Meta CEO Mark Zuckerberg posted on Instagram Reels on January 18 that the company’s “future roadmap” for AI requires building a “massive computing infrastructure” that will be powered by 350,000 Nvidia H100 chips by the end of 2024.
 |
| Mark Zuckerberg says AI is Meta's top investment priority in 2024. |
Zuckerberg did not say how many GPUs he bought, but the Nvidia H100 is due to hit the market in late 2022 and is in extremely limited supply. Analysts at Raymond James estimate Nvidia is selling the H100 for $25,000 to $30,000. On eBay, the chip can fetch more than $40,000. If Meta gets a lower price, the cost would be close to $9 billion.
In addition, Meta CEO also revealed that if other GPUs are included, the computing power of the infrastructure is equivalent to 600,000 H100. In December 2023, Meta, OpenAI and Microsoft said they would use AMD's new Instinct MI300X AI chip.
Meta needs so many “heavy” computer chips because the company is pursuing artificial general intelligence (AGI), which Zuckerberg calls a “long-term vision.” OpenAI and Google’s DeepMind are also working on AGI – a futuristic AI that is expected to rival human intelligence.
Meta chief scientist Yann LeCun also emphasized the importance of GPUs at an event last month in San Francisco, US. He said that if you want AGI, you have to buy more GPUs. “There is an AI race and he (Nvidia CEO Jensen Huang) is providing the weapons.”
In its Q3 2023 earnings report, Meta said its total spending in 2024 would range from $94 billion to $99 billion, partly to expand its computing capabilities. In a meeting with analysts, the group’s chief executive also confirmed that “in terms of investment priorities, AI is the largest investment area in 2024, both in terms of technology and computing resources.”
Mark Zuckerberg said Meta plans to “responsibly open source” the “general AI” model it is developing, an approach the company has taken with its large language product line Llama.
Meta is currently training Llama 3, while also supporting the Fundamental AI Research (FAIR) team and the GenAI research team to work more closely together.
Source
Comment (0)