[ad_1]
- Meta reportedly plans to own more than 340,000 of Nvidia’s highly popular H100 chips this year.
- Meta estimates it will have a total stockpile of 600,000 chips by the end of 2024, The Verge reported.
- CEO Mark Zuckerberg aims to make Meta an AI-first company.
Meta is planning to stockpile hundreds or thousands of popular semiconductors to gain an edge in the AI war.
Mark Zuckerberg has said that by the end of 2024, the company will have more than 340,000 Nvidia H100 GPUs (the primary chip that companies like OpenAI use to train and deploy AI models like ChatGPT). told The Verge.
That’s not all. CEO Zuckerberg expects Meta to accumulate a total of 600,000 GPUs by the end of the year, including Nvidia’s A100 and other AI chips.
The details emerged as companies scramble to acquire Nvidia’s limited supply of AI chips to build the best products. The company’s products are in huge demand from large corporations and startups alike, and the company’s stock price has increased more than 200% in the past 12 months. Overall, global demand for semiconductors is outpacing production.
And Meta seems to be ahead of AI advances on the chip front. Nvidia shipped 150,000 H100 chips to Meta in 2023, according to data from Omdia Research obtained by The Verge. He estimates that Omdia is three times more than companies like Google, Amazon, and Oracle. Only Microsoft delivered a similar amount to Meta.
The H100 chips will cost between $25,000 and $40,000 each, CNBC reported in November, citing estimates from Raymond James.
Meta’s chip pile illustrates the company’s strategy to become an AI-first company. Last July, Meta publicly released Llama 2, a large-scale language model that competes with OpenAI’s GPT-4 model that powers ChatGPT. In September, Meta released the latest version of its Ray-Ban smart glasses with an AI assistant. In the same month, the company also announced a celebrity AI chatbot available on Messenger and WhatsApp. Zuckerberg said he is currently training Llama 3.
Mr. Zuckerberg is aiming to enable even smarter AI technology, referred to by some experts as “AGI” or “artificial general intelligence.” According to The Verge, the CEO hasn’t yet given clarity on what that will look like or when it will actually happen, but he believes it will eventually happen as a gradual process.
“The truth is, I don’t know if certain thresholds are felt that deeply.”
Meta did not respond to Business Insider’s request for comment prior to publication. Nvidia declined to comment.
Although the semiconductor shortage will continue into 2023, things could improve for tech buyers this year.
Last September, Microsoft Chief Technology Officer Kevin Scott said Microsoft was making it easier to obtain Nvidia chips, CNBC reported.
Still, Zuckerberg said Meta’s current fleet of NVIDIA AI chips could pose a problem for other companies looking to capitalize on the hype.
“We’ve built the ability to do this at a larger scale than any other individual company,” he told The Verge. “I think a lot of people might not understand that.”
[ad_2]
Source link
