Ten years after its founding with the goal of building a novel chip to challenge Nvidia, Cerebras went public in May this year, with its market value once approaching 100 billion US dollars. CEO Andrew Feldman stated the company is willing to cooperate with every hyperscaler to improve AI inference speed, except for Nvidia. The company's story traces back to 2014, when Andrew Ng left Stanford AI Lab and joined Baidu, which invested 300 million US dollars to establish an AI lab in Silicon Valley. The lab aimed to generate commercial value from deep learning, focusing on physical-level applications like autonomous driving, robotics, and speech recognition. Greg Diamos, now an architect at TensorWave ScalarLM and co-founder of MLCommons, was among the top AI researchers Ng brought to Baidu. Diamos, who previously built the CUDA software architecture at Nvidia when it was a neglected internal project, noted that CUDA was almost a joke at the time. The Baidu team subsequently published a paper titled *Deep Speech*, an end-to-end speech recognition model. Cerebras' early investors, researchers who first proved its products could be deployed, and senior product executives discussed the company's transformation over the past two years and its moves after the IPO.