Nvidia to put $3 billion into Meta shares, AMD adds $1 billion

Nvidia, the world's biggest maker of chips for artificial intelligence, said on Tuesday, October 27, 2026, that it will invest $3 billion in Meta Platforms' new share offering, as part of a wider partnership on Meta's Muse family of AI models. Chipmaker AMD is putting in a further $1 billion, according to reports. The commitments came a day after Meta laid out plans to raise up to about $95 billion to pay for its AI data-centre build-out, starting with a share sale of about $35 billion in common and preferred stock.

Building AI is now as much a money problem as a technology problem. Training and running large AI models takes huge data centres full of specialised chips, along with the power and cooling to run them. Meta is among the biggest spenders, and it has chosen to raise part of the money by selling new shares, not only by borrowing or using cash from its advertising business. When a company sells a large block of new shares, existing shareholders worry that the extra supply will push the price down. A big, committed buyer helps calm that fear.

Nvidia is not an ordinary buyer. Meta is one of its largest customers, and a large part of the money Meta raises will be spent on AI hardware, much of it from Nvidia. Critics call this circular financing: a supplier invests in a customer, which then uses the money to buy the supplier's products. Supporters say it is a normal way for close partners to share the risk and reward of a long project. Either way, the arrangement ties the fortunes of the two companies more tightly together, and it shows how central a few chip suppliers have become to the AI build-out.

The Muse part of the deal is about software as well as hardware. Inference is the work of running a trained model to answer a request, such as a chatbot reply. Training is the earlier, heavier stage of teaching the model. By working together on both, Meta hopes to make its models faster and cheaper to run on Nvidia's chips, and Nvidia gets a large customer that tunes its software to Nvidia's hardware.

For Indian businesses, the scale of this spending matters in two ways. It keeps the global price of AI computing high while demand runs ahead of supply, which affects what Indian firms pay to rent AI capacity from cloud providers. It also shows that the race to build AI infrastructure is pulling in very large sums of equity capital, a trend that data-centre and technology services companies in India are watching closely.

Short-term: Meta's share sale gets a committed anchor of $4 billion from two chipmakers, which may steady investor nerves around the offering. Nvidia and AMD gain a financial stake in one of their largest customers.

Long-term: If AI spending keeps paying off, the partnership could lower Meta's cost of running its models and secure years of orders for Nvidia. If returns from AI disappoint, the circular links between suppliers and customers mean losses could spread across the companies involved.

Who is affected: Meta's existing and new shareholders; Nvidia and AMD investors; companies that rent AI computing power from cloud providers; and data-centre builders and technology firms worldwide, including in India, that depend on the same supply of AI chips.

Questions and Answers