SK Hynix has become the first memory maker to certify its high-density 256GB DDR5 RDIMMs for Intel’s Xeon 6 server platform. The new modules, based on 32Gb DRAM chips, promise a 16% boost in AI inference speed and 18% better power efficiency. This is a key move to address the massive memory demands
Google is launching a major internal initiative called “TorchTPU” to make its AI chips fully compatible with the PyTorch software framework. The move, developed with Meta’s help, is a direct attempt to erode Nvidia’s software moat and make Google’s TPUs a viable alternative for developers.
NVIDIA is seeing huge demand in China for its last-gen H200 AI GPUs after the US lifted a sales ban. However, the company must pay 25% of each sale to the US government, potentially forcing a price hike. Chinese authorities still need to finalize approval for the chips next week.
Broadcom just reported a monster year fueled by AI chips and VMware, with total revenue hitting $64 billion. The CEO revealed Anthropic is behind a new $11 billion order, and the company’s AI-related backlog now stands at a staggering $73 billion.
Qualcomm is aggressively pushing into automotive and other markets to diversify its revenue. The company’s strategy hinges on AI-powered platforms, and it’s already seeing results with a $45 billion pipeline and quarterly auto revenue surpassing $1 billion.
A new rumor claims Nvidia is targeting a Q1 2027 announcement for its next-gen RTX 60-series GPUs, led by an RTX 6090. The cards would use the Rubin architecture and promise big performance leaps. But with the RTX 50 SUPER series in flux, does this timeline actually make sense?
Nvidia is rolling out new, optional software that lets customers monitor their AI GPU fleets, providing location and health data. The announcement comes as the U.S. tightens export controls on advanced chips. Nvidia is adamant the software is read-only and contains no remote disable function.
AMD’s next-generation laptop processors, the Ryzen AI 400 series, have been spotted in an unreleased chipset driver. This confirms the “Gorgon Point” family as a refresh of the current Ryzen AI 300 lineup.
Johny Srouji, the Apple exec behind the company’s powerful in-house chips, has shot down rumors of his departure. In an internal memo, he stated he has no plans to leave, countering a weekend report from Bloomberg’s Mark Gurman.
Apple’s senior VP of hardware technologies, Johny Srouji, is reportedly “seriously considering” leaving the company. His potential departure follows a wave of other executive exits announced for 2025 and 2026.
