AI Chips news and analysis on Web Pulse: 6163 curated articles tracking the latest AI Chips developments, tools, and research, updated continuously from vetted sources.
Advanced Micro Devices Inc. announced its next-generation AMD Instinct MI400 Series GPUs and sixth-generation AI infrastructure at its Advancing AI 2026 event in San Francisco, targeting frontier models, agentic workload…
AMD has dismissed Nvidia's CUDA platform as a 'non-event', arguing that companies are increasingly programming at higher abstraction levels where the underlying hardware architecture does not matter. The statement comes …
AMD launched its Helios rackscale system at its Advancing AI event in San Francisco, directly challenging Nvidia's dominance in the AI data center market. CEO Lisa Su announced the platform is in full production, claimin…
AMD introduced the Instinct MI455X GPU with 432GB of HBM4, a 50% increase over NVIDIA's B300 or Rubin, and the 72-GPU Helios rack achieving 2.9 exaFLOPS of FP4, at its Advancing AI event. The company said OpenAI, Meta, A…
AMD CEO Lisa Su unveiled the Helios AI platform at a San Francisco event to compete with Nvidia's Vera Rubin, claiming AMD has an edge in efficiency and cost. The platform includes the new Instinct MI455X GPU and Venice …
AMD unveiled OTel 2.0, an open-source AI model for telecom operators co-developed with AT&T and Microsoft, and AT&T has already processed over one trillion tokens on AMD Instinct GPUs via Microsoft Foundry. Dan McNamara,…
AMD announced the Instinct MI430X GPU for AI supercomputing at its AI event in San Francisco, claiming HPC FP64 compute of 288 teraflops (8.7 times Nvidia Vera Rubin), HBM4 capacity of 432 GB (50% better than Vera Rubin)…
Google has started selling its custom Tensor Processing Units (TPUs) to customers for the first time, delivering them to customer data centers in the second quarter, CFO Anat Ashkenazi disclosed during Alphabet's earning…
Broadcom's Tomahawk 6 switch family is shipping in production volume, while its Davisson co-packaged-optics variant is available to early customers, though Broadcom has not publicly quantified CPO production volumes or n…
AMD and Cerebras announced on July 23, 2026, a joint inference platform that combines AMD Helios racks with Cerebras Wafer-Scale Engine systems, targeting up to 5x higher tokens per second per watt by routing prompt proc…
AMD CEO Lisa Su and Cerebras CEO Andrew Feldman announced a partnership to develop ultra low-latency AI systems, combining AMD's traditional AI chips with Cerebras' wafer-scale integration technology. The first products …
AMD announced the Instinct MI455X, its first GPU designed for rack-scale AI deployments, based on the new CDNA5 architecture and featuring 256 Work Group Processors, 432 GB of HBM4 memory at 23.3 TB/sec, and peak compute…
AMD CEO Lisa Su announced a partnership with chip startup Cerebras to pursue disaggregated inference, splitting AI workloads across different hardware types, as AMD claims its Helios server system delivers up to 30% more…
AMD and Cerebras Systems announced a technical partnership to develop a joint Helios rack-scale AI inference platform, integrating Cerebras' Wafer-Scale Engine with AMD's Helios system to deliver up to five-times higher …
Meta and OpenAI each committed 6 gigawatts of AMD accelerator capacity at AMD's Advancing AI 2026 conference, marking the clearest evidence yet that hyperscalers are building a real alternative to Nvidia. AMD also unveil…
VAST Data is expanding its collaboration with AMD to support AMD EPYC and Instinct processors in its AI OS software, aiming to build and operate high-performance AI factories at scale. The deal includes VAST selecting 6t…
AMD unveiled its next-generation EPYC 9006 Series 'Venice' server CPUs on Thursday, targeting AI agent workloads and high-performance computing. CEO Lisa Su announced full production and said customer demand is the stron…
AMD launched Helios, its first rack-scale AI platform, claiming the 72-GPU system outperforms Nvidia's Vera Rubin by 15-25% in AI training and offers 50% more HBM4 memory and scale-out bandwidth. The Instinct MI455X GPU,…
AMD announced the EPYC 9006 "Venice" server processors at Advancing AI 2026, promising up to 256 cores per socket, support for DDR5-8000 and MRDIMM-12800 memory, PCIe Gen 6, and roughly 20% more performance per core over…
AMD argues that agentic AI workloads require a CPU-to-GPU ratio of 1:1 or CPU-heavy, flipping the traditional GPU-dominant ratio of 1:4 or 1:8, according to technical analysis published by the chipmaker in mid-2026. The …