Zuckerberg sells eight years of residency and buys a share class Meta was forced to abandon in 2017 Meta Platforms, Inc. (NASDAQ: META) · chairman & chief executive
- What he did
- Signed an agreement dated 19 October 2026 with Meta’s board. Meta creates a new non-voting Class C, distributed one-for-one to every Class A and Class B holder; the new charter takes effect 19 January 2027 and trading is expected late that month. A further one-for-one Class C dividend goes to all holders each time the stock doubles from $580.60, provided annual returns are at least 10%. Zuckerberg may sell 20% of his new Class C immediately “for any reason”; the remaining 80% is stapled to his voting shares and released over eight years. He withdrew his pay request, keeps the $1 salary he has taken since 2013, and commits to keep his primary residence in California, or wherever Meta is headquartered, for eight years.
- The document
- Meta SEC filing of 21 October 2026 (agreement file dated meta-20261019), released the same day as Q3 results. Board approval 19 October; Chan Zuckerberg entities holding 56% of the vote signed off 20 October; the deal itself was inked 14 September.
- The backdrop, as the filing tells it
- California Proposition 40, on the ballot 3 November 2026, would levy a one-time 5% tax on the net worth of state residents worth over $1bn, with liability fixed by residency on 1 January 2026. On a net worth of about $250bn that is roughly $12.5bn. Most of it sits in a 13% Meta stake; selling Class B to pay would take his board voting interest from about 61% to about 53%. In July he raised the measure repeatedly with lead independent director Robert M. Kimmitt; on 3 August his attorneys wrote to a board special committee demanding a revised compensation agreement; on 9 September he told a board dinner he planned to list his $110m Palo Alto estate, and that he would need to begin establishing residency outside California by mid-October.
Consensus read
A retention deal, and a win for the tax campaign. The founder threatened to leave over a wealth tax; the board bought him enough liquidity to pay a $12.5bn bill without selling voting stock, and locked him into the state for eight years in exchange.
Gabriel Zucman: “This SEC filing is proof that the campaign for Proposition 40 is working. We are at a turning point.” The steelman is genuinely strong. The negotiation really did begin with his Prop 40 concerns and the filing says so in its own words. A founder paying $12.5bn out of a 13% stake really does have a control problem, and a special committee really does have a defensible interest in keeping its chief executive near Menlo Park.
Contrarian read
The tax is the occasion, not the cause. Meta has wanted a non-voting class since 2016 and was made to drop it in September 2017 under a Delaware fiduciary suit. The structure was never the obstacle — the record was.
What changed in 2026 is that an external, one-time, precisely quantifiable personal liability hands a special committee a business purpose — retain the founder — that the 2016 “fund philanthropy” rationale could not sustain in court. The eight-year residency covenant is the consideration manufactured to make that record hold.
What Meta actually bought is a permanent, control-neutral equity currency, for a company whose free cash flow has just gone negative against a $135–145bn capex year.
Evidence that discriminates
- The ratchet. A one-for-one Class C dividend repeats every time the stock doubles from $580.60, indefinitely, subject only to a 10% annual-return floor. A one-time tax cannot explain a perpetual, price-triggered issuance mechanism. This is the single fact the retention read cannot absorb: nothing about a bill you pay once requires a machine that fires again at $1,161, and again at $2,322.
- It is unconditional. Prop 40 is voted on 3 November 2026. The charter takes effect 19 January 2027 and the 20% sale right applies “for any reason.” Nothing in the reported terms unwinds if the measure fails. A tax bridge would be contingent on the tax existing.
- The company has the funding problem, not only the man. Q3 2026 free cash flow was negative $2.12bn — capex $34.91bn against operating cash flow of $32.79bn. Full-year capex guidance is $135–145bn, implying $49–59bn in Q4 alone. Long-term debt rose $22.2bn in nine months to $80.9bn. There have been no buybacks in 2026. CFO Susan Li said on the Q3 call she wanted “access to a diverse pool of funding sources.” A new non-voting class dilutes economics without diluting votes — which is exactly what a founder-controlled company needs to issue equity at scale.
- The opening ask. On 4 August he proposed 20 Class C shares per share. The committee called it excessive the next day. An ask twenty times the settled figure is not an ask calibrated to a $12.5bn liability; it is an opening bid on capital structure.
- The cheap alternative he did not take. Borrowing against stock solves a $12.5bn liability in days, with no corporate action, no committee, no charter amendment and no disclosure beyond a pledge line in the proxy. Larry Ellison did precisely that in the same week (item 02 below). Meta instead ran an eleven-week, counsel-driven process and rewrote the charter. When a party declines the fast, cheap, off-the-shelf instrument and buys the slow, expensive, permanent one, the permanent one is what they came for.
What does not discriminate, and why I am not using it: the eight-year covenant is widely read as proof the tax drove the deal. Against Prop 40 specifically it is close to worthless — liability is fixed by residency on 1 January 2026, already nine months past, so he could move on 4 November and still owe. Inference, labelled as such: its real value is against future wealth taxes in other jurisdictions, a worry the filing records him raising. That cuts both ways and separates nothing, so it stays out of the case.
Adjacent measurement check — the “30% of the base has already left” claim
The Jaros–Rauh NBER working paper (October 2026, Hoover Institution) puts Prop 40 collections at $28–36bn against proponents’ $100bn, and reports that departures have “already eliminated nearly 30% of the proposed tax base.” The base is fixed by residency on 1 January 2026. Departures after that date cannot shrink it. The paper’s own caveat concedes that announced moves do not establish liability. The denominator is not a choice variable, and a 30% erosion attributed to a decision that cannot affect it is the error, not the conclusion.
Boll, Saez and Zucman make the mirror objection from the other side: it is inconsistent to model the tax as one-time but the emigration as permanent, and on their arithmetic it would take about 25 years of lost income tax to match a $100bn windfall even if every billionaire left for good. Saez gave unpaid advice on the measure’s design, which is worth knowing. California’s own Legislative Analyst’s Office expects tens of billions over several years with annual income-tax losses under $1bn.
The same disease, different patient: the Tax Justice Network’s 19 October review of Bloomberg’s $160bn “UK billionaire exodus” found that 7 of the 14 listed individuals fail a basic condition — wrong window, or wealth attributed to a relative who never left — and that only $31.5bn (20%) of the headline was ever UK-based to begin with. When a flight statistic is quoted at you this season, ask what the denominator is and when the window opened.
What would falsify me
The Class C issuance, the doubling ratchet, or the 20% sale right terminating or unwinding if Prop 40 fails on 3 November. That is the cleanest test and it resolves in twelve days.
An aggregate monetisation cap at or near $12.5bn rather than materially above it — 20% of his new Class C is reported as about 10% of his financial stake, which on a $250bn net worth is far more than the tax; if the filing in fact caps it near the liability, my read is wrong.
Meta stating that Class C will not be used as acquisition, compensation or financing currency. Or a Delaware fiduciary suit that forces abandonment within months, as in 2017 — that would show the record was no stronger this time, and that the tax rationale bought nothing.
Pattern — against his own history 2016: Class C proposed, on philanthropy grounds. September 2017: abandoned under a shareholder suit, days before he was to testify, and he sold shares to fund the philanthropy instead. 2026: the same structure, a different justification, board approval and 56% of the vote in writing. One prior attempt, one failure, one success. Separately: he has never previously accepted a contractual restriction on where he lives, and he gave up a compensation increase to get this — a founder who trades cash pay for structure is telling you which one he values.