Meta plans up to $95B in stock sales for AI infrastructure

Meta plans to sell new stock to help finance its AI infrastructure buildout, according to The Information and details from a Meta announcement cited by The Decoder. The plan could raise as much as $95 billion, making clear how expensive the next phase of AI competition has become.

For creators, marketers, freelancers and small businesses, this is not a new product launch. But it is a useful signal: the companies behind everyday AI assistants and creative tools are increasingly competing on compute capacity, not just interface design or model branding.

How Meta plans to raise the money

According to The Information, Meta plans to sell up to $95 billion in shares. The financing is split across several types of stock, rather than a single offering.

The structure described in the source includes:

• About $22.5 billion from sales of Class A common stock, according to Meta’s announcement.
• Another $12.5 billion from depositary shares, with each depositary share representing one-twentieth of a new preferred share.
• A further $60 billion from continuing open-market sales after new Class C shares begin trading.

The preferred shares will automatically convert into common stock in August 2029 unless they are converted earlier, according to the report. The Class C shares are expected to begin trading no earlier than the first quarter of 2027.

That timeline matters because it suggests Meta is not treating AI infrastructure as a short, one-off purchase. The company is setting up a financing structure that can support spending over time.

The voting structure keeps Zuckerberg in control

Meta already has two voting share classes. Class A shares carry one vote each, while Class B shares carry ten votes each. Entities controlled by CEO Mark Zuckerberg hold most of the Class B shares, according to the source.

The newly announced Class C shares carry no voting rights. The effect is straightforward: Meta can raise capital by selling stock without reducing Zuckerberg’s voting control in the same way a new voting share issuance would.

For investors and partners, that makes the raise partly about governance as well as AI. Meta is seeking very large sums for infrastructure while preserving the control structure that has long shaped the company’s decision-making.

For users of AI products, the governance detail may feel distant. But it can influence how quickly and aggressively a platform pursues infrastructure-heavy AI work, because leadership can commit to long-term spending without the same pressure that might come from a more dispersed voting base.

What this says about the AI race

The source does not name specific data centers, chips or Meta AI product upgrades tied to the stock sales. It also does not give a product roadmap. What it does show is that Meta sees AI infrastructure as important enough to support with a potential financing plan of up to $95 billion.

That puts the move in the same broad market context as other major AI companies competing to train, run and distribute advanced models. Readers comparing services such as ChatGPT, Google Gemini and Claude should remember that the visible chatbot is only the front end. Behind it is a large and costly layer of computing infrastructure.

Emonarc has also covered adjacent signs of the same pressure, including Anthropic’s compute deals and OpenAI’s reported valuation target while seeking up to $30B. Meta’s latest plan points in the same direction: AI capability is increasingly tied to access to capital and hardware.

What creators and small teams should watch

There is no immediate change here for the average user. The report is about financing, not pricing, availability or a new Meta feature.

Still, small teams should pay attention to three practical questions. First, does Meta’s spending translate into more capable AI products inside its existing platforms? Second, does the company use infrastructure scale to make AI features more widely available? Third, do large platform investments increase pressure on smaller AI vendors that cannot raise or spend at the same level?

The chip industry angle is also notable. According to Meta’s announcement cited in the source, Nvidia, AMD and Broadcom have expressed non-binding interest in buying up to $5 billion in Meta stock. The source does not say whether those companies will ultimately buy the shares, but their interest underlines how closely AI infrastructure, semiconductor suppliers and platform companies are becoming linked.

For now, the most concrete takeaway is financial: Meta is preparing a major stock-selling program to fund AI infrastructure while keeping voting control largely intact. Product implications may come later, but the capital plan itself shows how much of the AI race is being fought before users ever see a new button.