Meta's AI Spending Pushes Free Cash Flow Negative, Triggers Board Conflict and Strategic Shifts

Meta reported $2.12 billion in negative free cash flow for the third quarter of 2026 — its first cash burn in more than a decade — as capital expenditures hit $34.91 billion [7] and the company raised its full-year spending forecast to $135–145 billion [8]. The financial strain has sparked a boardroom confrontation over CEO Mark Zuckerberg's compensation, prompted a new non-voting share class to fund AI bets, and coincided with an acquisition and a delayed policy change after publisher backlash.

Negative cash flow and a higher spending ceiling

Meta's Q3 results showed revenue of approximately $63.3 billion, with the Family of Apps segment contributing roughly $62.8 billion [2]. Free cash flow swung from a positive $10.63 billion in Q3 2025 to negative $2.12 billion [8]. Consensus estimates place full-year 2026 capex near $140 billion, nearly double the 2025 level [2]. The company increased the lower end of its 2026 capex guidance by $5 billion, citing a "sweeping plan to put its viral Muse agent at the heart of Meta's AI strategy" [1].

Investors reacted sharply: shares fell more than 7% on October 22, wiping over $124 billion from market capitalization [3], and the cost of insuring Meta's debt against default hit an all-time high [1]. "More spending, negative free cash flow, not enough revenue being generated from the spend (is) concerning investors that this is just sort of a bit of a money pit," said Robert Pavlik of Dakota Wealth [8].

Board conflict over compensation and control

The spending surge has fueled a confrontation between Zuckerberg and Meta's board. According to the Wall Street Journal, Zuckerberg demanded a "massive stock-heavy pay raise," leading to what the paper called a "war over power and money" [4]. A three-member negotiation committee prepared a contingency plan in early September that included the possibility of their own resignation if Zuckerberg refused to back down [4].

The board's special committee described "robust negotiations" that "were protracted and, at several points, contentious" [3]. It concluded that Zuckerberg's "continued leadership role ... is critical at this current inflection point" and that the benefits "substantially outweigh any countervailing considerations" [3]. RBC analyst Brad Erickson characterized the outcome as "the board saying, hey he's our guy, we're gonna stick with this" [3].

Class C shares: funding, control, and a wealth-tax hedge

To finance its AI buildout, Meta will create a new class of non-voting Class C shares, aiming to raise "hundreds of billions of dollars over several years" with issuances beginning as early as fall 2026 [1]. The shares take effect January 19, 2027, via a one-time dividend of one Class C share for each outstanding Class A and Class B share [7].

The structure serves multiple purposes. It protects Zuckerberg's roughly 60.6% voting control while he owns about 13.4% of the company [7]. It also addresses a potential California wealth tax: voters decide next month on a one-time 5% levy on billionaires' assets as of January 1, 2026 [3]. Zuckerberg told the board he might leave California if the tax passed; the Class C award lets him sell shares to pay any liability without diluting his voting power [3]. He pledged to remain in California for eight years as part of the agreement [3].

Two drawbacks accompany the new shares. Meta expects Class C shares to trade at a discount to voting stock, increasing dilution for existing shareholders in future raises [7]. They also cannot meet voting-stock requirements for certain tax-deferred acquisitions, limiting their utility in dealmaking [7].

Muse adoption and Reality Labs losses

Meta's new AI assistant Muse shows early traction: approximately 10.9 million downloads since its September launch and an estimated 2.5 million daily active users by mid-October [2]. The company frames Muse as a path to turn a technology popular with businesses automating coding into a "consumer moneymaker" [1]. However, downloads and daily active users are "early indicators of product adoption, not measures of profitability" [2]. Monetization, retention, and the cost of serving a large user base remain unproven [2].

Reality Labs continues to weigh on results, with a projected Q3 operating loss of roughly $5.7 billion and a potential full-year loss near $20.7 billion [2]. The segment is distinct from AI infrastructure but raises the same capital-allocation questions for investors [2].

Sundial acquisition and scraping-terms delay

Amid the spending push, Meta acquired Sundial, an AI analytics platform founded by former Meta employees. Co-founder Chandra Narayanan will lead "data analytics tooling and agents work," according to chief data officer Alex Shultz [5]. The deal size was not disclosed.

Separately, Meta delayed controversial changes to its commercial terms that would have explicitly allowed scraping business users' websites for AI training without a contract or compensation. The effective date moved from October 30 to November 20 "to allow more time for us to explain to our partners and stakeholders what's changing and what's not" [6]. The revised terms still state Meta's "automated software" can grab "any content" for "advertising and business products" [6]. The News/Media Alliance called the move "an exercise of market power and is anticompetitive" and warned it "risks undermining the robust licensing market for publishers' content" [6].

What remains unknown

• Whether Muse can convert early adoption into recurring revenue at margins that justify the infrastructure spend.
• The size and terms of the Sundial acquisition.
• Whether the Class C share discount will materialize as Meta expects and how much dilution shareholders will absorb.
• The outcome of California's wealth-tax vote and its practical impact on Zuckerberg's holdings.
• Whether the delayed terms-of-service change will satisfy publishers or merely postpone the conflict.

Meta's advertising engine remains powerful, but the gap between AI spending and demonstrable returns from that spending has widened — and the board, bondholders, and publishers are all pushing back.