IBM CEO Arvind Krishna has note of caution as the all the big tech companies such as Google, Microsoft and others are pouring billions into building AI infrastructure. As reported by Fortune, speaking at the Decoder podcast earlier this month, Krishan argued that the huge investments made in building massive data centres is unlikely to deliver the returns the tech giants are expecting. Krishna pointed out that even a simple calculation shows the math doesn’t add up. He highlighted that building a data center capable of consuming just one gigawatt of power costs around $80 billion. If a single company commits to around 20 to 30 gigawatts, would mean capital expenditure of around $1.5 trillion which roughly equates to the current market cap of Tesla.So, collectively, the hyperscalers could add up to 100 gigawatts of capacity, but Krishna estimated that would require $8 trillion in investment. “It’s my view that there’s no way you’re going to get a return on that, because $8 trillion of capex means you need roughly $800 billion of profit just to pay for the interest,” he said.

Rise in the demand of energy

The data from Goldman Sachs underscores the challenge, The global data center market presently consumes around 55 gigawatts of power out of which only 14% is dedicated to AI. By the year 2027 which demand can go up to 84 gigawatts, driven by AI workloads. Krishna has now warned that the chips powering these centers could quickly become obsolete, forcing the companies to refresh infrastructure in every five years. “You’ve got to use it all in five years, because at that point, you’ve got to throw it away and refill it,” he explained.

The race towards AGI

Speaking at the podcast, Krishan further suggested that part of the motivation behind these massive investments is the race to achieve artificial general intelligence (AGI). However, he still believes that the odds of reaching AGI with current large language model technology are slim. “There’s at most a 1% chance this feat can be accomplished with our current technology,” he said.Skeptical of AGI timelines, Krishan was also clear about AI’s enterprise potential. “I think it’s incredibly useful for enterprise. I think it’s going to unlock trillions of dollars of productivity in the enterprise, just to be absolutely clear,” he added.Despite the warnings made by Krishna, the companies like Google, Microsoft and others are still accelerating spending. Google parent Alphabet recently raised its 2025 capital expenditure outlook to $91–93 billion, while Amazon lifted its estimate to $125 billion. Industry-wide, AI infrastructure investments are expected to reach $380 billion this year alone.

IBM CEO Arvind Krishna says not due to AI, many companies are laying off

IBM CEO Arvind Krishna has refuted the narrative that the current wave of mass layoffs across the tech industry is primarily due to the artificial intelligence (AI) technology, arguing that the job cuts are largely a “natural correction” following over-hiring during the pandemic years. Krishna, however, acknowledged that AI will eventually lead to some job displacement, but suggested that many companies are currently laying off staff because they overhired between 2020 and 2023.In an interview with The Verge, Krishna used strong terms to describe the rapid staffing increases across the industry during the lockdown era.“If you look at the total employment numbers, I think people gorged on employment… during the pandemic and the year after,” Krishna explained. He noted that some companies saw their employee count spike by “30, 40, 50, 100 percent” from 2020 to 2023.“There is going to be some natural correction. Business is never completely optimized. I think in engineering terms, it’s an underdamped system. When there’s a need, it goes above. Now, it has to correct. It’s probably going to go below what’s needed, and then it’ll hit the correct equilibrium, depending on market demand and growth,” the CEO noted.