Tag
The article discusses Tencent's AI capex constraints due to NVIDIA chip shortages and its recent shift to using Kunlun chips, analyzing the company's valuation and strategic positioning in the AI landscape.
The article analyzes the AI inference ASIC architectures of Groq, SambaNova, Tenstorrent, and Cerebras, highlighting Cerebras's unique wafer-scale engine design. It discusses the benefits of deterministic latency and high bandwidth for LLM inference, while noting challenges like yield, cost, and KV cache bottlenecks.
Google unveiled eighth-gen TPUs (8t/8i) and a new Gemini Enterprise Agent Platform at Cloud Next, while revealing 75% of new Google code is now AI-generated.
Andrew Ng discusses how U.S. policies are driving allies toward sovereign AI and open-source models, referencing DeepSeek, Qwen, and K2 Think as examples. He argues that open-source AI can help nations reduce reliance on U.S. technology.