Mark Zuckerberg’s dedication to artificial intelligence (AI) is demonstrated by his considerable investment into Mets’s AI chip development, aiming to bolster its AI data processing capabilities.
As reported by Reuters, Meta is assessing its first internally developed chip, representing a critical milestone in its strategy to fabricate custom silicon and reduce dependence on suppliers like Nvidia. The company has commenced limited use of this chip and plans to enhance production for broader application pending successful evaluations.
This progress is remarkable, as Meta currently employs around 350,000 Nvidia H100 chips for its AI projects, each priced at approximately $25,000. Although these chips are acquired in bulk at discounted rates, Meta intends to significantly augment its investments in AI infrastructure, estimating approximately $65 billion for 2025 to expand data centers for new AI chip integrations.
Meta’s aspirations include improving its processing capabilities with tailored chips, potentially unlocking external opportunities in light of the high demand and restricted availability of H100s due to the current AI surge. While Nvidia has recently succeeded in reducing delivery wait times for H100 units, suggesting market stabilization, the ability to manufacture AI hardware internally could provide Meta with a substantial competitive edge in the future.
The ainewsarticles.com article you just read is a brief synopsis; the original article can be found here: Read the Full Article…
