AI Industry's Profit Problem: High Costs, Low Returns
Despite the stratospheric ascent of leading AI companies like OpenAI, Microsoft, and Nvidia on the stock market, a deeper examination reveals a more complex and inconvenient truth: the artificial intelligence industry is currently far from profitable, and its path to financial viability remains profoundly uncertain. Even the most ardent AI proponents, in recent interviews with the New York Times, have struggled to articulate a clear route to profitability for the technology and its surrounding ecosystem.
Andrew McAfee, an MIT research scientist and founder of an AI consultancy, highlighted this disconnect, lamenting that while AI’s “raw technological horsepower is terrific,” it alone will not dictate the pace of AI’s economic transformation. A primary impediment to AI becoming the lucrative investment many hoped for is its staggering operational cost, which is projected to escalate further as operations expand.
A report earlier this year from McKinsey underscored this financial challenge, estimating that by 2030, AI data centers will need to spend an astonishing $6.7 trillion on computing power to meet surging demand. When contrasted with estimates from software company Hartinger, which peg the AI industry’s total market size at approximately $305.9 billion by the close of 2025, the sheer scale of the required investment becomes starkly apparent. It is difficult to envision trillions of dollars flowing into the industry within the next five years, let alone beyond that. As tech journalist Ed Zitron observed earlier this year, OpenAI, for instance, reportedly spent its entire $4 billion revenue on running and training its models.
While AI’s relentless hype cycles often promise groundbreaking models that will propel the world closer to artificial general intelligence (AGI) or human-level intelligence, AI companies have consistently fallen short of these ambitious goals. A notable example of these diminishing returns was the highly anticipated launch of OpenAI’s GPT-5, which ultimately proved to be a disappointment. With each successive model that fails to deliver a significant leap forward, a growing chorus of critics suggests that AI progress may have reached a plateau.
This sobering reality has begun to ripple through the executive suites of companies that eagerly invested in AI. A report released in May by work management software company Asana, based on a survey of nearly 4,000 IT professionals, found that 29 percent — or roughly one in three — companies that adopted AI in 2024 now regret their decision. As the report succinctly put it, “the 2024 rush to deploy AI has given rise to a sobering reality.”
Such regrets are already translating into concrete actions. A similar survey conducted in March by S&P Global Market Intelligence, polling 1,000 companies that had invested in AI, revealed that a striking 42 percent had already abandoned their AI endeavors. This marks a significant increase from the 17 percent who halted their AI projects in 2024. Lori Beer, Chief Information Officer at JPMorgan, told the New York Times that following the bank’s decision to restrict staff use of ChatGPT, she has since curtailed hundreds of other AI projects. “We’re absolutely shutting things down,” Beer stated, emphasizing, “We’re not afraid to shut things down. We don’t think it’s a bad thing. I think it’s a smart thing.”
While AI proponents like McAfee might frame these investment failures as part of the innovative process, arguing that “innovation is a process of failing fairly regularly,” the mounting evidence of colossal costs and limited tangible returns makes it increasingly difficult to view the AI industry as anything other than a massively inflated bubble poised for a correction.