Meta Sells No-Vote Stock After Its First Cash Shortfall

Meta spent $34.91 billion on servers, data centers, and network gear in the third quarter, more than the $32.79 billion of cash the business produced. Free cash flow went negative by $2.12 billion, the first quarterly shortfall since the 2012 IPO.

Revenue still hit a record $64.60 billion, up 26 percent and more than $1 billion above the Wall Street forecast. Shares still fell 7 percent on Thursday after the board created non-voting Class C stock so Meta can raise money without touching Mark Zuckerberg’s control.

Record Ad Sales Still Left a Cash Hole

The advertising machine did what it was built to do. Family daily active people averaged 3.64 billion in September, up 3 percent. Ad impressions rose 10 percent and the average price per ad rose 14 percent. Family of Apps advertising revenue was $63.1 billion, also up 26 percent.

GAAP earnings looked worse than the cash line because of a legal bill, not because ads failed. In August, Meta reached an agreement with 52 state attorneys general on teen mental health and social media cases, with payments of up to about $18.59 billion. The third quarter took a $9.89 billion charge for that deal. Net income was $9.86 billion, or $3.83 a share. Strip out the charge and operating income would have been $22.88 billion, with diluted earnings of $7.40 a share and a 35 percent operating margin instead of 20 percent.

That charge hits the income statement. It is not why cash went negative. Operating cash flow was still $32.79 billion, up 9 percent. Capital expenditures, including principal payments on finance leases, were $34.91 billion, about 80 percent higher than a year earlier. A year ago, free cash flow in the same quarter was $10.6 billion. The third-quarter results and cash-flow tables put the gap in one place.

THE QUARTER’S CASH MATH

Full-year capital spending, including those lease payments, is now guided at $135 billion to $145 billion, narrowed from $130 billion to $145 billion by lifting the floor $5 billion. Nine-month purchases of property and equipment were $83.06 billion, and finance-lease principal took another $2.77 billion, or about $85.83 billion already spent. Hitting even the low end of the year means a fourth quarter of about $49 billion to $59 billion, heavier than the quarter that just went negative.

Zuckerberg, the founder, chairman, and chief executive, said in the release that investments are paying off and the core business is strong. Susan Li, the chief financial officer, guided fourth-quarter revenue to $73 billion to $76 billion, with foreign exchange a 1 percent headwind, and full-year expenses to $175 billion to $179 billion including the legal charge. Reality Labs losses are pegged at $18 billion to $19 billion this year and around $15 billion next year as money is shifted toward AI. Headcount was 66,775 on September 30, down 15 percent from a year earlier, so the cash drain is hardware and power, not a hiring binge.

Class C Shares Carry the Check, Not the Vote

The first distribution of the new stock raises no cash. Holders of Class A and Class B shares are due one Class C share for each share they own once an amended Delaware charter takes effect on January 19, 2027, a move Meta says should trade like a two-for-one split. Voting power does not change on that day. The point is what Meta can sell after that.

The board approved the plan on October 19 on the unanimous recommendation of a special committee of independent directors. The next day, seven Zuckerberg- and Chan-linked entities that owned 315,826,951 Class B shares, about 56.1 percent of the vote, approved it by written consent. No other shareholder was asked. The non-voting Class C recapitalization terms say the new stock is for capital raising, employee pay, acquisitions, and general corporate purposes, and that Meta expects to offer Class C shares from time to time.

HOW THE NEW STOCK IS BUILT

• The dividend: One Class C share for each outstanding Class A and Class B share after the charter takes effect on January 19, 2027.
• The vote: Class C carries no votes, so later sales do not shrink the existing Class A and Class B voting split.
• The founder: As of October 1, Zuckerberg held about 60.6 percent of the vote on about 13.4 percent of the economic interest, through dual-class stock in which Class B carries 10 votes.
• The capacity: Moelis, adviser to the special committee, told directors Meta could raise about $700 billion of equity over several years under the current setup before his vote fell below 50 percent.
• The forecast: Management told those advisers that long-term cases showed a need to raise hundreds of billions of dollars over several years, beginning as early as fall 2026.

Li later called that hundreds-of-billions figure a theoretical maximum used to test when founder control would break, not a description of present equity plans, and said Meta had nothing to announce on that front. The filing still exists for a reason. Class C is a pipe that can take in cash from people who get an economic claim and no say, which is the whole design.

The special committee also wrote that Meta’s Class A stock is up more than 1,700 percent since the IPO, against about 495 percent for the S&P 500, and that keeping Zuckerberg in charge is part of why the new class exists. Public holders who already own the one-vote stock did not get a ballot on that conclusion.

Default Insurance Hits a Record High

Equity sold off. Credit markets sent a quieter bill. Five-year credit default swaps, contracts that pay if Meta misses a debt payment, widened 7.6 basis points on Thursday to a record 114.79 basis points, according to Markit. Protection on other big AI spenders moved with it, including Oracle, Microsoft, and Broadcom.

Steve Sosnick, chief market analyst at Interactive Brokers, said the market has been questioning whether AI spending is producing the returns it first expected. Shay Boloor, chief market strategist at Futurum Equities, said the jump in swaps could have more to do with banks hedging Meta’s AI debt than with a live default scare. Both can be true at once: nobody is pricing a Facebook bankruptcy, and somebody is paying up to lay off a much larger debt book.

Morgan Stanley Research has put AI-linked debt issuance at an estimated $466 billion so far in 2026, against $101 billion in the same stretch of 2025, and said weekly CDS trading on a cluster of those names, Meta among them, averaged more than $9 billion in September, up from less than $2 billion in July. The people writing that insurance are now a constituency Meta has to live with, even if they never appear on an earnings slide.

Piper Sandler trimmed its price target to $850 on higher costs and kept an overweight rating. Analysts can stay constructive on Muse and still watch the swap screen, because those are different books.

What Muse Has to Earn

The product that is supposed to justify the servers is Muse, Meta’s personal agent, launched in early September in the United States and Canada. Zuckerberg told the earnings call it was one of the company’s most successful launches, with more than 10 million people setting up agents in about six weeks, and that Meta has been number 1 in the App Store almost every day.

The most common use case for Muse is shopping. With more than 25% of people who are using Muse engaging in this with an average purchase size of over $200.

Mark Zuckerberg, founder and CEO, Q3 2026 earnings call

He said about 25 percent of people who have set up Muse use it on a given day, more than twice the early run rate of Threads, and that Plaid had not seen connector activity like this on any other service. In the first week Meta had to ask Stripe to raise credit limits. Beyond shopping, the common hooks are productivity, email, and calendar. Li said Meta is not yet monetizing Muse outside subscriptions. Zuckerberg wants business fees on agent-led transactions, optional subscriptions for heavier users, and, over time, more revenue per American user than the Family of Apps, which he put at several hundred dollars a year.

He also said Muse is compute-intensive to run, which is why it has not launched worldwide, and that Meta expects to improve cost efficiency by more than 10 times in the next year so the product can be profitable once revenue is at scale. He wants capacity for more than a billion people over the next few years and talks about a large-scale Muse network in which agents transact with other agents. That is the wager the cash hole is funding.

Early users are already hitting a different limit. People testing the agent have watched it ask for account passwords even when they were logged in, then walked away rather than hand over keys. Shopping volume can be real and still collide with trust. Meta is building a confidential virtual machine it says even the company will not be able to read. Until that ships, the agent’s appetite for access is part of the product, and part of the cost of scaling it.

Buybacks Stopped and Bond Sales Took Over

Meta used to buy its own stock with leftover cash. In the first nine months of 2025 it spent $26.25 billion on Class A repurchases. In the first nine months of 2026 it spent nothing. Dividend and dividend-equivalent payments in the third quarter were $1.36 billion, so the capital-return program is now a dividend, not a shrink of the share count.

Debt filled the hole. In May, Meta issued $25.00 billion of fixed-rate senior notes. The 10-Q shows $24.91 billion of net proceeds from long-term debt in the nine-month period. Face value of outstanding notes is $84.00 billion. Long-term debt on the balance sheet is $80.92 billion, plus $2.75 billion classified as current, against $82.45 billion of cash, cash equivalents, and marketable securities. The estimated fair value of those notes was $73.31 billion on September 30, below par. Interest expense on the notes, net of amounts capitalized, was $882 million in the quarter, against $205 million a year earlier.

The notes now run from 2022, 2023, 2024, November 2025 ($30 billion), and May 2026. Meta is no longer a company that throws off more cash than it can spend. It is a company that can still borrow, and that is building a second, voteless equity currency so it does not have to borrow all of it.

A Side Deal Tied to California’s Wealth Tax

The same week, Meta signed a contract with its own controlling shareholder. The Founder Alignment Agreement with Zuckerberg, dated October 20, limits how he can sell the Class C shares he will receive.

THE RECAP CALENDAR

1. October 19, 2026: The board, on a special committee’s unanimous recommendation, approves the restated charter and the Class C dividend.
2. October 20, 2026: Consenting Class B holders with 56.1 percent of the vote approve the charter, and Zuckerberg signs the Founder Alignment Agreement.
3. October 21, 2026: Third-quarter results and the Class C plan are disclosed with the earnings release.
4. November 3, 2026: California voters decide Proposition 40, a proposed one-time 5 percent wealth tax.
5. January 19, 2027: The restated charter takes effect and the one-for-one Class C dividend can be paid.

Under that agreement, 20 percent of the Class C shares Zuckerberg’s holder group receives can be sold freely, which the company says is about 10 percent of his overall economic stake. The other 80 percent are stapled: selling one generally means also selling or converting a voting share, or using credits from earlier sales. A further 2.5 percent of his original Class C allotment unlocks for each qualifying quarter he remains a full-time executive or executive chair, up to 32 quarters. He also committed to remain a California resident, or to live where Meta keeps its headquarters, for eight years, with an out for material unforeseen circumstances.

The information statement ties the free slice to Proposition 40. Zuckerberg had told directors he was weighing a move out of California if the tax passed, because selling high-vote Class B shares to pay a multi-billion-dollar bill could cut his voting interest to about 53 percent. The new stock lets him raise cash without touching that vote. The internet’s shortest summary of the Reuters post on that deal was that Meta had struck a bargain with Zuckerberg. That is literally the counterparty: the company is contracting with the person who already holds the votes that made the contract possible.

Leases and Contracts Already Top $810 Billion

The quarter’s $2.12 billion shortfall is small next to what is already signed and not yet on the balance sheet as debt.

OFF-BALANCE-SHEET BILLS, SEPTEMBER 30

• Leases not started: About $355.73 billion of operating and finance leases, mostly data centers, energy kit, colocations, and network gear, due to begin from the rest of 2026 through 2036, with terms of more than one year up to 30 years.
• Firm contracts: $454.52 billion of contractual commitments for cloud capacity, servers, network gear, energy, and Reality Labs hardware, with about $45.85 billion due in 2026 and $121.41 billion due in 2027.
• The combined total: About $810 billion of leases plus contracts, before a separate contingent cloud obligation of up to $14.72 billion over five years.
• Escrowed cash: $10.80 billion of money-market funds locked against multi-year infrastructure purchases and due to be released between 2028 and 2030.

Those commitments are why Class C is not a split for its own sake. The ad business still prints tens of billions of dollars a quarter. It no longer covers the build, the leases already signed, and the agent that cannot yet run worldwide without a 10-times cut in compute cost. Outside capital, some of it voteless, is being asked to close that gap while the founder’s 60.6 percent remains the vote that counts.

Fourth-quarter revenue is guided at $73 billion to $76 billion, and the spending range still tops out at $145 billion for the year. The cash hole is not a one-quarter accident. It is the funding model now, and the new shareholders who fill it will not get a ballot.

Disclaimer: This article is news reporting and analysis of Meta Platforms’ third-quarter results, capital-spending plans, and Class C recapitalization, and is for information only. It is not investment advice, a recommendation to buy or sell META or any related security, or tax or legal advice on dual-class stock, credit-default swaps, or California’s Proposition 40. Readers should consult a licensed financial adviser, tax professional, or securities lawyer before making any investment or tax decision. Figures, ratings, and deal terms reflect company filings and market data as of the dates named in the piece and can change with later filings, votes, and trading.