Keep pulling the thread on Jensen Huang.
Jensen Huang predicts that the number of AI agents and the number of software tool users will grow exponentially.
Jensen Huang predicts that the use of software tools from companies like Synopsys will skyrocket due to their adoption by AI agents.
According to its latest filings, NVIDIA had almost $100 billion in purchase commitments with foundries, memory suppliers, and packaging companies.
Semi-Analysis has reported that NVIDIA will have $250 billion in purchase commitments for scarce components.
According to some analysis, AI chip production will consume 60% of TSMC's N3 node capacity this year and is projected to consume 86% next year.
Jensen Huang views restrictive energy policies as a more significant long-term bottleneck to the growth of the AI industry than semiconductor manufacturing capacity.
NVIDIA's Blackwell GPU architecture is 30 to 50 times more energy efficient than the previous Hopper architecture for some workloads.
Two of the world's top three AI models, Claude and Gemini, were trained on Google's TPUs.
Jensen Huang claims that NVIDIA's computing stack offers the best performance per Total Cost of Ownership (TCO) and the highest tokens per watt of any platform in the world.
Anthropic recently announced a multi-gigawatt deal with Broadcom and Google for TPUs.
Jensen Huang asserts that Anthropic's use of Google TPUs and Amazon Trainium is almost entirely due to large, direct investments made by Google and AWS into the company.
Jensen Huang states his past mistake was not realizing that foundation model labs like Anthropic required multi-billion dollar investments directly from their compute suppliers, as traditional VC funding was insufficient.