AI Industry Expansion Need not Cost $7 Trillion, Says NVIDIA CEO
Last week, OpenAI chief, Sam Altman, proposed an initiative to raise $7 trillion from investors for the construction of several new factories for TSMC. This would substantially increase the production of accelerators for artificial intelligence (AI) systems. Contrarily, NVIDIA’s CEO and founder believes that lower expenditures can still foster industry growth.
AI Accelerator Production: A Question of Quantity
Altman’s primary idea revolves around creating a sufficient number of production lines for computation accelerators, as current factory capabilities do not meet industry needs. According to Bloomberg, NVIDIA’s CEO, Jensen Huang, agrees that data processing center construction spending is unlikely to slow over the next few years. However, he is confident this can be achieved with less funds compared to what Altman is asking.
Computational Power over Number
“We are at the beginning of a new era. Currently, the number of functioning data centers worldwide approximates $1 trillion. Over the next four or five years, we will see the amount of data processing centers working with software worldwide equating to $2 trillion,” stated Jensen Huang. Adding that it’s not necessary to acquire more computers, Huang emphasized that computers will become more efficient, thus the overall increase in their number will not be substantial.