AMD and Cerebras Systems have partnered to create a new AI inference solution that integrates AMD Helios rackscale systems with Cerebras Wafer-Scale Engine technology. This collaboration aims to deliver ultra-low latency and high throughput for advanced AI applications by combining AMD's scalable throughput with Cerebras' fast token generation. The joint solution is designed to optimize AI workloads with different latency and throughput needs, offering up to 5 times higher efficiency in tokens per second per watt. Cerebras plans to deploy AMD Helios in its data centers, with availability through Cerebras Cloud expected in the second half of 2026.
AMD reported strong AI-driven revenue growth with its Instinct GPUs and EPYC CPUs, doubling its Data Center revenue in 2026 and expecting further growth in 2027. However, Broadcom poses a significant threat by co-designing custom AI chips with major ...
VanEck Fabless Semiconductor ETF (SMHX) is rated a Buy due to its top holdings' strong positions across AI infrastructure, supporting robust revenue and free cash flow growth. The ETF's concentrated portfolio, with over 45% in NVIDIA, Broadcom, and A...
Nvidia's latest AI chips deliver significantly higher performance than China's best, Huawei's Ascend 910C, which is less than half as powerful as Nvidia's three-year-old H200 chip. Nvidia has since advanced through multiple chip generations, with its...
The August jobs report showed stronger-than-expected job growth with 162,000 jobs added and unemployment steady at 4.1%, easing concerns about labor market weakness. This shifts the Federal Open Market Committee's (FOMC) attention fully to inflation ...
A TradingView analysis warns that AMD stock might face a severe correction of up to 90% by early December 2029, based on a long-term historical trading pattern. This bearish scenario compares AMD's current position in its ascending trading channel to...
