The AI Bubble
Part II: Growing Skepticism
We’re now in a phase of the AI bubble in which many market actors are starting to ask if we’re in one.
By Aiden Singh, September 2, 2025
Aiden Singh is the Founder & Publisher of the Social Science Encyclopedia.
In May 2024, I published an article calling a bubble in AI. The essay argued that, despite widespread optimistic pronouncements, generative AI - while an incremental technological advance - was not a harbinger of a world imminently transformed by artificial general intelligence [AGI]. And I argued that market participants had become overly euphoric about investing in companies even remotely connected to AI.
Since then, others have joined me in calling an AI bubble. In January 2025, Ray Dalio, founder of Bridgewater Associates, told the Financial Times that, “Where we are in the cycle right now is very similar to where we were between 1998 or 1999”, a reference to the late 90s dot-com bubble. In a July 16, 2025 note, Torsten Sløk, chief economist at Apollo Global Management, suggested that the AI bubble today is bigger than the 90s dot-com bubble. And in August, hedge funder Howard Marks stated that we are “in the early days of a bubble”.
And this growing chorus of concern that investor expectations for generative AI have become unmoored from reality are not limited to Wall Street. Over in Silicon Valley, Sam Altman, the CEO of OpenAI - the company behind ChatGPT which lies at the very heart of the AI hype - has recently stated that there may be a bubble in the AI sector. As if summarizing the argument of my May 2024 essay, Altman reportedly stated, “When bubbles happen, smart people get overexcited about a kernel of truth.”
This is in stark contrast to earlier comments by Altman proclaiming that OpenAI was on the verge of achieving AGI. As recently as January of this year Altman said, “We are now confident we know how to build AGI.” And back in September 2023, he posted that OpenAI had, “AGI achieved internally”.
The media too have shifted the tone of their coverage of the AI sector.
In my Part I essay, I highlighted news articles espousing the ‘industrial revolution’ recent innovations in AI were supposedly on the cusp of engendering, like a May 2023 piece in the Guardian entitled AI ‘could be as transformative as Industrial Revolution’, a May 2024 piece in the Wall Street Journal entitled AI Is Driving ‘the Next Industrial Revolution.’ Wall Street Is Cashing In, and a November 2023 article in Barrons that proclaimed, “The rise of artificial intelligence heralds a new stage of the Industrial Revolution, one where machines think, learn, self-replicate, and can master many tasks that were once reserved for humans.”
In contrast, more recent articles on the topic of AI have expressed concern that the market’s expectations for the technology have grown unduly exuberant.
For example, the Guardian recently ran an op-ed with the downright alarming headline, “Is the AI Bubble about to burst - and send the stock market into freefall?”. Likewise, reporting on Nvidia’s most recent earnings release, a Guardian article reads “Nvidia sets fresh sales record amid fears of an AI bubble and Trump’s trade wars”.
Other outlets too have recently struck a tone of consternation. For example, The Hill ran a piece on August 27, 2025 entitled AI bubble worries grow ahead of Nvidia earnings. Forbes ran a piece on the 26th entitled Investors worry about An AI Bubble. And PBS News Hour ran a segment on August 25 asking “Is a bubble forming as AI investments drive economic growth?”
Typifying the waning expectations for an imminent AI-led industrial revolution, Business Insider ran an August 25 piece entitled 3 reasons everyone is talking about an AI bubble. The article reads, “Some are wondering whether large language models are actually powerful enough to develop the long-desired superintelligence”, echoing the argument I made over a year ago that the emergence of generative AI did not imply that we were on the cusp of artificial general intelligence.
The same outlet on the same day ran a second article with the opening sentence, “The race to build artificial general intelligence is colliding with a harsh reality: Large language models might be maxed out”. It continues, “And there's growing concern that LLMs may be nearing their plateau, far from a technology capable of evolving into AGI.”
Also on August 25th, Fortune Magazine ran an article that captures the changing vibe of the conversation around generative AI. It opens, “Once upon a time—meaning, um, as recently as earlier this year—Silicon Valley couldn’t stop talking about AGI.” Now, the conversation is more measured.
What once was unfettered hype about the transformational promise of generative AI has become consternation that generative AI may not be the path towards artificial general intelligence and that AI related asset prices may be excessive. We’re now in a phase of the AI bubble in which many market actors are starting to ask if we’re in one.
- - - - - - -
*A special thanks to my friends historian Ray Ojserkis for many interesting conversations about speculative bubbles that have occurred throughout human history and whether we are currently in an AI bubble & to computer engineer Anton Kosiakin with whom I have had many insightful discussions about the technological limitations of generative AI.
——————