Meta Plans a $35 Billion Stock Sale to Fund Its AI Buildout
Meta plans a $35 billion stock offering to help fund its AI buildout, The Wall Street Journal reported on October 26, 2026. The same day, CNBC reported that Meta plans to raise $95 billion to fund AI infrastructure and superintelligence. The WSJ put the total financing plan at up to $95 billion, which makes the equity sale a $35 billion slice of a much larger program.
The equity component already has anchor interest. The WSJ reported that Nvidia, Advanced Micro Devices and Broadcom have indicated interest in buying up to $5 billion of common stock in the offering. On October 27, Reuters reported that Nvidia will invest $3 billion in Meta's equity offering. Reuters also reported that Nvidia and Meta will partner to expand their work on Meta's Muse assistant and the Muse family of models. The Reuters figure is the only one of the three chipmakers' commitments reported as firm.
Nvidia has also announced a strategic partnership with Marvell, the data infrastructure semiconductor firm, that includes a $2 billion investment in Marvell, as the WSJ reported in an earlier article. That deal is separate from the Meta transaction.
In my view, the Marvell deal is worth a second look for the same reason as the Meta one: a large chip supplier is putting its own balance sheet into a company it sells to.
For a non-specialist, the mechanics are simple. A stock offering sells new shares. The company gets cash, and existing holders own a slightly smaller slice of the business. The per-share cost depends on how many shares are issued and the price they fetch. Equity has no fixed coupon, meaning no required interest payments, but it permanently dilutes the claims of current shareholders. A company raising tens of billions for capital spending has to weigh that trade against borrowing capacity, credit ratings, and the cash its existing operations generate.
The $95 billion ceiling is, to my mind, the bigger number. If the offering is $35 billion, the remaining sum the plan contemplates is roughly $60 billion. Investors evaluating the equity piece will want to know how the rest is funded, since debt issuance and free cash flow carry different risks for the same capital-spending program.
The detail I would watch most closely is the participation of chip vendors. Nvidia, AMD and Broadcom sell the accelerators and networking hardware that make AI data centers run. When a supplier invests equity in a large buyer of its products, the revenue it books from that buyer and the return on its stake become linked. That does not make the investment unsound. It does complicate the usual reading of supplier demand, because part of that demand is now financed by the supplier's own shareholders' capital. Analysts will be trying to separate genuine end-customer spending from spending partly recycled through equity.
To put it plainly, two things are established: a $35 billion offering is planned, and Nvidia will invest $3 billion in it. The final size, the pricing, and whether the AMD and Broadcom interest turns into purchases are still prospective. Anyone tracking Meta's share count, its capital-spending guidance, or the credit profile of the broader AI infrastructure build should treat $95 billion as the frame and $35 billion as the first concrete piece.
