Meta's quarter matters less for its income statement than for the change in its role in its own stock. For years the company's cash flow made it a reliable buyer of its own shares, absorbing employee dilution and then some. In 3Q'26 free cash flow turned negative, net cash disappeared, and management launched a non-voting Class C share that the 10-Q says will be used "to conduct equity offerings from time to time in the future." When a company's treasury goes from absorbing supply to creating it, holders start pricing the supply schedule ahead of the earnings line. That repricing of supply, more than any doubt about the advertising machine, is what spooked investors.
A Core Business With Nothing to Apologize For
The operating results deserve to be stated plainly before the funding question takes over. Revenue was $64.6 billion, up 26% year over year. Ad revenue grew 26% to $63.1 billion, with impressions up 10% and price per ad up 14%, so both volume and pricing are contributing. The 4Q'26 revenue guide of $73-76 billion implies roughly 24% growth at the midpoint, which is a very small deceleration for a business of this size.
The GAAP operating income decline of 37% to $13.0 billion is entirely accounted for by the $9.9 billion settlement with 52 state attorneys general. That charge had been known since August. Excluding it, operating income grew 11% at a 35% margin.
The engagement metrics support the revenue figures. Daily active people reached 3.64 billion. Instagram time spent per user hit an all-time high. Nearly 45% of recommended Reels are less than a day old, which shows the recommendation system is surfacing fresh inventory rather than recycling old content. Partnership ads passed a $15 billion run rate growing about 97%. "Other revenue" rose 59% to $1.1 billion on WhatsApp paid messaging and subscriptions.
There is one regional crack. For the second straight quarter, US & Canada was the only region where ad growth accelerated, while every other region decelerated. North America and Rest of World are growing at the same rate even though North America is more than three times larger. Meta did not attribute this to anything. Higher oil prices and softer macro conditions outside North America are a reasonable guess, but they are unconfirmed. Europe also has its own issues with personalized ad restrictions.
The Margin Problem Under the Healthy Top Line
The weaker part of the core is incremental profitability. Family of Apps added $13.4 billion of revenue year over year but only $2.0 billion of adjusted operating income. That is a 15% incremental margin, the same as in Q2. The adjusted FOA margin of 42.1% still looks fine, but the business is now converting roughly one dollar in seven of new revenue into new profit.
The expense lines show why:
- Total expenses excluding the charge grew 36% against 26% revenue growth.
- Cost of revenue grew 45%. It now includes third-party AI token costs, which management expects to fall as Meta's own models replace outside ones.
- Depreciation was $7.2 billion, up 44%.
- R&D rose 32% even though headcount fell 15%.
- Stock-based compensation grew 32% to $7.3 billion.
- G&A excluding the legal charge was about $5.0 billion, up 41%.
Full-year expenses are now guided to $175-179 billion, raised by roughly the size of the legal charge. At the guide midpoints, Q4 operating income lands near $24.6 billion. That is flat against 4Q'25, and the ends of the range span a 15% decline to a 14% increase. In other words, a business growing revenue 24% is guiding to roughly zero operating income growth in its seasonally strongest quarter.
There was real progress at Reality Labs, which lost $4.1 billion on $443 million of revenue. Management gave a loss path for the first time:
- 2026 losses of $18-19 billion, versus $19.2 billion in 2025.
- 2027 losses of "around $15 billion."
- "Trending towards profitability from there."
New VR devices will come only with a major breakthrough, and Muse will serve as the software layer on glasses priced near cost. For investors who already treat Reality Labs losses as capitalized investment, this matters. But a $15 billion annual loss in 2027 does not ease the cash question.
How the Cash Flow Statement Turned Meta Into a Seller
This is the core of the argument. Capex, including finance lease principal, was $34.9 billion against operating cash flow of $32.8 billion, giving free cash flow of -$2.1 billion.
The capex guide was narrowed to $135-145 billion, which implies $49-59 billion in the fourth quarter alone. That is a step up of $14-24 billion in a single quarter, so Q4 free cash flow will be clearly more negative. Cash and securities of $82.4 billion now sit below $83.7 billion of debt, which means the net cash position is gone.
The mechanism that links this to the stock price is ownership flow.
Stock compensation becomes uncovered dilution. Stock-based compensation at $7.3 billion a quarter is close to $29 billion of equity delivered to employees each year. When free cash flow comfortably exceeded that, repurchases could offset the share issuance and the share count drifted down. With free cash flow negative, any buyback large enough to offset that issuance has to be funded with debt. That is the funding channel Meta's own bankers are flagging.
The debt market is not cooperating. According to the filing, the special committee's bankers pointed over the summer to "less favorable debt market conditions for hyperscalers," tighter credit, and weaker demand for the July bond issued by Meta's Texas data center joint venture. If the debt route is getting more expensive at the margin, the remaining options are slower capex or equity.
Class C is built to be the instrument. Each existing holder receives one Class C share per share held, which works like a two-for-one split, with trading starting in late January 2027. The non-voting class can be used for capital raising, employee compensation and acquisitions without diluting voting control. Once that currency exists, issuing stock to fund the AI buildout no longer threatens control, and control was the main thing that previously made large equity issuance unlikely.
Live data on dealer or systematic positioning in the stock was not available for this article, so the crowding argument here rests on ownership structure rather than measured flows. On that basis, the risk to holders is not a single offering. It is that the market now has to assume a continuous source of new share supply, and that supply has a calendar: the January trading date for Class C and the 2027 capex guide due around the same time.
Two Motives for One Share Class
The share class has two possible explanations, and the market is reacting to one of them.
The first is Zuckerberg's taxes. Proposition 40, California's proposed wealth tax, is on the November 3 ballot. The filing says Zuckerberg told the lead independent director in July that selling enough Class B stock to pay a potential tax bill could reduce his voting stake to about 53%, and that he was considering moving his residence out of California. Class C solves this. He can sell 20% of his Class C shares immediately, and the rest only alongside an equal number of voting shares. In exchange, he commits for eight years to live wherever Meta is headquartered. This also creates supply, because the founder now has a tradable currency. But it is bounded and tied to a personal tax event, and markets usually absorb that kind of supply without changing the valuation framework.
The second explanation is the one that unsettled investors. Management presented three long-term forecast cases to the board in May. Each showed a need to raise hundreds of billions of dollars over several years, starting as early as fall 2026.
On the follow-up call, Li walked this back. He described the figure as "a theoretical maximum that was basically computed to determine what level of equity financing would trigger a loss of founder control," and said it was "not actually a description of our present equity financing needs or plans." That explanation is plausible. A committee deciding whether to approve a new share class needs to know where control would be lost, and the board's job was to protect control, not to approve a fundraising program.
The problem is that the 10-Q signed the same day states that Meta expects to use Class C shares for equity offerings in the future. Li also said, in answer to a different question, that a strong balance sheet will matter "as the market shakes out in terms of how this all gets funded." The source analyst puts the probability of an equity raise next year at 80%. That number is the analyst's judgment, not a disclosed plan. Still, it is a reasonable reading of documents that tell two different stories, and the legally binding document is the one that mentions offerings.
Where the Supply Thesis Could Break
The strongest counterargument is that capex peaks before any equity is needed. Three developments would support it:
- Lower cost to serve. Meta has started moving content understanding onto its own Muse models, and offline tests show per-unit processing costs falling by more than 80%. Token costs paid to third parties should also decline as internal models take over.
- Narrowing Reality Labs losses. Losses falling from $19.2 billion to around $15 billion return a few billion dollars of cash a year.
- A Q4 capex spike that does not repeat. If the $49-59 billion quarter reflects timing rather than a new run rate, operating cash flow growing alongside 24% revenue growth could bring free cash flow back above zero during 2027 without new stock.
There is a second, less obvious counterargument. A single, well-sized offering could be good for the stock. The selling pressure comes from uncertainty about how much stock will be issued and when. A defined raise that funds the buildout would remove that uncertainty in one step, and it would protect the credit profile that the hyperscaler bond market is currently pressuring.
Against this, the evidence on the operating line points the other way. Depreciation is growing 44% while incremental FOA margins sit at 15%. Each additional dollar of capex is lowering reported margins before it pays back. A company with that profile does not usually shrink its capex budget on its own initiative.
What Settles It Between Now and January
The thesis can be tested on specific dates.
- November 3, Proposition 40. The outcome determines how much of the founder's Class C supply is actually needed.
- Q4 free cash flow. It has to come in as negative as the capex guide implies. If it does, the gap between the 10-Q language and Li's walk-back resolves toward the 10-Q.
- The 2027 capex guide, due around the time Class C starts trading in late January. Guidance that holds or rises from the implied Q4 run rate, while debt demand for hyperscaler paper stays weak, leaves equity as the remaining funding source.
- The opposite case. If 2027 capex flattens and repurchases continue without new borrowing, the supply concern fades and the stock goes back to being valued on a business that is still growing revenue in the mid-20s.
If capex steps up again and repurchases quietly stop, the market will price Meta as a company that has to issue stock to fund its AI buildout, regardless of how strong the advertising business is.
