry ten votes each. Reporting drawn from company proxy statements has indicated that Mark Zuckerberg holds the overwhelming majority of Class B shares, translating an economic stake in the low-to-mid teens percent into voting control around 60%. Exact percentages shift with share sales and issuance, so any specific figure should be checked against the most recent proxy statement rather than treated as fixed.

The economic consequence is simple and enormous. A public company spending $130–145 billion a year on infrastructure, absorbing $19 billion of annual segment losses, and cutting 8,000 jobs to redirect capital would normally face activist pressure, board fights, and possibly a change in management. Meta faces none of that in any meaningful sense, because ordinary shareholders cannot outvote the founder.

Whether you think that is good or bad depends largely on whether the AI bet pays off. Concentrated control is what allowed the mobile pivot and the post-ATT rebuild. It is also what allowed $80 billion of Reality Labs losses to continue year after year. It is the same feature producing both outcomes.

What the Public Story Doesn’t Explain

The public narrative about Meta is a narrative about products and controversies. The financial narrative is about something less visible: a company with an extraordinarily profitable core business deliberately spending that profitability down to buy optionality on a technology whose returns it cannot yet measure.

As of August 2026, market data placed Meta’s market capitalization in the neighborhood of $1.4 trillion. That number is a market opinion about the future, not a measure of what the company earned. It is worth remembering that the same market valued Meta at a fraction of that in late 2022, when the fundamentals were far better than the sentiment.

Three things remain genuinely uncertain, and no honest analysis should pretend otherwise. First, whether AI capital expenditure produces a return above its cost of capital the disclosures do not currently permit that calculation. Second, whether the depreciation load from hundreds of billions in infrastructure permanently resets Meta’s margin profile lower. Third, what the youth-safety litigation ultimately costs.

What the Financial Journey Teaches

Strip this down to transferable lessons and a few stand out.

Ownership beats revenue. Instagram cost roughly $715 million and became foundational. The lesson is not that acquisitions are good it is that buying an asset before the market prices its potential is where the real money is made, and that requires conviction when the evidence is thin.

Distribution you don’t control is a liability, even when it’s free. Apple’s tracking change cost Meta an estimated $10 billion in a single year without Meta having done anything wrong. Any business dependent on a platform it does not own carries this risk on its balance sheet whether or not it appears there.

Profit and cash are different animals. Meta reported $15.85 billion of net income in the second quarter of 2026 and generated $784 million of free cash flow. Both numbers are correct. They describe entirely different realities, and anyone reading only one of them misunderstands the company.

Capital-intensive strategies change what a company is. A business that once returned tens of billions to shareholders now issues bonds and forms financing joint ventures. Fixed assets create depreciation, depreciation compresses margins, and debt creates obligations that must be paid regardless of whether the strategy works.

And finally: control determines whose patience matters. Meta can sustain losses that would end most management careers, because of a share structure decided long before any of this was contemplated. Governance is not a footnote in a financial story. Sometimes it is the whole plot.

The most interesting thing about Meta’s money right now is not how much of it there is. It is how much of it has been deliberately taken off the table and pushed into the middle. The company has made its bet. The receipts on whether it pays off are still, genuinely, not in.

Financial Disclaimer: This article is provided for informational and educational purposes only. It is based on publicly available information and the cited sources, and reflects the public record as of the dates indicated. It does not constitute financial, investment, legal, tax, or accounting advice, and it should not be relied upon as a recommendation to buy, sell, hold, or make any financial decision. Figures described as estimates, reported information, or inferences are labeled as such and have not been independently audited. Readers should consult a qualified professional before making financial decisions.

Sources

Meta Platforms, Inc., “Meta Reports Fourth Quarter and Full Year 2025 Results,” January 28, 2026. https://investor.atmeta.com/investor-news/press-release-details/2026/Meta-Reports-Fourth-Quarter-and-Full-Year-2025-Results/default.aspx

Meta Platforms, Inc., “Meta Reports First Quarter 2026 Results,” April 29, 2026. https://investor.atmeta.com/investor-news/press-release-details/2026/Meta-Reports-First-Quarter-2026-Results/default.aspx

Meta Platforms, Inc., “Meta Reports Second Quarter 2026 Results,” July 29, 2026. https://investor.atmeta.com/investor-news/press-release-details/2026/Meta-Reports-Second-Quarter-2026-Results/default.aspx

Meta Platforms, Inc., “Facebook to Acquire WhatsApp,” February 19, 2014. https://investor.atmeta.com/investor-news/press-release-details/2014/Facebook-to-Acquire-WhatsApp/default.aspx

Facebook, Inc., Form 8-K, “Closing of WhatsApp Acquisition,” U.S. Securities and Exchange Commission, October 2014. https://www.sec.gov/Archives/edgar/data/1326801/000132680114000037/fb_8-kxclosingxofxwhatsapp.htm

Meta Platforms, Inc., “Meta Announces Joint Venture With Funds Managed by Blue Owl Capital to Develop Hyperion Data Center,” October 21, 2025. https://about.fb.com/news/2025/10/meta-blue-owl-capital-develop-hyperion-data-center/

Reuters, “Meta cash flow craters as Zuckerberg doubles down on AI,” July 29, 2026. https://www.reuters.com/business/meta-narrows-annual-capex-forecast-ai-buildout-grows-2026-07-29/

Reuters, “Meta to raise $30 billion in its biggest bond sale,” October 30, 2025. https://www.reuters.com/business/meta-seeks-least-25-billion-bond-offering-bloomberg-reports-2025-10-30/

Reuters, “Meta says US states are seeking $1.4 trillion in penalties in August youth safety trial,” July 7, 2026. https://www.reuters.com/business/meta-says-us-states-are-seeking-14-trillion-penalties-august-youth-safety-trial-2026-07-07/

Reuters, “Meta pledge to use less personal data for ads gets EU nod,” December 8, 2025. https://www.reuters.com/business/meta-offer-choices-personal-facebook-instagram-ads-eu-says-2025-12-08/

Bloomberg, “Meta Sells $30 Billion of Bonds as AI Frenzy Fuels Record Demand,” October 30, 2025. https://www.bloomberg.com/news/articles/2025-10-30/meta-platforms-offers-six-part-bond-amid-ai-spending-rush

CNBC, “Meta’s Reality Labs posts $6.02 billion loss in fourth quarter,” January 28, 2026. https://www.cnbc.com/2026/01/28/metas-reality-labs-posts-6point02-billion-loss-in-fourth-quarter.html

CNBC, “Meta’s Reality Labs lost over $4 billion in first quarter,” April 29, 2026. https://www.cnbc.com/2026/04/29/metas-reality-labs-lost-over-4-billion-in-first-quarter.html

CNBC, “Meta Louisiana data center investment reaches $50 billion amid AI push,” July 13, 2026. https://www.cnbc.com/2026/07/13/meta-louisiana-data-center-investment-reaches-50-billion-amid-ai-push.html

CNBC, “Facebook says Apple iOS privacy change will cost $10 billion this year,” February 2, 2022. https://www.cnbc.com/2022/02/02/facebook-says-apple-ios-privacy-change-will-cost-10-billion-this-year.html

CNBC, “Meta faces state AG trial over child safety claims,” August 17, 2026. https://www.cnbc.com/2026/08/17/meta-attorneys-general-california-federal-trial-astronomical-consequences.html

NPR, “Judge sides with Meta in antitrust trial, will not spin off Instagram and WhatsApp,” November 18, 2025. https://www.npr.org/2025/11/18/nx-s1-5495626/meta-ftc-instagram-whatsapp-antitrust-ruling

BBC News, “Meta does not have social media monopoly, judge rules,” November 2025. https://www.bbc.com/news/articles/crklgrpdke8o

Associated Press, “Meta invests $14.3B in AI firm Scale and recruits its CEO,” June 2025. https://apnews.com/article/meta-ai-superintelligence-agi-scale-alexandr-wang-4b55aabf7ea018e38ffdccb66e37cf26

The New York Times, “Meta Lays Off 8,000 Employees, as A.I. Casualties Mount,” May 19, 2026. https://www.nytimes.com/2026/05/19/technology/meta-layoffs-ai.html

European Commission, “Commission finds Apple and Meta in breach of the Digital Markets Act,” April 2025. https://digital-strategy.ec.europa.eu/en/news/commission-finds-apple-and-meta-breach-digital-markets-act

The Next Web, “Facebook’s final Instagram purchase price: $715 million,” October 2012. https://thenextweb.com/news/final-cost-to-facebook-to-purchase-instagram-715-million

Auganix, “Reality Labs Burns $19 Billion in 2025,” 2026. https://www.auganix.org/xr-news-meta-reality-labs-2025-financial-report/

Macrotrends, “Meta Platforms Market Cap 2012–2026,” accessed August 2026. https://www.macrotrends.net/stocks/charts/META/meta-platforms/market-cap

TechCrunch, “Cracks are forming in Meta’s partnership with Scale AI,” August 29, 2025. https://techcrunch.com/2025/08/29/cracks-are-forming-in-metas-partnership-with-scale-ai/