NVIDIA Invests $3 Billion in Meta and Expands Muse Agent Collaboration

Zhidongxi reported on October 27 that today, Nvidia announced it will invest $3 billion (approximately 20 billion yuan) in Meta's equity offering. At the same time, the two parties will expand their collaboration on the personal agent product Muse and the Muse series models to improve the models' inference and training performance.

Nvidia announcement (Image source: X platform)

Nvidia's investment is part of Meta's larger financing plan. On October 26, Meta announced that it plans to conduct a round of fundraising of $95 billion (approximately 633.5 billion yuan), raising $35 billion through an underwritten public offering of stock on the one hand, and then raising another $60 billion through an at-the-market offering plan. In addition to Nvidia, AMD and Broadcom also intend to invest in Meta.

Meta expects its full-year capital expenditure to reach between $135 billion and $145 billion (approximately 901.3 billion to 968 billion yuan), and this round of financing will support its growing capital expenditures.

Nvidia's announcement shows that in addition to financing, Meta and Nvidia plan to carry out optimization cooperation across various links to maximize AI infrastructure performance and improve total cost of ownership. Currently, Meta's Muse models already run on Nvidia's infrastructure, and the Muse Wonder model achieved frontier-level benchmark performance when it was released in October this year.

The collaboration also involves the recently popular personal agent product Muse. In early September this year, Meta officially released Muse, a general-purpose agent with cross-web task-handling capabilities. Ten days after launch, Muse topped the overall free app chart on the U.S. App Store.

Meta's stock price surged due to Muse's popularity, once reaching a high of $779.82 per share, with its total market value approaching $2 trillion. After Muse, personal agents have also become a new battleground for AI giants. Recently, from OpenAI and Anthropic abroad to Alibaba, Ant, Manus, Tencent and other companies domestically, companies have successively launched agent products with similar approaches.

In this collaboration, Meta will consider adopting Nvidia's OpenShell secure agent reference architecture, including OpenShell for secure deployment and Sentry for monitoring agent behavior.

Meta also plans to expand the deployment of Nvidia infrastructure, including Vera Rubin and Vera CPU for next-generation AI agents, to maximize performance per watt. Meta will also deploy Nvidia Spectrum-X Ethernet networking and BlueField-4 to improve network efficiency and throughput.

On Meta's earnings call last week, Meta founder and CEO Mark Zuckerberg said that about a month and a half after Muse launched, more than 10 million people in the United States and Canada had set up the Muse personal agent, with a daily active usage rate of about 25%. Shopping is one of the most common use cases, with more than a quarter of users participating, and an average order value of more than $200.

Amid this financing, Meta's stock price has fallen for several consecutive days. As of the close on October 26, U.S. Eastern Time, Meta's stock price fell 1.88% to $630.56, with a total market value of $1.61 trillion (approximately 10.71 trillion yuan).

Meta stock price (Image source: Tencent Self-Selected Stocks)

Conclusion: Agents are moving toward ordinary users, and collaboration between the application and computing power layers is becoming increasingly important

For Nvidia, the popularization of agent applications is expected to further drive demand for model training, inference and underlying infrastructure.

This article is from the WeChat public account "Zhidongxi" (ID: zhidxcom), author: Chen Junda, editor: Li Shuiqing, published with authorization from 36Kr.