Pretty sure his criticality to NVDA is the fact that he founded it and has been its CEO for all of its existence. GPGPUs were entirely his vision. Back in 2008 when no one gave a shit about CUDA or GPGPUs he was the only one insisting that massive data processing on GPUs was going to be important for data centers. He kept funding that for another 15 years despite the target audience being niche scientific computing and the market not rewarding him for that. NVDA without him is going to be a different company. That’s obviously bound to happen sooner or later, but Apple without Steve is a different company than Apple with Steve. Could be better, could be worse, but it’ll surly be different.