AI Chips
5 articles RSS
Huawei Says Ascend Has Overtaken Nvidia in China, Pointing to Supply Rather Than Demand as the Limit
Huawei's rotating chairman Eric Xu says Ascend has surpassed Nvidia in China by Huawei's own data, while DeepSeek released an Ascend port of its DeepGEMM kernels.
Qualcomm Unveils Dragonfly Data Center Roadmap: A 250-Core C1000 Server CPU and AI300 Accelerator, With Meta Signed as a CPU Customer
At its June 24 Investor Day, Qualcomm detailed a multi-generation data center chip line built for agentic AI inference, anchored by the Dragonfly C1000 CPU and AI300 accelerator, and named Meta as a server-CPU customer.
Alibaba's T-Head Unveils Zhenwu M890 AI Accelerator With 144GB Memory and a 128-Chip Supernode, Claiming 3x Its Prior Generation
Alibaba's chip unit T-Head detailed the Zhenwu M890, a 144GB AI accelerator it says triples the prior 810E, alongside a 128-chip supernode server.
NVIDIA Hand-Delivers First Vera CPUs to Anthropic, OpenAI, and SpaceXAI as Agentic AI Drives a New CPU Moment
NVIDIA VP Ian Buck personally delivered the first production Vera CPU systems to Anthropic, OpenAI, SpaceXAI, and Oracle in California, marking the commercial arrival of NVIDIA's first purpose-built agentic AI processor.
Google Splits Its Eighth-Generation TPU Into Training and Inference Chips as Anthropic Locks in Up to a Million Units
At Cloud Next 2026, Google unveiled the TPU 8t and TPU 8i, its first-ever split TPU line, with Anthropic as anchor customer for up to one million chips across 2026 and 3.5 gigawatts in 2027.