Google's AI Chip TPU v5p is 2.8 Times Faster in Training LLMs than its Predecessor


Microsoft launches its new generation AI inference chip, Maia 200, with significantly improved performance, remarkable energy efficiency, and cost advantages. The chip uses advanced manufacturing processes and integrates over 100 billion transistors, offering Petaflops-level computing power.
Tesla CEO Musk sues OpenAI and Microsoft, seeking compensation of 930 billion yuan, accusing them of deviating from their non-profit mission and forming a deep partnership with Microsoft. Musk had invested 38 million US dollars and participated in the establishment. The lawsuit reflects his concerns about the future of the AI industry.
Baidu's AI chip subsidiary, Kunlun Chip, has secretly submitted an IPO application in Hong Kong, with an estimated valuation of $3 billion. After the spin-off, it will remain within Baidu's ecosystem. Its predecessor can be traced back to Baidu's internal chip department established in 2012.
Nvidia secures non-exclusive AI chip tech license from Groq for ~$20B, absorbing its core team to sidestep regulatory hurdles, a move seen as strategic defense against Google's TPU competition.....
Google is advancing the "TorchTPU" initiative, aimed at improving the compatibility of its TPU chips with the PyTorch framework, to reduce the cost for developers migrating from NVIDIA GPUs to Google's TPU. This move is intended to challenge NVIDIA's dominance in the AI chip market and break the deep integration between PyTorch and NVIDIA CUDA.