Nvidia 向 Meta 股权投资 30 亿美元，合作优化 Muse 模型训练与推理

What just happened? One day after Meta announced plans to raise $95 billion to fuel its AI spending, Nvidia has confirmed it will invest $3 billion in Meta's equity offering. Team Green calls the move a strengthening of the companies' AI partnership, one that will deepen their collaboration on the Muse AI models.

When Meta announced its stock offering yesterday, it was reported that Nvidia, AMD, and Broadcom had indicated interest in buying up to $5 billion worth of the company's shares collectively. The world's largest company by market cap has now confirmed its contribution.

The original fundraising plans comprised roughly $35 billion in initial offerings and a separate $60 billion share-sale program, expected after Meta's new nonvoting Class C shares begin trading in early 2027. That later program remains subject to final terms. Reuters reports that Meta increased the overall fundraising figure to $97.5 billion on Tuesday.

Nvidia says the expanded partnership will focus on improving training and inference performance across Muse and its family of models, while reducing the cost of running them. The companies claim Muse has delivered strong benchmark results and surging downloads since launch, though there are no included figures.

There is, unsurprisingly, more Nvidia hardware involved. Meta plans to expand its deployment of Vera Rubin systems and Vera CPUs, alongside Spectrum-X Ethernet networking and BlueField-4 technology. The aim is to squeeze more performance from each watt and improve the movement of data across its infrastructure.

The pair are also considering further work on Nvidia's OpenShell framework for deploying agents safely and Sentry for monitoring their behavior.

Muse sits at the center of Mark Zuckerberg's push for personal superintelligence. Meta describes it as an agent that can take action across everyday services, from sending emails to booking travel, with users deciding how much access to grant it.

Meta expects capital expenditure of $135 billion to $145 billion this year, nearly all of which is fueling its AI infrastructure buildout. The tech giant's latest quarterly results showed revenue rising 26% to $64.6 billion.

For Nvidia, investing in a customer that plans to deploy more of its hardware adds to more questions about circular financing across the AI industry, which comes down to chip suppliers backing the companies that buy their products.