Mark Zuckerberg spends billions of dollars buying AI chips from Nvidia.

In an Instagram Reels post on January 18th, Mark Zuckerberg stated that the company's "future roadmap" for AI requires building a "massive computing infrastructure." By the end of 2024, this infrastructure will be served by 350,000 Nvidia H100 chips.

Mark Zuckerberg says AI is Meta's top investment priority in 2024. (Photo: Bloomberg)

Meta's CEO didn't disclose how many graphics processing units (GPUs) they purchased, but the H100 is set to launch in late 2022 and supply is extremely limited. Analysts at Raymond James estimate Nvidia is selling the H100 for between $25,000 and $30,000. On eBay, the chip could fetch over $40,000. Even if Meta acquired it at a lower price, the cost would still be close to $9 billion.

In addition, Zuckerberg revealed that, including other GPUs, the computing power of the infrastructure is equivalent to 600,000 H100s. In December 2023, Meta, OpenAI, and Microsoft announced that they would be using AMD's new Instinct MI300X AI chip.

Meta needs such "heavy-duty" computer chips because the company is pursuing artificial general intelligence (AGI), which Zuckerberg calls the company's "long-term vision." OpenAI and Google DeepMind are also researching AGI – a type of AI of the future that rivals human intelligence.

Meta's chief scientist , Yann LeCun, emphasized the importance of GPUs at an event last month in San Francisco, USA. He argued that if you want AGI, you have to buy more GPUs. "There's an AI battle going on, and he (Nvidia CEO Jensen Huang) is providing the weapons."

In its Q3 2023 earnings report, Meta announced that total spending in 2024 would range from $94 billion to $99 billion, partly to expand its computing capacity. During a meeting with analysts, the company's head also affirmed that "in terms of investment priorities, AI will be the biggest investment area in 2024, both in terms of technology and computing resources."

Also on January 18th, Zuckerberg stated his intention to "responsibly" open-source the AGI that the company is developing, a similar approach to the larger Llama language family. Meta is training Llama 3 and will bring the Fundamental AI Research (FAIR) and GenAI research teams closer together.

Immediately after Zuckerberg's post, engineer LeCun wrote on X: "To speed things up, FAIR is now a sister organization to GenAI, the AI product division."

(According to CNBC)

Source