AI assisted chip design of the H100 and H200 Hopper AI chips. Jensen wants to use AI to explore combinatorially the entire chip design. He wants AI assisted cross modular codesign.

They would optimize all of the thousands of chip cores that make up the GPUs.

AI is also being used to optimize AI data centers. The new unit of computing is the data center.

Nvidia creates massive digital twins of their entire data centers before they are built and implemented.

They pre-staged everything in the supply chain and created prototypes of sections of the data center.

Nvidia has had a multiyaer project called data center as a product.

Taalas and Etched are two AI chip startups aiming to challenge Nvidia's dominance in the AI hardware market. Etched is using ASIC chips. Taalas is using eASIC chips. They are both putting the logic into the hardware instead of using software on top of GPUs. There is the potential to get 100 to 1000 times more energy efficient than more generalized hardware where the software handles the AI.

It is likely that Nvidia will also apply section of chips with ASIC or FPGA capability to get direct to hardware compute efficiency.

Fully AI designed and optimized chips and data centers will likely provide 10,000 times more performance.