Behind a mild-sounding headline sits a bigger story: a new non-voting stock that keeps Mark Zuckerberg firmly in charge, a capex floor that just went up, not down, and a market that has stopped flinching at the spending.
Read the headline quickly and you might picture Meta pumping the brakes. It is doing close to the opposite. When the company reported third-quarter results on October 21, it raised the bottom of its 2026 capital-spending range and unveiled a new share class built to keep funding an all-out AI push for years, all while handing founder Mark Zuckerberg an even firmer grip on control. “Tightens” is the right verb; it just does not mean what a casual reader assumes.
Q3 2026 at a Glance
| Metric | Q3 2026 (quarter ended Sept 30) |
| Revenue | $64.60B — up 26% year over year; beat the ~$63.3B estimate |
| Advertising revenue | $63.06B — up 26%, still the engine funding everything else |
| Diluted EPS | $3.83 — up from $1.05 a year earlier; beat the $3.57 consensus |
| EPS excluding charge | ~$7.40 — stripping out a one-time $9.89B legal charge |
| Legal charge | $9.89B — tied to a settlement with 52 state attorneys general |
| Free cash flow | –$2.12B — a deficit, versus +$10.63B a year earlier |
| Q4 revenue guidance | $73B–$76B — roughly in line with Wall Street at the midpoint |
| 2026 capex | $135B–$145B — floor raised from $130B; the range narrowed upward |
| Stock reaction | Up ~4–7% in extended trading |
The Headline, Decoded
Two things happened at once, and the shorthand flattens both. First, Meta did not trim its AI budget, it narrowed the 2026 capital-expenditure range to $135 billion to $145 billion, lifting the low end from $130 billion. “Tightening” here means removing the cheaper outcome, not the expensive one. The signal to investors is confidence, not restraint: management is telling the market that spending will land in the upper half of what it previously floated.
Second, the company is minting a new tool to pay for all of it. A fresh class of publicly listed, non-voting Class C stock gives Meta a currency for acquisitions, employee compensation and capital raising without diluting the voting power concentrated in Zuckerberg and other insiders. The two moves are the same story told twice: keep spending big on AI, and keep control while doing it.
What the New Class C Stock Actually Does
Meta’s board approved the new non-voting Class C shares effective January 19, 2027. Mechanically, every holder of Class A and Class B stock will receive one Class C share as a one-time stock dividend, a one-for-one distribution that behaves like a two-for-one split on the Class A price, while leaving voting power untouched. A special committee of independent directors recommended it unanimously, and shareholders approved it by written consent.
The structure will look familiar to anyone who has watched Alphabet. By paying people and buying companies in a stock that carries no vote, Meta can spend its equity freely without ever loosening founder control. That is the quiet purpose behind the corporate-finance language about “flexibility.”
How the Three Share Classes Compare
| Class | Voting power | Who holds it / purpose |
| Class A | 1 vote per share | Public shareholders, the shares that trade under META |
| Class B | 10 votes per share | Insiders, including Mark Zuckerberg, the control block |
| Class C (new) | None — non-voting | Distributed Jan 19, 2027; a spending currency for pay, M&A and talent that never dilutes control |
The Spending Picture
The scale is hard to overstate. A 2026 capital budget approaching $145 billion puts Meta among the heaviest spenders in corporate history, and the bill is already visible in the cash statement: free cash flow swung to a $2.12 billion deficit for the quarter, from a $10.63 billion surplus a year earlier. That is the price of building AI infrastructure faster than the revenue it will eventually produce.
Management’s framing is that the losses are migrating toward AI and away from the metaverse. Reality Labs, long the money pit — is now expected to narrow its losses to roughly $15 billion next year as investment shifts primarily to AI. CFO Susan Li told investors the company’s AI businesses should add incremental revenue in 2027 and become “a meaningful source of revenue diversification over time.” Notably, Meta still did not put a number on 2027 capex, the one figure the market most wants.
Why Investors Shrugged Off the Cash Burn?
A year ago, a cash-flow deficit and a rising capex floor might have sent the stock down. This time shares rose in after-hours trading. The difference is a product the market can point to: Muse, Meta’s new consumer AI agent, which signed up more than 10 million users across the US and Canada in about six weeks, with roughly a quarter of them using it daily. Zuckerberg says Facebook and Instagram are seeing “all-time highs in engagement,” and he expects Muse to eventually reach billions though he conceded it costs more to run than past launches and will be monetized sooner than Meta’s usual slow-build playbook.
The read from the Street was that the narrative has flipped: where Meta was once punished for spending, it is now being rewarded for having a story to attach the spending to. That is a fragile kind of approval, it lasts exactly as long as the product momentum does.
The Legal Cloud Over the Numbers
The reason reported earnings look messy is a $9.89 billion charge tied to a settlement with 52 state attorneys general over litigation about teen mental health, the gap between the $3.83 headline EPS and the roughly $7.40 the business earned underneath it. It is not the only overhang: Meta is also contending with a separate court ruling ordering a multibillion-dollar penalty related to how it shared user data. For now investors are treating these as one-time noise rather than a structural drag, but the legal tab is becoming a recurring line in Meta’s story, not a footnote.
What to Watch Next
- The 2027 capex number. Meta pointedly withheld it. When it lands, it will reset the whole debate about how long the spending ramp runs.
- Muse retention and monetization. Ten million sign-ups is a headline; daily use and the first revenue are the proof. A sooner-than-usual monetization promise is a clock Meta has started on itself.
- Whether Reality Labs losses actually narrow to ~$15 billion, the test of the claim that money is truly shifting from the metaverse to AI.
- How the Class C shares get used once they exist in January. Heavy issuance for acquisitions would confirm this was always about buying scale without ceding control.
Discussion (0)
You need a free TechJox account to join the discussion and report scams you spot.