Meta launched Meta Compute on 1 July 2026 to sell surplus AI training and inference capacity, backed by its custom MTIA chips and a planned 2026 capital expenditure of $125 billion to $145 billion, in…
Meta CEO Mark Zuckerberg's AI manifesto promotes the open-source Llama strategy as democratizing AI, but critics note that licensing terms restrict commercial use past a certain scale, making it a fre…
Meta Platforms is in advanced talks to lease computing power to Anthropic in a deal that could be worth $10 billion, according to a New York Times report. The agreement would make Meta a critical infr…
Sandisk stock rose after a leaked memo revealed a multi-year flash supply agreement with Meta Platforms, marking the first known long-term agreement (LTA) in the NAND area. The deal supports Meta's AI…
Meta plans to begin production of its proprietary 'Iris' AI chip in September 2026 as part of a strategy to double computing power to 14 gigawatts by 2027 and deliver personal super intelligence to us…
Meta plans to begin manufacturing its custom AI chip, codenamed Iris, in September as part of a push to reach 14 gigawatts of computing capacity by 2027. The chip, built with Broadcom and TSMC, is cen…
Meta's stock rose 8% after reports that it is exploring a cloud business to sell excess AI computing power and developing a custom AI chip called Iris, easing investor fears about its massive infrastr…
Meta plans to begin production of its in-house MTIA AI chip in September 2025, aiming to reduce reliance on Nvidia GPUs and double its data center computing capacity to 14 gigawatts by 2027, according…
Meta plans to put its own AI chip into production in September and aims to roughly double computing capacity across its data centers, according to Reuters. The chip, part of Meta's MTIA line, is desig…
According to a May 2026 report from Tom's Hardware, the custom AI ASIC market is being driven by major deals from Broadcom and the continued development of in-house chips like Google's TPUs and Meta's…
Meta researchers developed In-Kernel Broadcast Optimization (IKBO), a kernel-model-system co-design that eliminates redundant user-embedding replication during recommendation model inference. Deployed…