Meta Platforms, the parent company of Facebook, is reportedly planning to deploy a new iteration of its custom chip into its data centres, as per an internal document reviewed by news agency Reuters.
The chip, which was announced last year, is a part of Meta’s efforts to bolster its artificial intelligence (AI) capabilities. With this, the company aims to lessen its reliance on Nvidia chips and curb mounting expenses associated with AI development.
The move comes amid Meta’s goal to integrate generative AI products across its platforms such as Facebook, Instagram, WhatsApp, and hardware like Ray-Ban smart glasses.The company has been investing substantially in specialised chips and data centre reconfigurations to run these technologies.
What Meta has to say
A Meta spokesperson confirmed plans to commence production of the updated chip in 2024, saying that the chips complement commercially available GPUs, which Meta continues to procure in large volumes.
“We see our internally developed accelerators to be highly complementary to commercially available GPUs in delivering the optimal mix of performance and efficiency on Meta-specific workloads,” the spokesperson was quoted as saying.
The report said that a successful deployment of its own chip at the scale at which Meta operates may cut hundreds of millions of dollars in annual energy costs and billions in chip purchasing costs.
Meta to purchase 350,000 Nvidia GPUs
Earlier this year, it was reported that the company plans to acquire approximately 350,000 flagship “H100” processors from Nvidia by the end of the year. The new chip, internally dubbed “Artemis,” focuses on inference, contrasting with GPUs primarily used for training AI models.
While Meta’s earlier attempts faced challenges, experts anticipate that an inference chip like Artemis could enhance efficiency in processing recommendation models compared to energy-intensive Nvidia processors, potentially leading to substantial cost savings.
The chip, which was announced last year, is a part of Meta’s efforts to bolster its artificial intelligence (AI) capabilities. With this, the company aims to lessen its reliance on Nvidia chips and curb mounting expenses associated with AI development.
The move comes amid Meta’s goal to integrate generative AI products across its platforms such as Facebook, Instagram, WhatsApp, and hardware like Ray-Ban smart glasses.The company has been investing substantially in specialised chips and data centre reconfigurations to run these technologies.
What Meta has to say
A Meta spokesperson confirmed plans to commence production of the updated chip in 2024, saying that the chips complement commercially available GPUs, which Meta continues to procure in large volumes.
“We see our internally developed accelerators to be highly complementary to commercially available GPUs in delivering the optimal mix of performance and efficiency on Meta-specific workloads,” the spokesperson was quoted as saying.
The report said that a successful deployment of its own chip at the scale at which Meta operates may cut hundreds of millions of dollars in annual energy costs and billions in chip purchasing costs.
Meta to purchase 350,000 Nvidia GPUs
Earlier this year, it was reported that the company plans to acquire approximately 350,000 flagship “H100” processors from Nvidia by the end of the year. The new chip, internally dubbed “Artemis,” focuses on inference, contrasting with GPUs primarily used for training AI models.
While Meta’s earlier attempts faced challenges, experts anticipate that an inference chip like Artemis could enhance efficiency in processing recommendation models compared to energy-intensive Nvidia processors, potentially leading to substantial cost savings.
Denial of responsibility! Swift Telecast is an automatic aggregator of the all world’s media. In each content, the hyperlink to the primary source is specified. All trademarks belong to their rightful owners, all materials to their authors. If you are the owner of the content and do not want us to publish your materials, please contact us by email – swifttelecast.com. The content will be deleted within 24 hours.