During the Q2 earnings call for SpaceX, founder Elon Musk responded to a question by stating that the company will fully adopt the NVIDIA architecture platform.

"We expect computing power to exceed 2 gigawatts by the end of this year, and by the end of next year, our cumulative online computing power could be several times what it is now. For example, closer to 10 gigawatts, rather than 5 gigawatts. Looking ahead, we have decided to go entirely with NVIDIA's architecture, as we believe the design by Vera Bubin is the best architecture available," Musk said.

When discussing computing power bottlenecks during the call, Musk pointed out that the biggest pain point is memory, with demand far outpacing supply.

"The current limiting factor is memory. Memory production grows by about 20% annually. That's typically a remarkable pace for any large, mature industry. But ask yourself, is demand really growing at 20% per year? No, demand is growing at 200% or even higher. So, if the growth rate of demand far exceeds that of supply, the basic principles of economics tell us that prices should rise, not fall," Musk explained.