Meta’s Multi-Billion Dollar Investment in Nvidia AI Chips Unveiled by Mark Zuckerberg
In a recent Instagram Reels post, Meta CEO Mark Zuckerberg disclosed the company’s significant financial commitment to Nvidia’s cutting-edge computer chips, essential for advancing artificial intelligence (AI) research and projects.
Zuckerberg emphasized Meta’s dedication to building a robust AI infrastructure, stating that their “future roadmap” necessitates a “massive compute infrastructure.” By the close of 2024, Meta aims to integrate 350,000 H100 graphics cards from Nvidia into this evolving infrastructure.
While Zuckerberg didn’t reveal the current number of graphics processing units (GPUs) acquired, industry analysts estimate the cost of Nvidia’s H100 to range between $25,000 and $30,000, with prices potentially exceeding $40,000 on eBay. At the lower end of this range, Meta’s investment in H100 GPUs could approach a staggering $9 billion.
Furthermore, Zuckerberg outlined Meta’s plan to include “almost 600k H100 equivalents of compute” when factoring in other GPUs. In December, major tech players such as Meta, OpenAI, and Microsoft announced their adoption of AMD’s new Instinct MI300X AI computer chips.
The substantial investment in these high-performance computer chips aligns with Meta’s strategic focus on artificial general intelligence (AGI), described by Zuckerberg as a “long-term vision” for the company. AGI represents a futuristic form of AI comparable to human-level intelligence, a field also explored by research entities like OpenAI and Google’s DeepMind unit.
Meta’s Chief Scientist Yann LeCun underscored the pivotal role of GPUs during a recent media event, asserting, “[If] you think AGI is in, the more GPUs you have to buy.” LeCun characterized the current landscape as an “AI war,” with Nvidia CEO Jensen Huang being a key supplier of the necessary “weapons.”
In Meta’s third-quarter earnings report, the company projected total expenses for 2024 in the range of $94 billion to $99 billion, attributing a significant portion of these expenditures to computing expansion. Zuckerberg affirmed AI as Meta’s primary investment area for 2024, encompassing both engineering and computing resources.
Zuckerberg also revealed Meta’s commitment to “open source responsibly” its forthcoming “general intelligence” and reiterated this approach with its Llama family of large language models. The company is currently engaged in training Llama 3 and fostering closer collaboration between its Fundamental AI Research team (FAIR) and GenAI research team to accelerate progress in the evolving AI landscape.
JUST IN:
Facebook $META CEO Mark Zuckerberg just released this video giving some updates on Meta’s AI efforts
Zuckerberg said they will have ~350K Nvidia $NVDA H100s by the end of this year or ~600K H100 equivalents of compute if you include other GPUs pic.twitter.com/RIqNb72nA5
— Evan (@StockMKTNewz) January 18, 2024