On October 26, Meta Platforms announced plans to raise a total of $95 billion through equity offerings to fund its artificial intelligence infrastructure expansion and superintelligence research. The news came right after the company raised its full-year capital expenditure forecast to a record high, causing its stock to fall slightly in morning trading.
This large-scale fundraising plan reflects the dizzying pace at which AI infrastructure costs are climbing, similar to the recent case where SpaceX borrowed $40 billion to buy GPUs.
Meta’s Two-Stage Fundraising Structure
According to the official announcement, Meta will roll out the offering in two specific stages. In the near term, the company will issue $35 billion of mandatory convertible preferred stock and Class A common stock through underwriters. Subsequently, the group plans to raise an additional $60 billion through an at-the-market equity offering program.
Notably, three leading semiconductor suppliers—Nvidia, AMD, and Broadcom—have expressed interest in purchasing up to $5 billion worth of shares in this offering. However, this represents preliminary interest and does not constitute a finalized purchase commitment. The plan follows other capital structure considerations, such as the non-voting stock issuance plan that the group previously discussed.
Capital Expenditure Pressure and Supercluster Data Center Construction
The new capital is designed to support the expansion strategy for its portfolio of applications, the Muse assistant platform, and its data center operations. Meta is currently constructing four data center superclusters, each with a capacity exceeding 1 gigawatt, aiming to nearly double its total computing capacity in 2026 compared to the prior year.
Meta’s third-quarter earnings report last week showed revenue reaching $64.60 billion, up 26% year-over-year. Nevertheless, investors expressed caution as free cash flow turned negative and management lifted the lower end of its full-year capital expenditure forecast from $130 billion to between $135 billion and $145 billion.
This situation highlights that even tech giants cannot rely solely on advertising cash flows to compete in the infrastructure race, but must instead combine equity and debt markets amid the increasingly apparent threat of a hardware credit squeeze.
Whether continuously pouring tens of billions of dollars into gigawatt data centers will yield commensurate margins, or continue to erode cash flows for major tech firms in the coming years, remains an open question.