Meta's $95 Billion AI Funding Plan Puts Dilution in Focus
Meta launched $35 billion of equity offerings and may add a $60 billion Class C program. The plan diversifies AI funding but introduces dilution and execution risk for shareholders.
Meta is not raising the full $95 billion immediately. The company launched about $35 billion of underwritten equity offerings on October 26, while a separate $60 billion at-the-market program for a new non-voting share class remains a future plan with no agreement, final terms or assurance that it will proceed. That distinction matters because only part of the headline amount is now being marketed.
The initial package combines approximately $22.5 billion of Class A common stock with about $12.5 billion of depositary shares representing mandatory convertible preferred stock. Meta said the preferred securities are scheduled to convert around August 15, 2029 into a variable number of common shares. The common-stock leg uses forward sale agreements, so Meta does not expect to receive those proceeds at closing. It can settle the agreements on dates it chooses through November 19, 2027, subject to the contracts' terms.
Meta plans to begin the separate $60 billion program only after its Class C shares start trading, which the company expects in the first quarter of 2027. Those shares carry no voting rights, so issuing them can raise capital without giving new investors voting power. Existing holders would still face economic dilution if Meta sells the shares. The company explicitly said the program's timing, size and terms have not been determined.
The financing plan shows how rapidly Meta's AI spending has outgrown even its formidable internal cash generation. Meta expects $135 billion to $145 billion of capital expenditures in 2026. It generated $97 billion of operating cash flow during the first nine months and ended the third quarter with about $82 billion of cash, cash equivalents and marketable securities. Third-quarter revenue rose 26% to $64.6 billion, while net income was $9.9 billion.
Meta says the spending is supporting four superclusters with more than one gigawatt of capacity each and is helping the company nearly double infrastructure capacity in 2026. Management also argues that AI is already improving engagement, advertising performance and monetization. The funding question is therefore not whether the core business can produce cash, but whether returns from the next wave of infrastructure can exceed the cost of a much larger capital base.
The SEC prospectus used an assumed forward price of $642.62 for the Class A offering and reported 2.214 billion Class A shares outstanding at September 30. Dividing $22.5 billion by that assumed price produces roughly 35 million shares, or about 1.6% of the existing Class A count, in a BTI calculation. The actual share count and proceeds will depend on pricing, adjustments and settlement. This estimate also excludes the preferred conversion, any underwriter options and the possible $60 billion Class C program.
Nvidia, AMD and Broadcom indicated interest in buying up to $5 billion of the Class A offering in aggregate. The SEC filing says those indications are not binding, and the chipmakers may buy more, fewer or no shares. Investors should treat that interest as a signal of strategic alignment, not as a completed commitment.
Equity funding diversifies Meta's capital sources and can limit further balance-sheet strain, but it transfers more of the AI build-out's risk to shareholders through potential dilution. The central test is now measurable: Meta must turn its expanded compute capacity into durable cash returns fast enough to justify issuing equity despite having a highly profitable advertising business.



