Jensen made basically two visionary choices. The first was to promote GPGPU with CUDA across their entire lineup including consumer hardware. The second was to specifically invest heavily in supporting machine learning way before anyone else in chips cared. CuDNN came out in 2014, just two years after AlexNet. Three years before the transformer. Eight years before ChatGPT. Before PyTorch. Before TensorFlow. Other chip CEOs probably hadn't even heard of AlexNet in 2014.

Assuming these decisions are attributable to him personally, then he absolutely deserves the description "visionary". That said, he's made plenty of other choices I don't like, and also invested in plenty of silly stuff that isn't panning out for them. I don't really think Nvidia needs a visionary right now. They have the best business in the world and they just need to execute. Regardless of whether he deserves to be ousted, the company will be fine without him for the next decade at least.