IA en un minutoNewsBusiness and products
NVIDIA will invest $3 billion in Meta and expand its AI alliance
IA en un minuto newsroom · Editor: Jon Elgezabal
In 30 seconds
NVIDIA will contribute 3 billion dollars to the new shares Meta is putting on the market, a proposed transaction that still has to close. Meta and NVIDIA will extend their work on Muse, Meta's own AI models, to train and run them with better performance. Meta also wants to roll out more NVIDIA hardware, Vera Rubin and the Vera chips, built for the next generation of agents.
NVIDIA is going to put money into Meta: 3 billion dollars. According to NVIDIA's announcement of October 26, the company will take part in Meta's share offering with that amount. It is a proposed transaction, not a closed one, and its completion depends on conditions.
According to Forbes España, which carries the text together with Europa Press, Meta said on Monday that it wants to raise up to 95 billion dollars to finance its AI investments, through offerings of common stock and convertible preferred securities. It expects to sell about 35 billion through public offerings and, in 2027, to launch an at-the-market sales program of approximately 60 billion. Meta added that NVIDIA, AMD and Broadcom showed interest, "individually and not jointly", in buying up to 5 billion in total.
The two companies will expand their collaboration on Muse, the Meta family of AI models, to improve the performance of training and inference, and will tune their workflows to do more for less and cut the total cost of ownership. The announcement says that Muse and its agents, which run on NVIDIA infrastructure, have reached top-tier results in benchmark tests and a sharp rise in downloads since launch.
Meta plans to expand its deployment of NVIDIA infrastructure, with Vera Rubin systems and Vera CPUs, so that next-generation AI agents deliver the most performance per watt. It will also deploy Spectrum-X Ethernet networking and BlueField-4 to improve network efficiency and throughput. The companies are also looking at working together on NVIDIA's OpenShell reference architecture: OpenShell, for deploying agents securely, and Sentry, for monitoring their behavior. They are considering it; it is not a closed agreement.
Jensen Huang, NVIDIA's founder and chief executive, said in the announcement that AI agents "represent a new computing paradigm".
Why it matters · analysis and opinion
The announcement mixes two moves that are worth separating. The commercial one: Meta will keep buying NVIDIA infrastructure and, on top of that, its supplier would become a shareholder. That double relationship says something about the market: whoever makes the chips also finances whoever needs them, while Meta pays for its AI investments through share issues. The technical one is halfway there: OpenShell and Sentry are being considered, not agreed. For a company that wants to deploy agents, the useful signal is that security and monitoring of their behavior are already treated as part of the infrastructure rather than as a later add-on. What remains unknown is whether the transaction closes and on what conditions.
Source: StockTitan · Written with the help of AI: how we make the news
