IBM's Shocking 'Today’s Number' Will Leave Sam Altman Shaking - Are You Prepared?

In a recent discussion on the Decoder podcast, IBM CEO Arvind Krishna delivered a stark warning regarding the escalating costs associated with the pursuit of artificial general intelligence (AGI). As tech giants race to develop massive data centers, Krishna highlighted that the spending required could reach unprecedented levels, potentially threatening even the most established players in the industry, including OpenAI and its CEO, Sam Altman.

Krishna claimed that building a single 1-gigawatt data center could cost around $80 billion, a figure he referred to as “today’s number.” Companies aiming to establish 20–30 gigawatts of infrastructure could face staggering expenditures of approximately $1.5 trillion at current costs. This immense financial burden is compounded by the rapid obsolescence of AI chips, which need replacement every five years, further straining the budgets of these companies.

The Debate Over Returns

During the podcast, Krishna estimated that a total investment of $8 trillion aimed at achieving AGI would require about $800 billion in profits just to cover interest payments. This raises critical questions about whether such extensive spending can ever be justified. In stark contrast, Altman has expressed confidence in OpenAI’s ability to generate returns on its significant capital investments, which include a recent $1.4 trillion long-term buildout agreement with partners like Nvidia, Broadcom, Oracle, and Alphabet.

Despite these optimistic views, Krishna remains skeptical, arguing that the current infrastructure commitments reflect aspirations rather than guaranteed outcomes. He placed the likelihood of current technologies achieving AGI at a mere “zero to 1%,” suggesting that today’s large language models do not meet the criteria for true intelligence. While he acknowledged that existing tools provide substantial enterprise value, he cautioned against assuming they will lead directly to AGI breakthroughs.

In light of Krishna's remarks, it is noteworthy that major companies like Google and Microsoft continue to ramp up their investments. For instance, Google’s parent company, Alphabet, has increased its 2025 capital expenditure outlook to between $91 billion and $93 billion, while Amazon has boosted its estimate to $125 billion. Industry-wide, AI infrastructure investments are projected to hit $380 billion this year alone.

Krishna suggested that achieving AGI would necessitate a combination of large language models and forms of hard knowledge, although he admitted he remains uncertain about whether this approach would be adequate. For now, he advised the industry to focus on the productivity gains that current AI technologies already provide, while simultaneously recognizing that making trillion-dollar bets on infrastructure may not guarantee the breakthroughs many are hoping for.

This ongoing discussion highlights a critical juncture in the tech industry. As companies aggressively pursue the next frontier of AI, the financial implications of their strategies could redefine the landscape of technology as we know it. The balance between ambition and practicality will be essential as this competition unfolds.