Meta plans to buy 350,000 Nvidia GPUs to build general artificial intelligence

Meta plans to buy 350,000 Nvidia GPUs to build general artificial intelligence

Meta Platforms Inc. CEO Mark Zuckerberg revealed today that his company plans to buy about 350,000 high-end H100 GPUs from Nvidia Corp. to provide computing power to advance its goals in a new field of artificial intelligence.

The company is striving to become a leader in artificial general intelligence, or AGI, a potentially advanced artificial intelligence that can match the intelligence of humans and perform a wide range of different tasks.

Zuckerberg (pictured) said in an Instagram Reels video posted today that AGI may eventually be able to power a wide range of cutting-edge services and devices, including more advanced digital assistants, augmented reality glasses, and more. The company’s goal is to build “the best AI assistants, AI for creators, AI for businesses and more,” he said, adding that doing so will require advances in every area of ​​AI.

In order to achieve this, Meta will need to make a significant investment in AI computing infrastructure, hence its plans to acquire mountains of Nvidia’s powerful H100 GPUs. The H100 chip is Nvidia’s most advanced chip ever, and is said to be particularly adept at training large language models that support generative AI like ChatGPT.

“We’re building a tremendous amount of infrastructure to support this,” Zuckerberg said. “By the end of the year, we will have about 350,000 Nvidia H100 units, or about 600,000 compute-equivalent H100 units if you include other GPUs.”

It’s a staggering number, not least because of the enormous cost required to buy all that silicon. Although the CEO didn’t say how much the company intends to spend on its GPUs, it’s likely to be a staggering number. Furthermore, it’s unlikely that Meta has amassed all that much yet, as the H100 only went on sale in late 2022, and supply has been limited.

Analysts at Raymond James estimate that Nvidia is selling the H100 for $25,000 to $30,000, while it could fetch more than $40,000 on eBay. Although Meta will certainly get some credit for the bulk purchase, it will still likely spend upwards of $9 billion on GPUs.

Zuckerberg said 350,000 H100 GPUs will be added to the infrastructure, which will also contain about 600,000 H100-equivalent GPUs, which could mean the company is also buying competing chips in bulk, such as Advanced Micro Devices’ new Instinct M1300X AI accelerator. Inc.

The comments provide insight into how much money big tech companies are spending to stay at the forefront of AI development. Research conducted by Omdia indicates that other major players, such as Microsoft Corp. Google LLC and Inc. are all believed to have purchased between 50,000 and 150,000 H100 GPUs last year, and likely have similar plans to add them to their PC stacks.

The importance of GPUs to Meta’s ambitions was emphasized by chief AI scientist Yann LeCun during an interview last month. “If you believe general artificial intelligence exists, the more GPUs you have to buy,” he said.

Finally, Meta’s total expenses for 2024 will likely range between $94 billion and $99 billion, the company said in its latest quarterly financial report. Expansion in computing is expected to be one of the largest expenditures. “In terms of investment priorities, artificial intelligence will be our biggest investment area in 2024, whether in engineering or computer resources,” Zuckerberg said in a conference call with analysts.

In his Instagram post today, Zuckerberg revealed that Meta’s GPUs are already being used to good effect in training Llama 3, an open source alternative to OpenAI’s ChatGPT models. However, the company is taking a different approach to its rival, and plans to open source the Llama 3 model, just as it did with its predecessor, the Llama 2.

“This technology is so important and the opportunities are so great that we should open source and make it as widely available as possible,” Zuckerberg insisted.

Photo: Anthony Quintano/Flickr

Your upvote is important to us and helps us keep the content free.

One click below supports our mission of providing free, deep, relevant content.

Join our community on YouTube

Join a community of over 15,000 #CubeAlumni experts, including CEO Andy Jassy, ​​Dell Technologies Founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more notable figures and experts.

“TheCUBE is an important partner to the industry. You guys are really a part of our events and we really appreciate you coming and I know people appreciate the content you create too.” -Andy Jassy

Thank you

You may also like...

Leave a Reply