Nvidia invests $3B in Meta as AI partnership deepens

Nvidia plans to put $3 billion into Meta Platforms. The move ties the world’s most closely watched chipmaker even tighter to one of its biggest-spending AI customers.

The investment rides alongside a fundraising plan from Meta that is enormous by any standard. Meta intends to raise $95 billion to pay for its artificial intelligence infrastructure buildout, and Nvidia wants a seat at that table.

How the $95 billion raise works

Meta laid out the plan on Monday, and it comes in two stages. The first is $35 billion in underwritten public equity offerings, which means banks commit to selling a defined block of shares.

The second stage is a $60 billion at-the-market program. Think of it as a slow drip rather than a single pour: the company can sell shares into the open market over time at whatever the going price happens to be.

The initial tranche of roughly $35 billion includes $25 billion in Class A common stock priced at $625 per share. That offering was upsized, a signal that demand showed up.

The total raise could ultimately land at approximately $95 billion to $97.5 billion after the upsizing.

Meta’s announcement also said Nvidia, Advanced Micro Devices, and Broadcom were interested in buying up to $5 billion in shares. That interest is described as non-binding, so it remains an expression of intent rather than a signed check.

Nvidia then made its own commitment explicit: $3 billion. Its two chipmaking peers have not been tied to specific amounts in the source material.

Muse sits at the center of the partnership

The money is only part of the story. The two companies are broadening a partnership focused on optimizing Meta’s Muse AI models and agents.

Muse launched in September 2026 as a personal AI agent built to handle tasks across Meta’s various applications.

The partnership aims to improve how Muse is trained and how it runs inference on Nvidia hardware. That work will likely include greater use of Nvidia’s Vera Rubin GPUs and its Spectrum-X networking technology, per the research findings. Nobody has stated that as a finalized deployment plan, so treat it as the probable direction rather than a purchase order.

Both companies’ CEOs framed AI agents as a transformative new computing paradigm.

Meta’s spending appetite keeps growing

This raise does not exist in a vacuum. Meta has lifted its full-year capital expenditure guidance to between $135 billion and $145 billion, driven by planned AI supercluster builds.

Muse is regarded as a critical piece of Mark Zuckerberg’s long-term artificial intelligence strategy.

What this means for Nvidia, Meta and the AI trade

The most interesting wrinkle is the circularity. Nvidia is investing in a company that will spend heavily on infrastructure that likely includes Nvidia’s own chips and networking gear.

The involvement of AMD and Broadcom adds another layer. Three chip suppliers signaling interest in the same customer’s stock suggests they see value in staying close to Meta’s buildout, even if only Nvidia has named its figure so far.

Key things to track from here include how quickly Meta draws down the $60 billion at-the-market program and whether the non-binding interest from AMD and Broadcom turns into firm commitments. Also worth watching: whether Vera Rubin and Spectrum-X deployments actually scale inside Meta’s data centers as the research findings anticipate.