How Meta's Class C Deal Lets Zuckerberg Sell Without Losing Control

Meta's board has approved a new class of non-voting stock that the company says it can use for financing, acquisitions, and employee awards. It also agreed to give every existing Class A and Class B share one Class C share. The distribution does not immediately change anyone's voting percentage. Future Class C issuance can raise money without adding votes to the denominator.¹

The unusual part sits in a separate contract with Mark Zuckerberg. Meta calls it the Founder Alignment Agreement. It allows him to sell some of his new economic interest while keeping the high-vote shares that give him control, then releases more of the new stock as he continues to serve. The company's information statement says he held about 60.6% of votes and 13.4% of economic interests on October 1, 2026. He was not then on the verge of losing majority voting control.¹

The distinction matters to any founder board designing an equity raise. Creating a non-voting financing currency, granting its holders liquidity, and deciding how much stock a controller can sell without losing votes are three separate decisions. Meta negotiated them together. Its June proposal would have made the founder's Class B votes expire on specified events. The October agreement leaves those votes in place and puts conditions on the sale of new Class C shares instead.¹²

The proposed California wealth tax was a major reason the parties moved quickly. Zuckerberg told the committee he might establish residence elsewhere before the November 3 election if he could not pay a potential tax without selling voting shares. The eventual dividend is unconditional on that tax becoming law. A proposed change of address acquired its own share class.¹

Meta's stated business case is substantial. The special committee said the company needed an equity security it could issue at scale for AI investment, acquisitions, and compensation without reducing existing holders' voting percentages. A widely traded Class C would also help avoid a deep discount when Meta sells more of it. The committee said a small initial distribution might leave the new class too thinly traded and force the company to issue more shares to raise the same amount.¹ Reuters reported the plan alongside Meta's forecast of $135 billion to $145 billion in 2026 capital spending.³

There is a founder interest in the same structure. Zuckerberg's Class B stock carries ten votes per share; Class A carries one. Selling Class B ordinarily converts it to Class A and costs voting power. Selling a Class C share costs no vote because the share had none. Meta's information statement says the initial dividend could let Zuckerberg sell up to 20% of his new Class C shares, representing about 10% of his pre-dividend economic stake, without weakening his majority vote.¹

The committee treated that divergence as a real problem. It also judged Zuckerberg's continued leadership valuable. Its first governance proposal, delivered in June, would have converted all Class B stock to Class A within a year after his death, disability, or departure, or after his equity holdings fell below a specified half of their starting level. Founder counsel said the proposal was likely unacceptable. In August, Zuckerberg said he would not support an equity financing that diluted his control or agree to those sunset terms.¹

The tax was one pressure point, not the whole transaction. In early August, the committee floated a Class C dividend only if California Proposition 40 became law and a tax payment came due. By late August it offered an unconditional dividend of roughly one Class C share for nine existing shares. Its advisers estimated that plan could create a roughly $150 billion Class C market and about $20 billion of immediate founder liquidity. Zuckerberg's counsel argued the ratio was too low for a liquid market and did not address his broader liquidity concerns.¹

Both positions have force. A thinly traded non-voting class could sell at a discount, harming a company that intends to use it as currency. The committee also had reason to resist a large distribution that let its controller monetize a substantial interest without surrendering control. The record describes a financing negotiation in which the person whose votes were to be protected could refuse the financing terms. That bargaining position helps explain the final contract. Whether the board struck a fair bargain is a separate question.

The final one-for-one dividend gives each existing holder the same number of new Class C shares. Only Zuckerberg and his defined holder group face the agreement's stapling restriction. At the outset, 80% of their initial Class C allotment is stapled. They generally can sell a stapled Class C share only if they also sell a Class A or Class B share, convert a Class B share to Class A, or apply a credit earned from a prior sale or conversion. The agreement's ratio can adjust if there are too few remaining A and B shares to pair with the stapled C shares.²

That pairing gives a sale a potential voting cost. A sale or conversion of Class B removes nine of its ten votes; selling Class A removes its one vote. The details matter because the agreement permits pairing with either class and allows credits for earlier transactions. It does not promise that every Class C sale costs exactly ten votes, nor that a founder must sell two securities in one trade. A noncompliant sale can automatically convert the required Class B shares to Class A, subject to the contract's cure procedure.²

The remaining 20% of the initial Class C allotment is unstapled from day one. No linked voting-share sale is required for that portion. The agreement then releases another 2.5% of the original Class C allotment at the end of each qualifying quarter during which Zuckerberg serves full-time on Meta's executive team or as executive chair. Thirty-two qualifying quarters release the rest. They need not be consecutive; a quarter without qualifying service pauses the clock rather than confiscating the stock.²

The committee won a real limit on immediate liquidity. It did not win the Class B sunset it first proposed. That is the deal's governing asymmetry: the new stock's sale restriction runs down with service, while the old stock's votes have no comparable eight-year expiry in this agreement. A service condition can be a sensible retention device. It is a different protection from a promise that high-vote control will eventually end.

Here is the release schedule expressed as a fraction of Zuckerberg's initial Class C dividend. It assumes each quarter qualifies, the agreement is not amended, and no later Class C dividend changes the share count. It describes eligibility to sell without a paired A or B sale, not actual sales or a forecast of control.²

The arithmetic is 20% free initially plus 2.5 percentage points for each qualifying quarter, capped at 100%. The last column divides by two because the one-for-one initial dividend creates one new economic share for each existing one. It assumes equal economic value per A or B share and per C share after the dividend. Real market prices may differ: Meta itself warns that voting Class A could trade at a premium to non-voting Class C.¹

That table makes the negotiated concession easier to evaluate than the word "alignment." A founder who serves for eight qualifying years could become able to sell all of the initial Class C allotment without a linked sale of A or B. Meta states that this would permit sales of up to one-half of his economic stake from the initial dividend while retaining voting control, subject to the agreement and other applicable restrictions. It estimates he would still hold about $124 billion of economic interest at its October 1 reference price if he sold that entire allotment. Neither number predicts the future share price or says he intends to sell.¹

For a board, the relevant question is therefore not merely "Is 80% stapled?" Ask how quickly the stapled amount becomes free, what service counts, whether missed quarters catch up, and what happens after the final quarter. A large initial restriction can be temporary by design. That may be the right bargain. It should be valued as a temporary restriction.

The agreement has another release path. If Meta's Class C share price reaches successive doublings against a September 10, 2026 Class A reference price, and the specified annualized shareholder return reaches at least 10%, the board is to issue an additional one-for-one Class C dividend to all share classes, subject to legally available funds. The tests can be met on different dates. Both must be satisfied for a dividend. No additional dividend occurs after Zuckerberg's death.¹²

All shareholders receive that new economic stock on the same terms. If the agreement's continued-leadership condition is unmet when a dividend occurs, a formula keeps part of the founder's new portion stapled; satisfying it leaves that portion free. The committee described these contingent dividends and the service release schedule as part of the bargain in place of a performance-based compensation program; Zuckerberg withdrew his request for that program. Calling the dividend "compensation" would erase its distribution to every shareholder. Calling it a mere stock split would erase the separate founder sale conditions.¹²

The performance hurdle is a meaningful rival to the criticism that the founder simply received unlimited free liquidity. Any additional dividend requires unusually strong stock performance; the service condition governs how much of the founder's portion is free of stapling. But a hurdle for the next dividend does not restore a sunset for the voting stock. A founder board should model two paths separately: the steady release of the first dividend and the contingent creation of more non-voting shares. Calling both paths "alignment" conceals which term changes with time and which changes with performance.

The special committee could reject the recapitalization and spent months negotiating. Its members and advisers documented objections, changed terms, and obtained restrictions that apply only to the founder. The company says Class C helps it finance growth without an expensive compensation award, and it says directors do not simply rubber-stamp Zuckerberg. Those are material parts of the board's case.¹

Yet the charter change and equity-plan amendment did not go to a meeting of unaffiliated holders. On October 20, consenting holders of about 56.1% of Meta's voting power approved them by written consent. The November 2 information statement tells other shareholders they need take no action and that Meta is not asking for their vote. That is the disclosed approval route, not proof that the outside owners agreed with the exchange of economic flexibility for continued control.¹

The filing also reports Delaware litigation seeking to stop the recapitalization and alleging fiduciary breaches. Meta denies the claims and says it will defend the transaction. The filing does not establish the merits of either side. An editor or investor should keep the allegation separate from the contract's observable terms.¹

This is where the question in our dual-class sunset analysis becomes concrete. Who approves the new control benefit if the controller can pass the vote alone? Meta's case adds a further question: when a special committee negotiates a restriction in lieu of a sunset, who is allowed to decide whether eight years of release is enough? The legal vote can be valid while leaving that economic judgment to the board and controller.

If your company is considering a new non-voting class, do not start with the attractive phrase "everyone gets the same dividend." Put four dated columns in the board paper instead:

Then run the columns across the financing plan and a plausible founder liquidity plan. Use the actual charter and agreement alongside the cap table. A control wedge shows the present gap between votes and economics. The release schedule shows how that gap could change even if the founder receives no new voting share.

Meta's arrangement may fund growth on terms its board considers worthwhile. It also demonstrates a specific cost of doing so: future capital and founder liquidity can expand without corresponding voting dilution, while the founder's initial sale limits expire through service. Before you describe that as alignment, write down the date and condition on which each limit stops aligning anyone.