Sam Altman admits AI bubble, plans trillions for data centers
OpenAI CEO Sam Altman recently acknowledged that the artificial intelligence sector is experiencing a significant bubble, yet he remains undeterred, signaling his intention to dominate whatever landscape emerges after its inevitable correction. Speaking to a select group of reporters at a dinner, Altman defined a bubble as a phenomenon where intelligent individuals become overly enthusiastic about an idea that, despite its potential, contains an element of irrational exuberance. By this measure, he asserted, the AI industry is indeed in the throes of a rapidly expanding, potentially volatile bubble.
Drawing parallels to historical economic booms, Altman referenced the dot-com era, noting that while investors grew “overexcited,” the underlying technology – the internet – was undeniably transformative and crucial. He suggested that AI embodies a similar dichotomy: a period of investor frenzy coinciding with what he considers “the most important thing to happen in a very long time.” This perspective implies that just as the dot-com bust didn’t extinguish the internet but rather paved the way for trillion-dollar companies built on its foundational ideas like e-commerce and search, the AI sector is likely to survive an implosion that could wipe out countless startups.
One critical factor contributing to this potential bubble burst, according to Altman, is a looming shortage of graphics processing units (GPUs). This scarcity has already influenced OpenAI’s strategic decisions, leading the company to design its latest model, ChatGPT-5, with a primary focus on optimizing inference costs rather than raw computational power. Beyond GPUs, the massive investments poured into datacenters by tech giants and colocation providers also present a dual risk. While essential for AI’s growth, these very datacenters could become a bottleneck themselves, potentially ensnared in the wider economic repercussions of a bursting bubble.
Despite these acknowledgements, Altman’s demeanor suggests little concern about the bubble’s potential collapse; rather, he appears focused on shaping its aftermath. He revealed ambitious plans for OpenAI to invest “trillions of dollars on data center construction in the not very distant future.” This colossal expenditure is deemed necessary to unlock the full potential of OpenAI’s advancements, enabling the company to deploy “better models” and introduce “other kinds of new products and services” currently hampered by insufficient capacity.
However, the source of such immense capital remains a pertinent question. While OpenAI has secured tens of billions in commitments from major players like Microsoft, Softbank, and Oracle, this sum pales in comparison to the proposed multi-trillion-dollar investment. The company is projected to generate $10 billion in revenue this year, a significant leap from $5.5 billion last year. Yet, it also reported a $5 billion loss last year, as per Reuters. While impressive for a company that was largely off investors’ radar before the launch of ChatGPT, these figures are a mere fraction of the cash flows enjoyed by established tech behemoths like Microsoft, Amazon, or Google. Thus, while Altman readily admits to the existence of an AI bubble, his declared intent to pour trillions more into its foundational infrastructure suggests a deliberate strategy: to not only ride the wave but to actively inflate it, aiming to be the dominant force once the froth settles.