IBM CEO says there is 'no way' spending on AI data centers will pay off (businessinsider.com)

Article summary

IBM CEO Arvind Krishna expressed skepticism about the financial viability of spending trillions of dollars on AI data centers, citing the high costs of building and maintaining them. He estimated that it takes around $80 billion to fill up a one-gigawatt data center and questioned whether the revenue generated would be enough to justify the investment. Krishna also noted that the current technology may not be sufficient to achieve Artificial General Intelligence (AGI). The comments section discussed the validity of Krishna's math and the potential for future technological advancements to improve efficiency and reduce costs.

Main themes

• AI data center costs
• Financial viability
• Technological advancements
• Artificial General Intelligence
• Data center infrastructure
• GPU efficiency

What commenters say

• The IBM CEO's argument against investing in AI data centers is based on flawed assumptions and a lack of understanding of the industry.
• The high costs of building and maintaining AI data centers may not be justified by the potential revenue generated.
• Future technological advancements could improve the efficiency and reduce the costs of AI data centers, making them more financially viable.
• The current focus on achieving AGI may be misguided, and alternative approaches or technologies could be more effective.
• The cost of GPUs and other hardware is a significant factor in the overall cost of AI data centers, and improvements in power efficiency could be crucial to their financial viability.
• The potential for older GPUs to be repurposed or reused in other applications could reduce the need for frequent upgrades and replacements.
• The AI industry is driven by a 'bubble' of hype and investment, which may not be sustainable in the long term.
• The question of whether newer GPUs will be worth the upgrade is uncertain, and depends on factors such as power efficiency and performance gains.